Hugging Face's logo

Spaces:
pjramg
/
openvino_anomalib
Runtime error

App Files Community
pjramg commited on
Commit
b5e1135
·
1 Parent(s): 7095157

CVPR_2023_OpenVINO_Anomalib

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. app.py +132 -0
  2. bottle/examples/000.png +0 -0
  3. bottle/examples/007.png +0 -0
  4. bottle/examples/010.png +0 -0
  5. bottle/examples/018.png +0 -0
  6. bottle/run/weights/lightning/model.ckpt +3 -0
  7. bottle/run/weights/openvino/metadata.json +52 -0
  8. bottle/run/weights/openvino/model.bin +3 -0
  9. bottle/run/weights/openvino/model.mapping +387 -0
  10. bottle/run/weights/openvino/model.onnx +3 -0
  11. bottle/run/weights/openvino/model.xml +2703 -0
  12. cubes/examples/001.jpg +0 -0
  13. cubes/examples/002.jpg +0 -0
  14. cubes/examples/003.jpg +0 -0
  15. cubes/examples/004.jpg +0 -0
  16. cubes/examples/005.jpg +0 -0
  17. cubes/examples/006.jpg +0 -0
  18. cubes/examples/007.jpg +0 -0
  19. cubes/examples/008.jpg +0 -0
  20. cubes/run/weights/openvino/metadata.json +52 -0
  21. cubes/run/weights/openvino/model.bin +3 -0
  22. cubes/run/weights/openvino/model.mapping +387 -0
  23. cubes/run/weights/openvino/model.onnx +3 -0
  24. cubes/run/weights/openvino/model.xml +2703 -0
  25. grid/examples/001.png +0 -0
  26. grid/examples/005.png +0 -0
  27. grid/examples/006.png +0 -0
  28. grid/examples/007.png +0 -0
  29. grid/examples/009.png +0 -0
  30. grid/examples/010.png +0 -0
  31. grid/run/weights/lightning/model.ckpt +3 -0
  32. grid/run/weights/openvino/metadata.json +52 -0
  33. grid/run/weights/openvino/model.bin +3 -0
  34. grid/run/weights/openvino/model.mapping +387 -0
  35. grid/run/weights/openvino/model.onnx +3 -0
  36. grid/run/weights/openvino/model.xml +2703 -0
  37. pill/examples/001.png +0 -0
  38. pill/examples/007.png +0 -0
  39. pill/examples/013.png +0 -0
  40. pill/examples/014.png +0 -0
  41. pill/examples/019.png +0 -0
  42. pill/examples/021.png +0 -0
  43. pill/examples/022.png +0 -0
  44. pill/run/weights/lightning/model.ckpt +3 -0
  45. pill/run/weights/openvino/metadata.json +52 -0
  46. pill/run/weights/openvino/model.bin +3 -0
  47. pill/run/weights/openvino/model.mapping +387 -0
  48. pill/run/weights/openvino/model.onnx +3 -0
  49. pill/run/weights/openvino/model.xml +2703 -0
  50. requirements.txt +1 -0
app.py ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import time
2
+
3
+ import gradio as gr
4
+ import numpy as np
5
+ from pathlib import Path
6
+ import time
7
+ from anomalib.deploy import OpenVINOInferencer
8
+ from openvino.runtime import Core
9
+ # Initialize the Core
10
+ core = Core()
11
+
12
+ # Get the available devices
13
+ devices = core.available_devices
14
+
15
+ inferencer = None
16
+
17
+ example_list = [["bottle/examples/000.png", "anomaly_map", "bottle", "CPU"],
18
+ ["pill/examples/010.png", "heat_map", "pill", "CPU"],
19
+ ["zipper/examples/001.png", "pred_mask", "zipper", "CPU"],
20
+ ["grid/examples/005.png", "segmentations", "grid", "CPU"],
21
+ ["cubes/examples/005.jpg", "heat_map", "cubes", "CPU"]]
22
+
23
+ def OV_compilemodel(category_choice, device):
24
+ global inferencer
25
+ #Get the available models
26
+ openvino_model_path = Path.cwd() / category_choice / "run" / "weights" / "openvino" / "model.bin"
27
+ metadata_path = Path.cwd() / category_choice / "run" / "weights" / "openvino" / "metadata.json"
28
+
29
+ inferencer = OpenVINOInferencer(
30
+ path=openvino_model_path, # Path to the OpenVINO IR model.
31
+ metadata_path=metadata_path, # Path to the metadata file.
32
+ device=device, # We would like to run it on an Intel CPU.
33
+ config= {"INFERENCE_PRECISION_HINT": "f16" } if device != "CPU" else {}
34
+ )
35
+
36
+ return inferencer
37
+
38
+ def OV_inference(input_img, operation, category_choice, device):
39
+
40
+ start_time = time.time()
41
+ predictions = inferencer.predict(image=input_img)
42
+ stop_time = time.time()
43
+ inference_time = stop_time - start_time
44
+ confidence = predictions.pred_score
45
+
46
+ if operation == "original":
47
+ output_img1 = predictions.image
48
+ elif operation == "anomaly_map":
49
+ output_img1 = predictions.anomaly_map
50
+ elif operation == "heat_map":
51
+ output_img1 = predictions.heat_map
52
+ elif operation == "pred_mask":
53
+ output_img1 = predictions.pred_mask
54
+ elif operation == "segmentations":
55
+ output_img1 = predictions.segmentations
56
+ else:
57
+ output_img1 = predictions.image
58
+ return output_img1, round(inference_time*1000), round(confidence*100,2)
59
+
60
+ with gr.Blocks() as demo:
61
+ gr.Markdown(
62
+ """
63
+ <img align="left" width="150" src= "https://github.com/openvinotoolkit/anomalib/assets/10940214/7e61a627-d1b0-4ad4-b602-da9b348c0cbe">
64
+ <img align="right" width="150" src= "https://github.com/openvinotoolkit/anomalib/assets/10940214/5d6dd038-b40c-441f-ad38-1cf526137de2">
65
+
66
+ <h1 align="center"> 🚀 Anomaly detection 🚀 </h1>
67
+
68
+ Experience the power of the state-of-the-art anomaly detection with Anomalib-OpenVINO Anomaly detection toolbox. This interactive APP leverages the robust capabilities of Anomalib and OpenVINO.
69
+
70
+ All model are FP32 precision, if you select GPU it will automatically change precision to FP16. Using Anomalib you can also quantize your model in INT8 using NNCF.
71
+
72
+
73
+
74
+ ![](https://github.com/openvinotoolkit/anomalib/assets/10940214/ce78346f-4d27-4f99-bea7-75b87e2ac02a)
75
+
76
+
77
+
78
+ """
79
+ )
80
+
81
+ gr.Markdown("## 1. Select the category over you want to detect anormalities.")
82
+ category_choice = gr.Radio(["bottle", "grid", "pill", "zipper", "cubes"], label="Choose the category")
83
+
84
+ gr.Markdown(
85
+ """
86
+ ## 2. Select the Intel device
87
+ Device Name | CPU | GPU.0 | GPU.1
88
+ ------------- | ------------ |------------- | -------------
89
+ Intel Device | CPU | Integrated GPU | Discrete GPU
90
+
91
+
92
+ """
93
+ )
94
+ device_choice = gr.Dropdown(devices, label="Choose the device")
95
+
96
+ gr.Markdown("## 3. Compile the model")
97
+ compile_btn = gr.Button("Compile Model")
98
+
99
+ gr.Markdown("## 4. Choose the output you want to visualize.")
100
+ output_choice = gr.Radio(["original", "anomaly_map", "heat_map", "pred_mask", "segmentations"], label="Choose the output")
101
+
102
+ gr.Markdown("## 5. Drop the image in the input image box and run the inference")
103
+ with gr.Row():
104
+ with gr.Column():
105
+ image = gr.Image(type="numpy", label= "Input image")
106
+
107
+
108
+ with gr.Column():
109
+ output_img = gr.outputs.Image(type="numpy", label="Anomalib Output")
110
+
111
+ inference_btn = gr.Button("Run Inference")
112
+
113
+ with gr.Row():
114
+ # Create your output components
115
+ #output_prediction = gr.Textbox(label="Prediction")
116
+ output_confidence = gr.Textbox(label="Confidence [%]")
117
+ output_time = gr.Textbox(label="Inference Time [ms]")
118
+
119
+ gr.Markdown("Note: Change the image and run the inference again. If you want to change the object you need to recompile the model, that means you need to start from step 1.")
120
+ gr.Markdown("## Image Examples")
121
+
122
+ gr.Examples(
123
+ examples=example_list,
124
+ inputs=[image, output_choice, category_choice, device_choice],
125
+ outputs=[output_img, output_time, output_confidence],
126
+ fn=OV_inference,
127
+ )
128
+
129
+ compile_btn.click(OV_compilemodel, inputs=[category_choice, device_choice])
130
+ inference_btn.click(OV_inference, inputs=[image, output_choice], outputs=[output_img, output_time, output_confidence])
131
+
132
+ demo.launch(share=True, enable_queue=True)
bottle/examples/000.png ADDED
bottle/examples/007.png ADDED
bottle/examples/010.png ADDED
bottle/examples/018.png ADDED
bottle/run/weights/lightning/model.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30ef6e09877b0d7be1e0d8e7492f203ffaa33caea9b7009b5dc7137df8220815
3
+ size 186461500
bottle/run/weights/openvino/metadata.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "task": "segmentation",
3
+ "transform": {
4
+ "__version__": "1.3.0",
5
+ "transform": {
6
+ "__class_fullname__": "Compose",
7
+ "p": 1.0,
8
+ "transforms": [
9
+ {
10
+ "__class_fullname__": "Resize",
11
+ "always_apply": true,
12
+ "p": 1,
13
+ "height": 256,
14
+ "width": 256,
15
+ "interpolation": 1
16
+ },
17
+ {
18
+ "__class_fullname__": "Normalize",
19
+ "always_apply": false,
20
+ "p": 1.0,
21
+ "mean": [
22
+ 0.485,
23
+ 0.456,
24
+ 0.406
25
+ ],
26
+ "std": [
27
+ 0.229,
28
+ 0.224,
29
+ 0.225
30
+ ],
31
+ "max_pixel_value": 255.0
32
+ },
33
+ {
34
+ "__class_fullname__": "ToTensorV2",
35
+ "always_apply": true,
36
+ "p": 1.0,
37
+ "transpose_mask": false
38
+ }
39
+ ],
40
+ "bbox_params": null,
41
+ "keypoint_params": null,
42
+ "additional_targets": {
43
+ "image": "image",
44
+ "depth_image": "image"
45
+ }
46
+ }
47
+ },
48
+ "image_threshold": 14.50792121887207,
49
+ "pixel_threshold": 12.323184967041016,
50
+ "min": 0.09382443130016327,
51
+ "max": 54.39400100708008
52
+ }
bottle/run/weights/openvino/model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ee1cb6a5c082dd8669cb8ee94ac985324c49d5172ba0c3fe4f7e3cc612a8da7
3
+ size 176605956
bottle/run/weights/openvino/model.mapping ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <mapping>
3
+ <map>
4
+ <framework name="input" output_port_id="input" />
5
+ <IR name="input" output_port_id="0" />
6
+ </map>
7
+ <map>
8
+ <framework name="onnx::Conv_260" output_port_id="onnx::Conv_260" />
9
+ <IR name="onnx::Conv_260" output_port_id="0" />
10
+ </map>
11
+ <map>
12
+ <framework name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/conv1/Conv_output_0" />
13
+ <IR name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="2" />
14
+ </map>
15
+ <map>
16
+ <framework name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="/feature_extractor/feature_extractor/act1/Relu_output_0" />
17
+ <IR name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="1" />
18
+ </map>
19
+ <map>
20
+ <framework name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0" />
21
+ <IR name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="1" />
22
+ </map>
23
+ <map>
24
+ <framework name="onnx::Conv_263" output_port_id="onnx::Conv_263" />
25
+ <IR name="onnx::Conv_263" output_port_id="0" />
26
+ </map>
27
+ <map>
28
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0" />
29
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="2" />
30
+ </map>
31
+ <map>
32
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0" />
33
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="1" />
34
+ </map>
35
+ <map>
36
+ <framework name="onnx::Conv_266" output_port_id="onnx::Conv_266" />
37
+ <IR name="onnx::Conv_266" output_port_id="0" />
38
+ </map>
39
+ <map>
40
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0" />
41
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="2" />
42
+ </map>
43
+ <map>
44
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0" />
45
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="2" />
46
+ </map>
47
+ <map>
48
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0" />
49
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="1" />
50
+ </map>
51
+ <map>
52
+ <framework name="onnx::Conv_269" output_port_id="onnx::Conv_269" />
53
+ <IR name="onnx::Conv_269" output_port_id="0" />
54
+ </map>
55
+ <map>
56
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0" />
57
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="2" />
58
+ </map>
59
+ <map>
60
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0" />
61
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="1" />
62
+ </map>
63
+ <map>
64
+ <framework name="onnx::Conv_272" output_port_id="onnx::Conv_272" />
65
+ <IR name="onnx::Conv_272" output_port_id="0" />
66
+ </map>
67
+ <map>
68
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0" />
69
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="2" />
70
+ </map>
71
+ <map>
72
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0" />
73
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="2" />
74
+ </map>
75
+ <map>
76
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0" />
77
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="1" />
78
+ </map>
79
+ <map>
80
+ <framework name="onnx::Conv_275" output_port_id="onnx::Conv_275" />
81
+ <IR name="onnx::Conv_275" output_port_id="0" />
82
+ </map>
83
+ <map>
84
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0" />
85
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="2" />
86
+ </map>
87
+ <map>
88
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0" />
89
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="1" />
90
+ </map>
91
+ <map>
92
+ <framework name="onnx::Conv_278" output_port_id="onnx::Conv_278" />
93
+ <IR name="onnx::Conv_278" output_port_id="0" />
94
+ </map>
95
+ <map>
96
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0" />
97
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="2" />
98
+ </map>
99
+ <map>
100
+ <framework name="onnx::Conv_281" output_port_id="onnx::Conv_281" />
101
+ <IR name="onnx::Conv_281" output_port_id="0" />
102
+ </map>
103
+ <map>
104
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0" />
105
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="2" />
106
+ </map>
107
+ <map>
108
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0" />
109
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="2" />
110
+ </map>
111
+ <map>
112
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0" />
113
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="1" />
114
+ </map>
115
+ <map>
116
+ <framework name="onnx::Conv_284" output_port_id="onnx::Conv_284" />
117
+ <IR name="onnx::Conv_284" output_port_id="0" />
118
+ </map>
119
+ <map>
120
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0" />
121
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="2" />
122
+ </map>
123
+ <map>
124
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0" />
125
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="1" />
126
+ </map>
127
+ <map>
128
+ <framework name="onnx::Conv_287" output_port_id="onnx::Conv_287" />
129
+ <IR name="onnx::Conv_287" output_port_id="0" />
130
+ </map>
131
+ <map>
132
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0" />
133
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="2" />
134
+ </map>
135
+ <map>
136
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0" />
137
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="2" />
138
+ </map>
139
+ <map>
140
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0" />
141
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="1" />
142
+ </map>
143
+ <map>
144
+ <framework name="/Shape" output_port_id="/Shape_output_0" />
145
+ <IR name="/Shape" output_port_id="1" />
146
+ </map>
147
+ <map>
148
+ <framework name="/Constant_1" output_port_id="/Constant_1_output_0" />
149
+ <IR name="/Constant_1" output_port_id="0" />
150
+ </map>
151
+ <map>
152
+ <framework name="/Constant_2" output_port_id="/Constant_2_output_0" />
153
+ <IR name="/Constant_2" output_port_id="0" />
154
+ </map>
155
+ <map>
156
+ <framework name="/Slice" output_port_id="/Slice_output_0" />
157
+ <IR name="/Slice" output_port_id="4" />
158
+ </map>
159
+ <map>
160
+ <framework name="/Constant_3" output_port_id="/Constant_3_output_0" />
161
+ <IR name="/Constant_3" output_port_id="0" />
162
+ </map>
163
+ <map>
164
+ <framework name="/Concat" output_port_id="/Concat_output_0" />
165
+ <IR name="/Concat" output_port_id="2" />
166
+ </map>
167
+ <map>
168
+ <framework name="/Resize" output_port_id="/Resize_output_0" />
169
+ <IR name="/Resize" output_port_id="3" />
170
+ </map>
171
+ <map>
172
+ <framework name="/Concat_1" output_port_id="/Concat_1_output_0" />
173
+ <IR name="/Concat_1" output_port_id="2" />
174
+ </map>
175
+ <map>
176
+ <framework name="onnx::Conv_290" output_port_id="onnx::Conv_290" />
177
+ <IR name="onnx::Conv_290" output_port_id="0" />
178
+ </map>
179
+ <map>
180
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0" />
181
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="2" />
182
+ </map>
183
+ <map>
184
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0" />
185
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="1" />
186
+ </map>
187
+ <map>
188
+ <framework name="onnx::Conv_293" output_port_id="onnx::Conv_293" />
189
+ <IR name="onnx::Conv_293" output_port_id="0" />
190
+ </map>
191
+ <map>
192
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0" />
193
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="2" />
194
+ </map>
195
+ <map>
196
+ <framework name="onnx::Conv_296" output_port_id="onnx::Conv_296" />
197
+ <IR name="onnx::Conv_296" output_port_id="0" />
198
+ </map>
199
+ <map>
200
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0" />
201
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="2" />
202
+ </map>
203
+ <map>
204
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0" />
205
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="2" />
206
+ </map>
207
+ <map>
208
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0" />
209
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="1" />
210
+ </map>
211
+ <map>
212
+ <framework name="onnx::Conv_299" output_port_id="onnx::Conv_299" />
213
+ <IR name="onnx::Conv_299" output_port_id="0" />
214
+ </map>
215
+ <map>
216
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0" />
217
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="2" />
218
+ </map>
219
+ <map>
220
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0" />
221
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="1" />
222
+ </map>
223
+ <map>
224
+ <framework name="onnx::Conv_302" output_port_id="onnx::Conv_302" />
225
+ <IR name="onnx::Conv_302" output_port_id="0" />
226
+ </map>
227
+ <map>
228
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0" />
229
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="2" />
230
+ </map>
231
+ <map>
232
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0" />
233
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="2" />
234
+ </map>
235
+ <map>
236
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0" />
237
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="1" />
238
+ </map>
239
+ <map>
240
+ <framework name="/Shape_3" output_port_id="/Shape_3_output_0" />
241
+ <IR name="/Shape_3" output_port_id="1" />
242
+ </map>
243
+ <map>
244
+ <framework name="/Constant_9" output_port_id="/Constant_9_output_0" />
245
+ <IR name="/Constant_9" output_port_id="0" />
246
+ </map>
247
+ <map>
248
+ <framework name="/Constant_10" output_port_id="/Constant_10_output_0" />
249
+ <IR name="/Constant_10" output_port_id="0" />
250
+ </map>
251
+ <map>
252
+ <framework name="/Slice_1" output_port_id="/Slice_1_output_0" />
253
+ <IR name="/Slice_1" output_port_id="4" />
254
+ </map>
255
+ <map>
256
+ <framework name="/Shape_1" output_port_id="/Shape_1_output_0" />
257
+ <IR name="/Shape_1" output_port_id="1" />
258
+ </map>
259
+ <map>
260
+ <framework name="/Shape_1" output_port_id="/Shape_2_output_0" />
261
+ <IR name="/Shape_1" output_port_id="1" />
262
+ </map>
263
+ <map>
264
+ <framework name="Gather_6092" output_port_id="/Concat_2_output_0" />
265
+ <IR name="Gather_6092" output_port_id="3" />
266
+ </map>
267
+ <map>
268
+ <framework name="Gather_6092" output_port_id="/Cast_output_0" />
269
+ <IR name="Gather_6092" output_port_id="3" />
270
+ </map>
271
+ <map>
272
+ <framework name="/Concat_3" output_port_id="/Concat_3_output_0" />
273
+ <IR name="/Concat_3" output_port_id="2" />
274
+ </map>
275
+ <map>
276
+ <framework name="/Resize_1" output_port_id="/Resize_1_output_0" />
277
+ <IR name="/Resize_1" output_port_id="3" />
278
+ </map>
279
+ <map>
280
+ <framework name="/Concat_4" output_port_id="/Concat_4_output_0" />
281
+ <IR name="/Concat_4" output_port_id="2" />
282
+ </map>
283
+ <map>
284
+ <framework name="onnx::Gather_308" output_port_id="onnx::Gather_308" />
285
+ <IR name="onnx::Gather_308" output_port_id="0" />
286
+ </map>
287
+ <map>
288
+ <framework name="/Gather_2" output_port_id="/Gather_2_output_0" />
289
+ <IR name="/Gather_2" output_port_id="3" />
290
+ </map>
291
+ <map>
292
+ <framework name="/anomaly_map_generator/Constant" output_port_id="/anomaly_map_generator/Constant_output_0" />
293
+ <IR name="/anomaly_map_generator/Constant" output_port_id="0" />
294
+ </map>
295
+ <map>
296
+ <framework name="/anomaly_map_generator/Reshape" output_port_id="/anomaly_map_generator/Reshape_output_0" />
297
+ <IR name="/anomaly_map_generator/Reshape" output_port_id="2" />
298
+ </map>
299
+ <map>
300
+ <framework name="/anomaly_map_generator/Sub" output_port_id="/anomaly_map_generator/Sub_output_0" />
301
+ <IR name="/anomaly_map_generator/Sub" output_port_id="2" />
302
+ </map>
303
+ <map>
304
+ <framework name="/anomaly_map_generator/Transpose" output_port_id="/anomaly_map_generator/Transpose_output_0" />
305
+ <IR name="/anomaly_map_generator/Transpose" output_port_id="2" />
306
+ </map>
307
+ <map>
308
+ <framework name="onnx::MatMul_310" output_port_id="onnx::MatMul_310" />
309
+ <IR name="onnx::MatMul_310" output_port_id="0" />
310
+ </map>
311
+ <map>
312
+ <framework name="/anomaly_map_generator/MatMul" output_port_id="/anomaly_map_generator/MatMul_output_0" />
313
+ <IR name="/anomaly_map_generator/MatMul" output_port_id="2" />
314
+ </map>
315
+ <map>
316
+ <framework name="/anomaly_map_generator/Mul" output_port_id="/anomaly_map_generator/Mul_output_0" />
317
+ <IR name="/anomaly_map_generator/Mul" output_port_id="2" />
318
+ </map>
319
+ <map>
320
+ <framework name="/anomaly_map_generator/ReduceSum" output_port_id="/anomaly_map_generator/ReduceSum_output_0" />
321
+ <IR name="/anomaly_map_generator/ReduceSum" output_port_id="2" />
322
+ </map>
323
+ <map>
324
+ <framework name="/anomaly_map_generator/Constant_1" output_port_id="/anomaly_map_generator/Constant_1_output_0" />
325
+ <IR name="/anomaly_map_generator/Constant_1" output_port_id="0" />
326
+ </map>
327
+ <map>
328
+ <framework name="/anomaly_map_generator/Reshape_1" output_port_id="/anomaly_map_generator/Reshape_1_output_0" />
329
+ <IR name="/anomaly_map_generator/Reshape_1" output_port_id="2" />
330
+ </map>
331
+ <map>
332
+ <framework name="/anomaly_map_generator/Clip" output_port_id="/anomaly_map_generator/Clip_output_0" />
333
+ <IR name="/anomaly_map_generator/Clip" output_port_id="1" />
334
+ </map>
335
+ <map>
336
+ <framework name="/anomaly_map_generator/Sqrt" output_port_id="/anomaly_map_generator/Sqrt_output_0" />
337
+ <IR name="/anomaly_map_generator/Sqrt" output_port_id="1" />
338
+ </map>
339
+ <map>
340
+ <framework name="/anomaly_map_generator/Shape" output_port_id="/anomaly_map_generator/Shape_output_0" />
341
+ <IR name="/anomaly_map_generator/Shape" output_port_id="1" />
342
+ </map>
343
+ <map>
344
+ <framework name="/anomaly_map_generator/Constant_4" output_port_id="/anomaly_map_generator/Constant_4_output_0" />
345
+ <IR name="/anomaly_map_generator/Constant_4" output_port_id="0" />
346
+ </map>
347
+ <map>
348
+ <framework name="/anomaly_map_generator/Constant_5" output_port_id="/anomaly_map_generator/Constant_5_output_0" />
349
+ <IR name="/anomaly_map_generator/Constant_5" output_port_id="0" />
350
+ </map>
351
+ <map>
352
+ <framework name="/anomaly_map_generator/Slice" output_port_id="/anomaly_map_generator/Slice_output_0" />
353
+ <IR name="/anomaly_map_generator/Slice" output_port_id="4" />
354
+ </map>
355
+ <map>
356
+ <framework name="/anomaly_map_generator/Constant_6" output_port_id="/anomaly_map_generator/Constant_6_output_0" />
357
+ <IR name="/anomaly_map_generator/Constant_6" output_port_id="0" />
358
+ </map>
359
+ <map>
360
+ <framework name="/anomaly_map_generator/Concat" output_port_id="/anomaly_map_generator/Concat_output_0" />
361
+ <IR name="/anomaly_map_generator/Concat" output_port_id="2" />
362
+ </map>
363
+ <map>
364
+ <framework name="/anomaly_map_generator/Resize" output_port_id="/anomaly_map_generator/Resize_output_0" />
365
+ <IR name="/anomaly_map_generator/Resize" output_port_id="3" />
366
+ </map>
367
+ <map>
368
+ <framework name="/anomaly_map_generator/blur/Pad" output_port_id="/anomaly_map_generator/blur/Pad_output_0" />
369
+ <IR name="/anomaly_map_generator/blur/Pad" output_port_id="4" />
370
+ </map>
371
+ <map>
372
+ <framework name="anomaly_map_generator.blur.kernel" output_port_id="anomaly_map_generator.blur.kernel" />
373
+ <IR name="anomaly_map_generator.blur.kernel" output_port_id="0" />
374
+ </map>
375
+ <map>
376
+ <framework name="/anomaly_map_generator/blur/Conv" output_port_id="/anomaly_map_generator/blur/Conv_output_0" />
377
+ <IR name="/anomaly_map_generator/blur/Conv" output_port_id="2" />
378
+ </map>
379
+ <map>
380
+ <framework name="/anomaly_map_generator/blur/Constant_8" output_port_id="/anomaly_map_generator/blur/Constant_8_output_0" />
381
+ <IR name="/anomaly_map_generator/blur/Constant_8" output_port_id="0" />
382
+ </map>
383
+ <map>
384
+ <framework name="output" output_port_id="output" />
385
+ <IR name="output" output_port_id="2" />
386
+ </map>
387
+ </mapping>
bottle/run/weights/openvino/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87ab506e58426d2e495b1f4c4807843a0c9b161a2b045bccf4d3092a05665b85
3
+ size 176626689
bottle/run/weights/openvino/model.xml ADDED
@@ -0,0 +1,2703 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="torch_jit" version="11">
3
+ <layers>
4
+ <layer id="0" name="input" type="Parameter" version="opset1">
5
+ <data shape="1,3,256,256" element_type="f32" />
6
+ <output>
7
+ <port id="0" precision="FP32" names="input">
8
+ <dim>1</dim>
9
+ <dim>3</dim>
10
+ <dim>256</dim>
11
+ <dim>256</dim>
12
+ </port>
13
+ </output>
14
+ </layer>
15
+ <layer id="1" name="onnx::Conv_260" type="Const" version="opset1">
16
+ <data element_type="f32" shape="64, 3, 7, 7" offset="0" size="37632" />
17
+ <output>
18
+ <port id="0" precision="FP32" names="onnx::Conv_260">
19
+ <dim>64</dim>
20
+ <dim>3</dim>
21
+ <dim>7</dim>
22
+ <dim>7</dim>
23
+ </port>
24
+ </output>
25
+ </layer>
26
+ <layer id="2" name="/feature_extractor/feature_extractor/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
27
+ <data strides="2, 2" dilations="1, 1" pads_begin="3, 3" pads_end="3, 3" auto_pad="explicit" />
28
+ <input>
29
+ <port id="0" precision="FP32">
30
+ <dim>1</dim>
31
+ <dim>3</dim>
32
+ <dim>256</dim>
33
+ <dim>256</dim>
34
+ </port>
35
+ <port id="1" precision="FP32">
36
+ <dim>64</dim>
37
+ <dim>3</dim>
38
+ <dim>7</dim>
39
+ <dim>7</dim>
40
+ </port>
41
+ </input>
42
+ <output>
43
+ <port id="2" precision="FP32">
44
+ <dim>1</dim>
45
+ <dim>64</dim>
46
+ <dim>128</dim>
47
+ <dim>128</dim>
48
+ </port>
49
+ </output>
50
+ </layer>
51
+ <layer id="3" name="Reshape_55" type="Const" version="opset1">
52
+ <data element_type="f32" shape="1, 64, 1, 1" offset="37632" size="256" />
53
+ <output>
54
+ <port id="0" precision="FP32">
55
+ <dim>1</dim>
56
+ <dim>64</dim>
57
+ <dim>1</dim>
58
+ <dim>1</dim>
59
+ </port>
60
+ </output>
61
+ </layer>
62
+ <layer id="4" name="/feature_extractor/feature_extractor/conv1/Conv" type="Add" version="opset1">
63
+ <data auto_broadcast="numpy" />
64
+ <input>
65
+ <port id="0" precision="FP32">
66
+ <dim>1</dim>
67
+ <dim>64</dim>
68
+ <dim>128</dim>
69
+ <dim>128</dim>
70
+ </port>
71
+ <port id="1" precision="FP32">
72
+ <dim>1</dim>
73
+ <dim>64</dim>
74
+ <dim>1</dim>
75
+ <dim>1</dim>
76
+ </port>
77
+ </input>
78
+ <output>
79
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/conv1/Conv_output_0">
80
+ <dim>1</dim>
81
+ <dim>64</dim>
82
+ <dim>128</dim>
83
+ <dim>128</dim>
84
+ </port>
85
+ </output>
86
+ </layer>
87
+ <layer id="5" name="/feature_extractor/feature_extractor/act1/Relu" type="ReLU" version="opset1">
88
+ <input>
89
+ <port id="0" precision="FP32">
90
+ <dim>1</dim>
91
+ <dim>64</dim>
92
+ <dim>128</dim>
93
+ <dim>128</dim>
94
+ </port>
95
+ </input>
96
+ <output>
97
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/act1/Relu_output_0">
98
+ <dim>1</dim>
99
+ <dim>64</dim>
100
+ <dim>128</dim>
101
+ <dim>128</dim>
102
+ </port>
103
+ </output>
104
+ </layer>
105
+ <layer id="6" name="/feature_extractor/feature_extractor/maxpool/MaxPool" type="MaxPool" version="opset8">
106
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" kernel="3, 3" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
107
+ <input>
108
+ <port id="0" precision="FP32">
109
+ <dim>1</dim>
110
+ <dim>64</dim>
111
+ <dim>128</dim>
112
+ <dim>128</dim>
113
+ </port>
114
+ </input>
115
+ <output>
116
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0">
117
+ <dim>1</dim>
118
+ <dim>64</dim>
119
+ <dim>64</dim>
120
+ <dim>64</dim>
121
+ </port>
122
+ <port id="2" precision="I64">
123
+ <dim>1</dim>
124
+ <dim>64</dim>
125
+ <dim>64</dim>
126
+ <dim>64</dim>
127
+ </port>
128
+ </output>
129
+ </layer>
130
+ <layer id="7" name="onnx::Conv_263" type="Const" version="opset1">
131
+ <data element_type="f32" shape="64, 64, 3, 3" offset="37888" size="147456" />
132
+ <output>
133
+ <port id="0" precision="FP32" names="onnx::Conv_263">
134
+ <dim>64</dim>
135
+ <dim>64</dim>
136
+ <dim>3</dim>
137
+ <dim>3</dim>
138
+ </port>
139
+ </output>
140
+ </layer>
141
+ <layer id="8" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
142
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
143
+ <input>
144
+ <port id="0" precision="FP32">
145
+ <dim>1</dim>
146
+ <dim>64</dim>
147
+ <dim>64</dim>
148
+ <dim>64</dim>
149
+ </port>
150
+ <port id="1" precision="FP32">
151
+ <dim>64</dim>
152
+ <dim>64</dim>
153
+ <dim>3</dim>
154
+ <dim>3</dim>
155
+ </port>
156
+ </input>
157
+ <output>
158
+ <port id="2" precision="FP32">
159
+ <dim>1</dim>
160
+ <dim>64</dim>
161
+ <dim>64</dim>
162
+ <dim>64</dim>
163
+ </port>
164
+ </output>
165
+ </layer>
166
+ <layer id="9" name="Reshape_105" type="Const" version="opset1">
167
+ <data element_type="f32" shape="1, 64, 1, 1" offset="185344" size="256" />
168
+ <output>
169
+ <port id="0" precision="FP32">
170
+ <dim>1</dim>
171
+ <dim>64</dim>
172
+ <dim>1</dim>
173
+ <dim>1</dim>
174
+ </port>
175
+ </output>
176
+ </layer>
177
+ <layer id="10" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" type="Add" version="opset1">
178
+ <data auto_broadcast="numpy" />
179
+ <input>
180
+ <port id="0" precision="FP32">
181
+ <dim>1</dim>
182
+ <dim>64</dim>
183
+ <dim>64</dim>
184
+ <dim>64</dim>
185
+ </port>
186
+ <port id="1" precision="FP32">
187
+ <dim>1</dim>
188
+ <dim>64</dim>
189
+ <dim>1</dim>
190
+ <dim>1</dim>
191
+ </port>
192
+ </input>
193
+ <output>
194
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0">
195
+ <dim>1</dim>
196
+ <dim>64</dim>
197
+ <dim>64</dim>
198
+ <dim>64</dim>
199
+ </port>
200
+ </output>
201
+ </layer>
202
+ <layer id="11" name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" type="ReLU" version="opset1">
203
+ <input>
204
+ <port id="0" precision="FP32">
205
+ <dim>1</dim>
206
+ <dim>64</dim>
207
+ <dim>64</dim>
208
+ <dim>64</dim>
209
+ </port>
210
+ </input>
211
+ <output>
212
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0">
213
+ <dim>1</dim>
214
+ <dim>64</dim>
215
+ <dim>64</dim>
216
+ <dim>64</dim>
217
+ </port>
218
+ </output>
219
+ </layer>
220
+ <layer id="12" name="onnx::Conv_266" type="Const" version="opset1">
221
+ <data element_type="f32" shape="64, 64, 3, 3" offset="185600" size="147456" />
222
+ <output>
223
+ <port id="0" precision="FP32" names="onnx::Conv_266">
224
+ <dim>64</dim>
225
+ <dim>64</dim>
226
+ <dim>3</dim>
227
+ <dim>3</dim>
228
+ </port>
229
+ </output>
230
+ </layer>
231
+ <layer id="13" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
232
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
233
+ <input>
234
+ <port id="0" precision="FP32">
235
+ <dim>1</dim>
236
+ <dim>64</dim>
237
+ <dim>64</dim>
238
+ <dim>64</dim>
239
+ </port>
240
+ <port id="1" precision="FP32">
241
+ <dim>64</dim>
242
+ <dim>64</dim>
243
+ <dim>3</dim>
244
+ <dim>3</dim>
245
+ </port>
246
+ </input>
247
+ <output>
248
+ <port id="2" precision="FP32">
249
+ <dim>1</dim>
250
+ <dim>64</dim>
251
+ <dim>64</dim>
252
+ <dim>64</dim>
253
+ </port>
254
+ </output>
255
+ </layer>
256
+ <layer id="14" name="Reshape_154" type="Const" version="opset1">
257
+ <data element_type="f32" shape="1, 64, 1, 1" offset="333056" size="256" />
258
+ <output>
259
+ <port id="0" precision="FP32">
260
+ <dim>1</dim>
261
+ <dim>64</dim>
262
+ <dim>1</dim>
263
+ <dim>1</dim>
264
+ </port>
265
+ </output>
266
+ </layer>
267
+ <layer id="15" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" type="Add" version="opset1">
268
+ <data auto_broadcast="numpy" />
269
+ <input>
270
+ <port id="0" precision="FP32">
271
+ <dim>1</dim>
272
+ <dim>64</dim>
273
+ <dim>64</dim>
274
+ <dim>64</dim>
275
+ </port>
276
+ <port id="1" precision="FP32">
277
+ <dim>1</dim>
278
+ <dim>64</dim>
279
+ <dim>1</dim>
280
+ <dim>1</dim>
281
+ </port>
282
+ </input>
283
+ <output>
284
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0">
285
+ <dim>1</dim>
286
+ <dim>64</dim>
287
+ <dim>64</dim>
288
+ <dim>64</dim>
289
+ </port>
290
+ </output>
291
+ </layer>
292
+ <layer id="16" name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" type="Add" version="opset1">
293
+ <data auto_broadcast="numpy" />
294
+ <input>
295
+ <port id="0" precision="FP32">
296
+ <dim>1</dim>
297
+ <dim>64</dim>
298
+ <dim>64</dim>
299
+ <dim>64</dim>
300
+ </port>
301
+ <port id="1" precision="FP32">
302
+ <dim>1</dim>
303
+ <dim>64</dim>
304
+ <dim>64</dim>
305
+ <dim>64</dim>
306
+ </port>
307
+ </input>
308
+ <output>
309
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0">
310
+ <dim>1</dim>
311
+ <dim>64</dim>
312
+ <dim>64</dim>
313
+ <dim>64</dim>
314
+ </port>
315
+ </output>
316
+ </layer>
317
+ <layer id="17" name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" type="ReLU" version="opset1">
318
+ <input>
319
+ <port id="0" precision="FP32">
320
+ <dim>1</dim>
321
+ <dim>64</dim>
322
+ <dim>64</dim>
323
+ <dim>64</dim>
324
+ </port>
325
+ </input>
326
+ <output>
327
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0">
328
+ <dim>1</dim>
329
+ <dim>64</dim>
330
+ <dim>64</dim>
331
+ <dim>64</dim>
332
+ </port>
333
+ </output>
334
+ </layer>
335
+ <layer id="18" name="onnx::Conv_269" type="Const" version="opset1">
336
+ <data element_type="f32" shape="64, 64, 3, 3" offset="333312" size="147456" />
337
+ <output>
338
+ <port id="0" precision="FP32" names="onnx::Conv_269">
339
+ <dim>64</dim>
340
+ <dim>64</dim>
341
+ <dim>3</dim>
342
+ <dim>3</dim>
343
+ </port>
344
+ </output>
345
+ </layer>
346
+ <layer id="19" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
347
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
348
+ <input>
349
+ <port id="0" precision="FP32">
350
+ <dim>1</dim>
351
+ <dim>64</dim>
352
+ <dim>64</dim>
353
+ <dim>64</dim>
354
+ </port>
355
+ <port id="1" precision="FP32">
356
+ <dim>64</dim>
357
+ <dim>64</dim>
358
+ <dim>3</dim>
359
+ <dim>3</dim>
360
+ </port>
361
+ </input>
362
+ <output>
363
+ <port id="2" precision="FP32">
364
+ <dim>1</dim>
365
+ <dim>64</dim>
366
+ <dim>64</dim>
367
+ <dim>64</dim>
368
+ </port>
369
+ </output>
370
+ </layer>
371
+ <layer id="20" name="Reshape_204" type="Const" version="opset1">
372
+ <data element_type="f32" shape="1, 64, 1, 1" offset="480768" size="256" />
373
+ <output>
374
+ <port id="0" precision="FP32">
375
+ <dim>1</dim>
376
+ <dim>64</dim>
377
+ <dim>1</dim>
378
+ <dim>1</dim>
379
+ </port>
380
+ </output>
381
+ </layer>
382
+ <layer id="21" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" type="Add" version="opset1">
383
+ <data auto_broadcast="numpy" />
384
+ <input>
385
+ <port id="0" precision="FP32">
386
+ <dim>1</dim>
387
+ <dim>64</dim>
388
+ <dim>64</dim>
389
+ <dim>64</dim>
390
+ </port>
391
+ <port id="1" precision="FP32">
392
+ <dim>1</dim>
393
+ <dim>64</dim>
394
+ <dim>1</dim>
395
+ <dim>1</dim>
396
+ </port>
397
+ </input>
398
+ <output>
399
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0">
400
+ <dim>1</dim>
401
+ <dim>64</dim>
402
+ <dim>64</dim>
403
+ <dim>64</dim>
404
+ </port>
405
+ </output>
406
+ </layer>
407
+ <layer id="22" name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" type="ReLU" version="opset1">
408
+ <input>
409
+ <port id="0" precision="FP32">
410
+ <dim>1</dim>
411
+ <dim>64</dim>
412
+ <dim>64</dim>
413
+ <dim>64</dim>
414
+ </port>
415
+ </input>
416
+ <output>
417
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0">
418
+ <dim>1</dim>
419
+ <dim>64</dim>
420
+ <dim>64</dim>
421
+ <dim>64</dim>
422
+ </port>
423
+ </output>
424
+ </layer>
425
+ <layer id="23" name="onnx::Conv_272" type="Const" version="opset1">
426
+ <data element_type="f32" shape="64, 64, 3, 3" offset="481024" size="147456" />
427
+ <output>
428
+ <port id="0" precision="FP32" names="onnx::Conv_272">
429
+ <dim>64</dim>
430
+ <dim>64</dim>
431
+ <dim>3</dim>
432
+ <dim>3</dim>
433
+ </port>
434
+ </output>
435
+ </layer>
436
+ <layer id="24" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
437
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
438
+ <input>
439
+ <port id="0" precision="FP32">
440
+ <dim>1</dim>
441
+ <dim>64</dim>
442
+ <dim>64</dim>
443
+ <dim>64</dim>
444
+ </port>
445
+ <port id="1" precision="FP32">
446
+ <dim>64</dim>
447
+ <dim>64</dim>
448
+ <dim>3</dim>
449
+ <dim>3</dim>
450
+ </port>
451
+ </input>
452
+ <output>
453
+ <port id="2" precision="FP32">
454
+ <dim>1</dim>
455
+ <dim>64</dim>
456
+ <dim>64</dim>
457
+ <dim>64</dim>
458
+ </port>
459
+ </output>
460
+ </layer>
461
+ <layer id="25" name="Reshape_253" type="Const" version="opset1">
462
+ <data element_type="f32" shape="1, 64, 1, 1" offset="628480" size="256" />
463
+ <output>
464
+ <port id="0" precision="FP32">
465
+ <dim>1</dim>
466
+ <dim>64</dim>
467
+ <dim>1</dim>
468
+ <dim>1</dim>
469
+ </port>
470
+ </output>
471
+ </layer>
472
+ <layer id="26" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" type="Add" version="opset1">
473
+ <data auto_broadcast="numpy" />
474
+ <input>
475
+ <port id="0" precision="FP32">
476
+ <dim>1</dim>
477
+ <dim>64</dim>
478
+ <dim>64</dim>
479
+ <dim>64</dim>
480
+ </port>
481
+ <port id="1" precision="FP32">
482
+ <dim>1</dim>
483
+ <dim>64</dim>
484
+ <dim>1</dim>
485
+ <dim>1</dim>
486
+ </port>
487
+ </input>
488
+ <output>
489
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0">
490
+ <dim>1</dim>
491
+ <dim>64</dim>
492
+ <dim>64</dim>
493
+ <dim>64</dim>
494
+ </port>
495
+ </output>
496
+ </layer>
497
+ <layer id="27" name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" type="Add" version="opset1">
498
+ <data auto_broadcast="numpy" />
499
+ <input>
500
+ <port id="0" precision="FP32">
501
+ <dim>1</dim>
502
+ <dim>64</dim>
503
+ <dim>64</dim>
504
+ <dim>64</dim>
505
+ </port>
506
+ <port id="1" precision="FP32">
507
+ <dim>1</dim>
508
+ <dim>64</dim>
509
+ <dim>64</dim>
510
+ <dim>64</dim>
511
+ </port>
512
+ </input>
513
+ <output>
514
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0">
515
+ <dim>1</dim>
516
+ <dim>64</dim>
517
+ <dim>64</dim>
518
+ <dim>64</dim>
519
+ </port>
520
+ </output>
521
+ </layer>
522
+ <layer id="28" name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" type="ReLU" version="opset1">
523
+ <input>
524
+ <port id="0" precision="FP32">
525
+ <dim>1</dim>
526
+ <dim>64</dim>
527
+ <dim>64</dim>
528
+ <dim>64</dim>
529
+ </port>
530
+ </input>
531
+ <output>
532
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0">
533
+ <dim>1</dim>
534
+ <dim>64</dim>
535
+ <dim>64</dim>
536
+ <dim>64</dim>
537
+ </port>
538
+ </output>
539
+ </layer>
540
+ <layer id="29" name="onnx::Conv_275" type="Const" version="opset1">
541
+ <data element_type="f32" shape="128, 64, 3, 3" offset="628736" size="294912" />
542
+ <output>
543
+ <port id="0" precision="FP32" names="onnx::Conv_275">
544
+ <dim>128</dim>
545
+ <dim>64</dim>
546
+ <dim>3</dim>
547
+ <dim>3</dim>
548
+ </port>
549
+ </output>
550
+ </layer>
551
+ <layer id="30" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
552
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
553
+ <input>
554
+ <port id="0" precision="FP32">
555
+ <dim>1</dim>
556
+ <dim>64</dim>
557
+ <dim>64</dim>
558
+ <dim>64</dim>
559
+ </port>
560
+ <port id="1" precision="FP32">
561
+ <dim>128</dim>
562
+ <dim>64</dim>
563
+ <dim>3</dim>
564
+ <dim>3</dim>
565
+ </port>
566
+ </input>
567
+ <output>
568
+ <port id="2" precision="FP32">
569
+ <dim>1</dim>
570
+ <dim>128</dim>
571
+ <dim>32</dim>
572
+ <dim>32</dim>
573
+ </port>
574
+ </output>
575
+ </layer>
576
+ <layer id="31" name="Reshape_303" type="Const" version="opset1">
577
+ <data element_type="f32" shape="1, 128, 1, 1" offset="923648" size="512" />
578
+ <output>
579
+ <port id="0" precision="FP32">
580
+ <dim>1</dim>
581
+ <dim>128</dim>
582
+ <dim>1</dim>
583
+ <dim>1</dim>
584
+ </port>
585
+ </output>
586
+ </layer>
587
+ <layer id="32" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" type="Add" version="opset1">
588
+ <data auto_broadcast="numpy" />
589
+ <input>
590
+ <port id="0" precision="FP32">
591
+ <dim>1</dim>
592
+ <dim>128</dim>
593
+ <dim>32</dim>
594
+ <dim>32</dim>
595
+ </port>
596
+ <port id="1" precision="FP32">
597
+ <dim>1</dim>
598
+ <dim>128</dim>
599
+ <dim>1</dim>
600
+ <dim>1</dim>
601
+ </port>
602
+ </input>
603
+ <output>
604
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0">
605
+ <dim>1</dim>
606
+ <dim>128</dim>
607
+ <dim>32</dim>
608
+ <dim>32</dim>
609
+ </port>
610
+ </output>
611
+ </layer>
612
+ <layer id="33" name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" type="ReLU" version="opset1">
613
+ <input>
614
+ <port id="0" precision="FP32">
615
+ <dim>1</dim>
616
+ <dim>128</dim>
617
+ <dim>32</dim>
618
+ <dim>32</dim>
619
+ </port>
620
+ </input>
621
+ <output>
622
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0">
623
+ <dim>1</dim>
624
+ <dim>128</dim>
625
+ <dim>32</dim>
626
+ <dim>32</dim>
627
+ </port>
628
+ </output>
629
+ </layer>
630
+ <layer id="34" name="onnx::Conv_278" type="Const" version="opset1">
631
+ <data element_type="f32" shape="128, 128, 3, 3" offset="924160" size="589824" />
632
+ <output>
633
+ <port id="0" precision="FP32" names="onnx::Conv_278">
634
+ <dim>128</dim>
635
+ <dim>128</dim>
636
+ <dim>3</dim>
637
+ <dim>3</dim>
638
+ </port>
639
+ </output>
640
+ </layer>
641
+ <layer id="35" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
642
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
643
+ <input>
644
+ <port id="0" precision="FP32">
645
+ <dim>1</dim>
646
+ <dim>128</dim>
647
+ <dim>32</dim>
648
+ <dim>32</dim>
649
+ </port>
650
+ <port id="1" precision="FP32">
651
+ <dim>128</dim>
652
+ <dim>128</dim>
653
+ <dim>3</dim>
654
+ <dim>3</dim>
655
+ </port>
656
+ </input>
657
+ <output>
658
+ <port id="2" precision="FP32">
659
+ <dim>1</dim>
660
+ <dim>128</dim>
661
+ <dim>32</dim>
662
+ <dim>32</dim>
663
+ </port>
664
+ </output>
665
+ </layer>
666
+ <layer id="36" name="Reshape_352" type="Const" version="opset1">
667
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1513984" size="512" />
668
+ <output>
669
+ <port id="0" precision="FP32">
670
+ <dim>1</dim>
671
+ <dim>128</dim>
672
+ <dim>1</dim>
673
+ <dim>1</dim>
674
+ </port>
675
+ </output>
676
+ </layer>
677
+ <layer id="37" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" type="Add" version="opset1">
678
+ <data auto_broadcast="numpy" />
679
+ <input>
680
+ <port id="0" precision="FP32">
681
+ <dim>1</dim>
682
+ <dim>128</dim>
683
+ <dim>32</dim>
684
+ <dim>32</dim>
685
+ </port>
686
+ <port id="1" precision="FP32">
687
+ <dim>1</dim>
688
+ <dim>128</dim>
689
+ <dim>1</dim>
690
+ <dim>1</dim>
691
+ </port>
692
+ </input>
693
+ <output>
694
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0">
695
+ <dim>1</dim>
696
+ <dim>128</dim>
697
+ <dim>32</dim>
698
+ <dim>32</dim>
699
+ </port>
700
+ </output>
701
+ </layer>
702
+ <layer id="38" name="onnx::Conv_281" type="Const" version="opset1">
703
+ <data element_type="f32" shape="128, 64, 1, 1" offset="1514496" size="32768" />
704
+ <output>
705
+ <port id="0" precision="FP32" names="onnx::Conv_281">
706
+ <dim>128</dim>
707
+ <dim>64</dim>
708
+ <dim>1</dim>
709
+ <dim>1</dim>
710
+ </port>
711
+ </output>
712
+ </layer>
713
+ <layer id="39" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
714
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
715
+ <input>
716
+ <port id="0" precision="FP32">
717
+ <dim>1</dim>
718
+ <dim>64</dim>
719
+ <dim>64</dim>
720
+ <dim>64</dim>
721
+ </port>
722
+ <port id="1" precision="FP32">
723
+ <dim>128</dim>
724
+ <dim>64</dim>
725
+ <dim>1</dim>
726
+ <dim>1</dim>
727
+ </port>
728
+ </input>
729
+ <output>
730
+ <port id="2" precision="FP32">
731
+ <dim>1</dim>
732
+ <dim>128</dim>
733
+ <dim>32</dim>
734
+ <dim>32</dim>
735
+ </port>
736
+ </output>
737
+ </layer>
738
+ <layer id="40" name="Reshape_400" type="Const" version="opset1">
739
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1547264" size="512" />
740
+ <output>
741
+ <port id="0" precision="FP32">
742
+ <dim>1</dim>
743
+ <dim>128</dim>
744
+ <dim>1</dim>
745
+ <dim>1</dim>
746
+ </port>
747
+ </output>
748
+ </layer>
749
+ <layer id="41" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" type="Add" version="opset1">
750
+ <data auto_broadcast="numpy" />
751
+ <input>
752
+ <port id="0" precision="FP32">
753
+ <dim>1</dim>
754
+ <dim>128</dim>
755
+ <dim>32</dim>
756
+ <dim>32</dim>
757
+ </port>
758
+ <port id="1" precision="FP32">
759
+ <dim>1</dim>
760
+ <dim>128</dim>
761
+ <dim>1</dim>
762
+ <dim>1</dim>
763
+ </port>
764
+ </input>
765
+ <output>
766
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0">
767
+ <dim>1</dim>
768
+ <dim>128</dim>
769
+ <dim>32</dim>
770
+ <dim>32</dim>
771
+ </port>
772
+ </output>
773
+ </layer>
774
+ <layer id="42" name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" type="Add" version="opset1">
775
+ <data auto_broadcast="numpy" />
776
+ <input>
777
+ <port id="0" precision="FP32">
778
+ <dim>1</dim>
779
+ <dim>128</dim>
780
+ <dim>32</dim>
781
+ <dim>32</dim>
782
+ </port>
783
+ <port id="1" precision="FP32">
784
+ <dim>1</dim>
785
+ <dim>128</dim>
786
+ <dim>32</dim>
787
+ <dim>32</dim>
788
+ </port>
789
+ </input>
790
+ <output>
791
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0">
792
+ <dim>1</dim>
793
+ <dim>128</dim>
794
+ <dim>32</dim>
795
+ <dim>32</dim>
796
+ </port>
797
+ </output>
798
+ </layer>
799
+ <layer id="43" name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" type="ReLU" version="opset1">
800
+ <input>
801
+ <port id="0" precision="FP32">
802
+ <dim>1</dim>
803
+ <dim>128</dim>
804
+ <dim>32</dim>
805
+ <dim>32</dim>
806
+ </port>
807
+ </input>
808
+ <output>
809
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0">
810
+ <dim>1</dim>
811
+ <dim>128</dim>
812
+ <dim>32</dim>
813
+ <dim>32</dim>
814
+ </port>
815
+ </output>
816
+ </layer>
817
+ <layer id="44" name="onnx::Conv_284" type="Const" version="opset1">
818
+ <data element_type="f32" shape="128, 128, 3, 3" offset="1547776" size="589824" />
819
+ <output>
820
+ <port id="0" precision="FP32" names="onnx::Conv_284">
821
+ <dim>128</dim>
822
+ <dim>128</dim>
823
+ <dim>3</dim>
824
+ <dim>3</dim>
825
+ </port>
826
+ </output>
827
+ </layer>
828
+ <layer id="45" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
829
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
830
+ <input>
831
+ <port id="0" precision="FP32">
832
+ <dim>1</dim>
833
+ <dim>128</dim>
834
+ <dim>32</dim>
835
+ <dim>32</dim>
836
+ </port>
837
+ <port id="1" precision="FP32">
838
+ <dim>128</dim>
839
+ <dim>128</dim>
840
+ <dim>3</dim>
841
+ <dim>3</dim>
842
+ </port>
843
+ </input>
844
+ <output>
845
+ <port id="2" precision="FP32">
846
+ <dim>1</dim>
847
+ <dim>128</dim>
848
+ <dim>32</dim>
849
+ <dim>32</dim>
850
+ </port>
851
+ </output>
852
+ </layer>
853
+ <layer id="46" name="Reshape_450" type="Const" version="opset1">
854
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2137600" size="512" />
855
+ <output>
856
+ <port id="0" precision="FP32">
857
+ <dim>1</dim>
858
+ <dim>128</dim>
859
+ <dim>1</dim>
860
+ <dim>1</dim>
861
+ </port>
862
+ </output>
863
+ </layer>
864
+ <layer id="47" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" type="Add" version="opset1">
865
+ <data auto_broadcast="numpy" />
866
+ <input>
867
+ <port id="0" precision="FP32">
868
+ <dim>1</dim>
869
+ <dim>128</dim>
870
+ <dim>32</dim>
871
+ <dim>32</dim>
872
+ </port>
873
+ <port id="1" precision="FP32">
874
+ <dim>1</dim>
875
+ <dim>128</dim>
876
+ <dim>1</dim>
877
+ <dim>1</dim>
878
+ </port>
879
+ </input>
880
+ <output>
881
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0">
882
+ <dim>1</dim>
883
+ <dim>128</dim>
884
+ <dim>32</dim>
885
+ <dim>32</dim>
886
+ </port>
887
+ </output>
888
+ </layer>
889
+ <layer id="48" name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" type="ReLU" version="opset1">
890
+ <input>
891
+ <port id="0" precision="FP32">
892
+ <dim>1</dim>
893
+ <dim>128</dim>
894
+ <dim>32</dim>
895
+ <dim>32</dim>
896
+ </port>
897
+ </input>
898
+ <output>
899
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0">
900
+ <dim>1</dim>
901
+ <dim>128</dim>
902
+ <dim>32</dim>
903
+ <dim>32</dim>
904
+ </port>
905
+ </output>
906
+ </layer>
907
+ <layer id="49" name="onnx::Conv_287" type="Const" version="opset1">
908
+ <data element_type="f32" shape="128, 128, 3, 3" offset="2138112" size="589824" />
909
+ <output>
910
+ <port id="0" precision="FP32" names="onnx::Conv_287">
911
+ <dim>128</dim>
912
+ <dim>128</dim>
913
+ <dim>3</dim>
914
+ <dim>3</dim>
915
+ </port>
916
+ </output>
917
+ </layer>
918
+ <layer id="50" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
919
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
920
+ <input>
921
+ <port id="0" precision="FP32">
922
+ <dim>1</dim>
923
+ <dim>128</dim>
924
+ <dim>32</dim>
925
+ <dim>32</dim>
926
+ </port>
927
+ <port id="1" precision="FP32">
928
+ <dim>128</dim>
929
+ <dim>128</dim>
930
+ <dim>3</dim>
931
+ <dim>3</dim>
932
+ </port>
933
+ </input>
934
+ <output>
935
+ <port id="2" precision="FP32">
936
+ <dim>1</dim>
937
+ <dim>128</dim>
938
+ <dim>32</dim>
939
+ <dim>32</dim>
940
+ </port>
941
+ </output>
942
+ </layer>
943
+ <layer id="51" name="Reshape_499" type="Const" version="opset1">
944
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2727936" size="512" />
945
+ <output>
946
+ <port id="0" precision="FP32">
947
+ <dim>1</dim>
948
+ <dim>128</dim>
949
+ <dim>1</dim>
950
+ <dim>1</dim>
951
+ </port>
952
+ </output>
953
+ </layer>
954
+ <layer id="52" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" type="Add" version="opset1">
955
+ <data auto_broadcast="numpy" />
956
+ <input>
957
+ <port id="0" precision="FP32">
958
+ <dim>1</dim>
959
+ <dim>128</dim>
960
+ <dim>32</dim>
961
+ <dim>32</dim>
962
+ </port>
963
+ <port id="1" precision="FP32">
964
+ <dim>1</dim>
965
+ <dim>128</dim>
966
+ <dim>1</dim>
967
+ <dim>1</dim>
968
+ </port>
969
+ </input>
970
+ <output>
971
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0">
972
+ <dim>1</dim>
973
+ <dim>128</dim>
974
+ <dim>32</dim>
975
+ <dim>32</dim>
976
+ </port>
977
+ </output>
978
+ </layer>
979
+ <layer id="53" name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" type="Add" version="opset1">
980
+ <data auto_broadcast="numpy" />
981
+ <input>
982
+ <port id="0" precision="FP32">
983
+ <dim>1</dim>
984
+ <dim>128</dim>
985
+ <dim>32</dim>
986
+ <dim>32</dim>
987
+ </port>
988
+ <port id="1" precision="FP32">
989
+ <dim>1</dim>
990
+ <dim>128</dim>
991
+ <dim>32</dim>
992
+ <dim>32</dim>
993
+ </port>
994
+ </input>
995
+ <output>
996
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0">
997
+ <dim>1</dim>
998
+ <dim>128</dim>
999
+ <dim>32</dim>
1000
+ <dim>32</dim>
1001
+ </port>
1002
+ </output>
1003
+ </layer>
1004
+ <layer id="54" name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" type="ReLU" version="opset1">
1005
+ <input>
1006
+ <port id="0" precision="FP32">
1007
+ <dim>1</dim>
1008
+ <dim>128</dim>
1009
+ <dim>32</dim>
1010
+ <dim>32</dim>
1011
+ </port>
1012
+ </input>
1013
+ <output>
1014
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0">
1015
+ <dim>1</dim>
1016
+ <dim>128</dim>
1017
+ <dim>32</dim>
1018
+ <dim>32</dim>
1019
+ </port>
1020
+ </output>
1021
+ </layer>
1022
+ <layer id="55" name="/Shape" type="ShapeOf" version="opset3">
1023
+ <data output_type="i64" />
1024
+ <input>
1025
+ <port id="0" precision="FP32">
1026
+ <dim>1</dim>
1027
+ <dim>128</dim>
1028
+ <dim>32</dim>
1029
+ <dim>32</dim>
1030
+ </port>
1031
+ </input>
1032
+ <output>
1033
+ <port id="1" precision="I64" names="/Shape_output_0">
1034
+ <dim>4</dim>
1035
+ </port>
1036
+ </output>
1037
+ </layer>
1038
+ <layer id="56" name="/Constant_1" type="Const" version="opset1">
1039
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1040
+ <output>
1041
+ <port id="0" precision="I64" names="/Constant_1_output_0">
1042
+ <dim>1</dim>
1043
+ </port>
1044
+ </output>
1045
+ </layer>
1046
+ <layer id="57" name="/Constant_2" type="Const" version="opset1">
1047
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1048
+ <output>
1049
+ <port id="0" precision="I64" names="/Constant_2_output_0">
1050
+ <dim>1</dim>
1051
+ </port>
1052
+ </output>
1053
+ </layer>
1054
+ <layer id="58" name="Broadcast_781" type="Const" version="opset1">
1055
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1056
+ <output>
1057
+ <port id="0" precision="I64">
1058
+ <dim>1</dim>
1059
+ </port>
1060
+ </output>
1061
+ </layer>
1062
+ <layer id="59" name="/Slice" type="StridedSlice" version="opset1">
1063
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1064
+ <input>
1065
+ <port id="0" precision="I64">
1066
+ <dim>4</dim>
1067
+ </port>
1068
+ <port id="1" precision="I64">
1069
+ <dim>1</dim>
1070
+ </port>
1071
+ <port id="2" precision="I64">
1072
+ <dim>1</dim>
1073
+ </port>
1074
+ <port id="3" precision="I64">
1075
+ <dim>1</dim>
1076
+ </port>
1077
+ </input>
1078
+ <output>
1079
+ <port id="4" precision="I64" names="/Slice_output_0">
1080
+ <dim>2</dim>
1081
+ </port>
1082
+ </output>
1083
+ </layer>
1084
+ <layer id="60" name="/Constant_3" type="Const" version="opset1">
1085
+ <data element_type="i64" shape="2" offset="2728472" size="16" />
1086
+ <output>
1087
+ <port id="0" precision="I64" names="/Constant_3_output_0">
1088
+ <dim>2</dim>
1089
+ </port>
1090
+ </output>
1091
+ </layer>
1092
+ <layer id="61" name="/Concat" type="Concat" version="opset1">
1093
+ <data axis="0" />
1094
+ <input>
1095
+ <port id="0" precision="I64">
1096
+ <dim>2</dim>
1097
+ </port>
1098
+ <port id="1" precision="I64">
1099
+ <dim>2</dim>
1100
+ </port>
1101
+ </input>
1102
+ <output>
1103
+ <port id="2" precision="I64" names="/Concat_output_0">
1104
+ <dim>4</dim>
1105
+ </port>
1106
+ </output>
1107
+ </layer>
1108
+ <layer id="62" name="Convert_804" type="Convert" version="opset1">
1109
+ <data destination_type="f32" />
1110
+ <input>
1111
+ <port id="0" precision="I64">
1112
+ <dim>4</dim>
1113
+ </port>
1114
+ </input>
1115
+ <output>
1116
+ <port id="1" precision="FP32">
1117
+ <dim>4</dim>
1118
+ </port>
1119
+ </output>
1120
+ </layer>
1121
+ <layer id="63" name="Convert_803" type="Convert" version="opset1">
1122
+ <data destination_type="f32" />
1123
+ <input>
1124
+ <port id="0" precision="I64">
1125
+ <dim>4</dim>
1126
+ </port>
1127
+ </input>
1128
+ <output>
1129
+ <port id="1" precision="FP32">
1130
+ <dim>4</dim>
1131
+ </port>
1132
+ </output>
1133
+ </layer>
1134
+ <layer id="64" name="Divide_805" type="Divide" version="opset1">
1135
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1136
+ <input>
1137
+ <port id="0" precision="FP32">
1138
+ <dim>4</dim>
1139
+ </port>
1140
+ <port id="1" precision="FP32">
1141
+ <dim>4</dim>
1142
+ </port>
1143
+ </input>
1144
+ <output>
1145
+ <port id="2" precision="FP32">
1146
+ <dim>4</dim>
1147
+ </port>
1148
+ </output>
1149
+ </layer>
1150
+ <layer id="65" name="Constant_6582" type="Const" version="opset1">
1151
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1152
+ <output>
1153
+ <port id="0" precision="FP32">
1154
+ <dim>1</dim>
1155
+ </port>
1156
+ </output>
1157
+ </layer>
1158
+ <layer id="66" name="Add_807" type="Add" version="opset1">
1159
+ <data auto_broadcast="numpy" />
1160
+ <input>
1161
+ <port id="0" precision="FP32">
1162
+ <dim>4</dim>
1163
+ </port>
1164
+ <port id="1" precision="FP32">
1165
+ <dim>1</dim>
1166
+ </port>
1167
+ </input>
1168
+ <output>
1169
+ <port id="2" precision="FP32">
1170
+ <dim>4</dim>
1171
+ </port>
1172
+ </output>
1173
+ </layer>
1174
+ <layer id="67" name="/Resize" type="Interpolate" version="opset4">
1175
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1176
+ <input>
1177
+ <port id="0" precision="FP32">
1178
+ <dim>1</dim>
1179
+ <dim>128</dim>
1180
+ <dim>32</dim>
1181
+ <dim>32</dim>
1182
+ </port>
1183
+ <port id="1" precision="I64">
1184
+ <dim>4</dim>
1185
+ </port>
1186
+ <port id="2" precision="FP32">
1187
+ <dim>4</dim>
1188
+ </port>
1189
+ </input>
1190
+ <output>
1191
+ <port id="3" precision="FP32" names="/Resize_output_0">
1192
+ <dim>1</dim>
1193
+ <dim>128</dim>
1194
+ <dim>64</dim>
1195
+ <dim>64</dim>
1196
+ </port>
1197
+ </output>
1198
+ </layer>
1199
+ <layer id="68" name="/Concat_1" type="Concat" version="opset1">
1200
+ <data axis="1" />
1201
+ <input>
1202
+ <port id="0" precision="FP32">
1203
+ <dim>1</dim>
1204
+ <dim>64</dim>
1205
+ <dim>64</dim>
1206
+ <dim>64</dim>
1207
+ </port>
1208
+ <port id="1" precision="FP32">
1209
+ <dim>1</dim>
1210
+ <dim>128</dim>
1211
+ <dim>64</dim>
1212
+ <dim>64</dim>
1213
+ </port>
1214
+ </input>
1215
+ <output>
1216
+ <port id="2" precision="FP32" names="/Concat_1_output_0">
1217
+ <dim>1</dim>
1218
+ <dim>192</dim>
1219
+ <dim>64</dim>
1220
+ <dim>64</dim>
1221
+ </port>
1222
+ </output>
1223
+ </layer>
1224
+ <layer id="69" name="onnx::Conv_290" type="Const" version="opset1">
1225
+ <data element_type="f32" shape="256, 128, 3, 3" offset="2728492" size="1179648" />
1226
+ <output>
1227
+ <port id="0" precision="FP32" names="onnx::Conv_290">
1228
+ <dim>256</dim>
1229
+ <dim>128</dim>
1230
+ <dim>3</dim>
1231
+ <dim>3</dim>
1232
+ </port>
1233
+ </output>
1234
+ </layer>
1235
+ <layer id="70" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1236
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1237
+ <input>
1238
+ <port id="0" precision="FP32">
1239
+ <dim>1</dim>
1240
+ <dim>128</dim>
1241
+ <dim>32</dim>
1242
+ <dim>32</dim>
1243
+ </port>
1244
+ <port id="1" precision="FP32">
1245
+ <dim>256</dim>
1246
+ <dim>128</dim>
1247
+ <dim>3</dim>
1248
+ <dim>3</dim>
1249
+ </port>
1250
+ </input>
1251
+ <output>
1252
+ <port id="2" precision="FP32">
1253
+ <dim>1</dim>
1254
+ <dim>256</dim>
1255
+ <dim>16</dim>
1256
+ <dim>16</dim>
1257
+ </port>
1258
+ </output>
1259
+ </layer>
1260
+ <layer id="71" name="Reshape_549" type="Const" version="opset1">
1261
+ <data element_type="f32" shape="1, 256, 1, 1" offset="3908140" size="1024" />
1262
+ <output>
1263
+ <port id="0" precision="FP32">
1264
+ <dim>1</dim>
1265
+ <dim>256</dim>
1266
+ <dim>1</dim>
1267
+ <dim>1</dim>
1268
+ </port>
1269
+ </output>
1270
+ </layer>
1271
+ <layer id="72" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" type="Add" version="opset1">
1272
+ <data auto_broadcast="numpy" />
1273
+ <input>
1274
+ <port id="0" precision="FP32">
1275
+ <dim>1</dim>
1276
+ <dim>256</dim>
1277
+ <dim>16</dim>
1278
+ <dim>16</dim>
1279
+ </port>
1280
+ <port id="1" precision="FP32">
1281
+ <dim>1</dim>
1282
+ <dim>256</dim>
1283
+ <dim>1</dim>
1284
+ <dim>1</dim>
1285
+ </port>
1286
+ </input>
1287
+ <output>
1288
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0">
1289
+ <dim>1</dim>
1290
+ <dim>256</dim>
1291
+ <dim>16</dim>
1292
+ <dim>16</dim>
1293
+ </port>
1294
+ </output>
1295
+ </layer>
1296
+ <layer id="73" name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" type="ReLU" version="opset1">
1297
+ <input>
1298
+ <port id="0" precision="FP32">
1299
+ <dim>1</dim>
1300
+ <dim>256</dim>
1301
+ <dim>16</dim>
1302
+ <dim>16</dim>
1303
+ </port>
1304
+ </input>
1305
+ <output>
1306
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0">
1307
+ <dim>1</dim>
1308
+ <dim>256</dim>
1309
+ <dim>16</dim>
1310
+ <dim>16</dim>
1311
+ </port>
1312
+ </output>
1313
+ </layer>
1314
+ <layer id="74" name="onnx::Conv_293" type="Const" version="opset1">
1315
+ <data element_type="f32" shape="256, 256, 3, 3" offset="3909164" size="2359296" />
1316
+ <output>
1317
+ <port id="0" precision="FP32" names="onnx::Conv_293">
1318
+ <dim>256</dim>
1319
+ <dim>256</dim>
1320
+ <dim>3</dim>
1321
+ <dim>3</dim>
1322
+ </port>
1323
+ </output>
1324
+ </layer>
1325
+ <layer id="75" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1326
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1327
+ <input>
1328
+ <port id="0" precision="FP32">
1329
+ <dim>1</dim>
1330
+ <dim>256</dim>
1331
+ <dim>16</dim>
1332
+ <dim>16</dim>
1333
+ </port>
1334
+ <port id="1" precision="FP32">
1335
+ <dim>256</dim>
1336
+ <dim>256</dim>
1337
+ <dim>3</dim>
1338
+ <dim>3</dim>
1339
+ </port>
1340
+ </input>
1341
+ <output>
1342
+ <port id="2" precision="FP32">
1343
+ <dim>1</dim>
1344
+ <dim>256</dim>
1345
+ <dim>16</dim>
1346
+ <dim>16</dim>
1347
+ </port>
1348
+ </output>
1349
+ </layer>
1350
+ <layer id="76" name="Reshape_598" type="Const" version="opset1">
1351
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6268460" size="1024" />
1352
+ <output>
1353
+ <port id="0" precision="FP32">
1354
+ <dim>1</dim>
1355
+ <dim>256</dim>
1356
+ <dim>1</dim>
1357
+ <dim>1</dim>
1358
+ </port>
1359
+ </output>
1360
+ </layer>
1361
+ <layer id="77" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" type="Add" version="opset1">
1362
+ <data auto_broadcast="numpy" />
1363
+ <input>
1364
+ <port id="0" precision="FP32">
1365
+ <dim>1</dim>
1366
+ <dim>256</dim>
1367
+ <dim>16</dim>
1368
+ <dim>16</dim>
1369
+ </port>
1370
+ <port id="1" precision="FP32">
1371
+ <dim>1</dim>
1372
+ <dim>256</dim>
1373
+ <dim>1</dim>
1374
+ <dim>1</dim>
1375
+ </port>
1376
+ </input>
1377
+ <output>
1378
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0">
1379
+ <dim>1</dim>
1380
+ <dim>256</dim>
1381
+ <dim>16</dim>
1382
+ <dim>16</dim>
1383
+ </port>
1384
+ </output>
1385
+ </layer>
1386
+ <layer id="78" name="onnx::Conv_296" type="Const" version="opset1">
1387
+ <data element_type="f32" shape="256, 128, 1, 1" offset="6269484" size="131072" />
1388
+ <output>
1389
+ <port id="0" precision="FP32" names="onnx::Conv_296">
1390
+ <dim>256</dim>
1391
+ <dim>128</dim>
1392
+ <dim>1</dim>
1393
+ <dim>1</dim>
1394
+ </port>
1395
+ </output>
1396
+ </layer>
1397
+ <layer id="79" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
1398
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
1399
+ <input>
1400
+ <port id="0" precision="FP32">
1401
+ <dim>1</dim>
1402
+ <dim>128</dim>
1403
+ <dim>32</dim>
1404
+ <dim>32</dim>
1405
+ </port>
1406
+ <port id="1" precision="FP32">
1407
+ <dim>256</dim>
1408
+ <dim>128</dim>
1409
+ <dim>1</dim>
1410
+ <dim>1</dim>
1411
+ </port>
1412
+ </input>
1413
+ <output>
1414
+ <port id="2" precision="FP32">
1415
+ <dim>1</dim>
1416
+ <dim>256</dim>
1417
+ <dim>16</dim>
1418
+ <dim>16</dim>
1419
+ </port>
1420
+ </output>
1421
+ </layer>
1422
+ <layer id="80" name="Reshape_646" type="Const" version="opset1">
1423
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6400556" size="1024" />
1424
+ <output>
1425
+ <port id="0" precision="FP32">
1426
+ <dim>1</dim>
1427
+ <dim>256</dim>
1428
+ <dim>1</dim>
1429
+ <dim>1</dim>
1430
+ </port>
1431
+ </output>
1432
+ </layer>
1433
+ <layer id="81" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" type="Add" version="opset1">
1434
+ <data auto_broadcast="numpy" />
1435
+ <input>
1436
+ <port id="0" precision="FP32">
1437
+ <dim>1</dim>
1438
+ <dim>256</dim>
1439
+ <dim>16</dim>
1440
+ <dim>16</dim>
1441
+ </port>
1442
+ <port id="1" precision="FP32">
1443
+ <dim>1</dim>
1444
+ <dim>256</dim>
1445
+ <dim>1</dim>
1446
+ <dim>1</dim>
1447
+ </port>
1448
+ </input>
1449
+ <output>
1450
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0">
1451
+ <dim>1</dim>
1452
+ <dim>256</dim>
1453
+ <dim>16</dim>
1454
+ <dim>16</dim>
1455
+ </port>
1456
+ </output>
1457
+ </layer>
1458
+ <layer id="82" name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" type="Add" version="opset1">
1459
+ <data auto_broadcast="numpy" />
1460
+ <input>
1461
+ <port id="0" precision="FP32">
1462
+ <dim>1</dim>
1463
+ <dim>256</dim>
1464
+ <dim>16</dim>
1465
+ <dim>16</dim>
1466
+ </port>
1467
+ <port id="1" precision="FP32">
1468
+ <dim>1</dim>
1469
+ <dim>256</dim>
1470
+ <dim>16</dim>
1471
+ <dim>16</dim>
1472
+ </port>
1473
+ </input>
1474
+ <output>
1475
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0">
1476
+ <dim>1</dim>
1477
+ <dim>256</dim>
1478
+ <dim>16</dim>
1479
+ <dim>16</dim>
1480
+ </port>
1481
+ </output>
1482
+ </layer>
1483
+ <layer id="83" name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" type="ReLU" version="opset1">
1484
+ <input>
1485
+ <port id="0" precision="FP32">
1486
+ <dim>1</dim>
1487
+ <dim>256</dim>
1488
+ <dim>16</dim>
1489
+ <dim>16</dim>
1490
+ </port>
1491
+ </input>
1492
+ <output>
1493
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0">
1494
+ <dim>1</dim>
1495
+ <dim>256</dim>
1496
+ <dim>16</dim>
1497
+ <dim>16</dim>
1498
+ </port>
1499
+ </output>
1500
+ </layer>
1501
+ <layer id="84" name="onnx::Conv_299" type="Const" version="opset1">
1502
+ <data element_type="f32" shape="256, 256, 3, 3" offset="6401580" size="2359296" />
1503
+ <output>
1504
+ <port id="0" precision="FP32" names="onnx::Conv_299">
1505
+ <dim>256</dim>
1506
+ <dim>256</dim>
1507
+ <dim>3</dim>
1508
+ <dim>3</dim>
1509
+ </port>
1510
+ </output>
1511
+ </layer>
1512
+ <layer id="85" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1513
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1514
+ <input>
1515
+ <port id="0" precision="FP32">
1516
+ <dim>1</dim>
1517
+ <dim>256</dim>
1518
+ <dim>16</dim>
1519
+ <dim>16</dim>
1520
+ </port>
1521
+ <port id="1" precision="FP32">
1522
+ <dim>256</dim>
1523
+ <dim>256</dim>
1524
+ <dim>3</dim>
1525
+ <dim>3</dim>
1526
+ </port>
1527
+ </input>
1528
+ <output>
1529
+ <port id="2" precision="FP32">
1530
+ <dim>1</dim>
1531
+ <dim>256</dim>
1532
+ <dim>16</dim>
1533
+ <dim>16</dim>
1534
+ </port>
1535
+ </output>
1536
+ </layer>
1537
+ <layer id="86" name="Reshape_696" type="Const" version="opset1">
1538
+ <data element_type="f32" shape="1, 256, 1, 1" offset="8760876" size="1024" />
1539
+ <output>
1540
+ <port id="0" precision="FP32">
1541
+ <dim>1</dim>
1542
+ <dim>256</dim>
1543
+ <dim>1</dim>
1544
+ <dim>1</dim>
1545
+ </port>
1546
+ </output>
1547
+ </layer>
1548
+ <layer id="87" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" type="Add" version="opset1">
1549
+ <data auto_broadcast="numpy" />
1550
+ <input>
1551
+ <port id="0" precision="FP32">
1552
+ <dim>1</dim>
1553
+ <dim>256</dim>
1554
+ <dim>16</dim>
1555
+ <dim>16</dim>
1556
+ </port>
1557
+ <port id="1" precision="FP32">
1558
+ <dim>1</dim>
1559
+ <dim>256</dim>
1560
+ <dim>1</dim>
1561
+ <dim>1</dim>
1562
+ </port>
1563
+ </input>
1564
+ <output>
1565
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0">
1566
+ <dim>1</dim>
1567
+ <dim>256</dim>
1568
+ <dim>16</dim>
1569
+ <dim>16</dim>
1570
+ </port>
1571
+ </output>
1572
+ </layer>
1573
+ <layer id="88" name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" type="ReLU" version="opset1">
1574
+ <input>
1575
+ <port id="0" precision="FP32">
1576
+ <dim>1</dim>
1577
+ <dim>256</dim>
1578
+ <dim>16</dim>
1579
+ <dim>16</dim>
1580
+ </port>
1581
+ </input>
1582
+ <output>
1583
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0">
1584
+ <dim>1</dim>
1585
+ <dim>256</dim>
1586
+ <dim>16</dim>
1587
+ <dim>16</dim>
1588
+ </port>
1589
+ </output>
1590
+ </layer>
1591
+ <layer id="89" name="onnx::Conv_302" type="Const" version="opset1">
1592
+ <data element_type="f32" shape="256, 256, 3, 3" offset="8761900" size="2359296" />
1593
+ <output>
1594
+ <port id="0" precision="FP32" names="onnx::Conv_302">
1595
+ <dim>256</dim>
1596
+ <dim>256</dim>
1597
+ <dim>3</dim>
1598
+ <dim>3</dim>
1599
+ </port>
1600
+ </output>
1601
+ </layer>
1602
+ <layer id="90" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1603
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1604
+ <input>
1605
+ <port id="0" precision="FP32">
1606
+ <dim>1</dim>
1607
+ <dim>256</dim>
1608
+ <dim>16</dim>
1609
+ <dim>16</dim>
1610
+ </port>
1611
+ <port id="1" precision="FP32">
1612
+ <dim>256</dim>
1613
+ <dim>256</dim>
1614
+ <dim>3</dim>
1615
+ <dim>3</dim>
1616
+ </port>
1617
+ </input>
1618
+ <output>
1619
+ <port id="2" precision="FP32">
1620
+ <dim>1</dim>
1621
+ <dim>256</dim>
1622
+ <dim>16</dim>
1623
+ <dim>16</dim>
1624
+ </port>
1625
+ </output>
1626
+ </layer>
1627
+ <layer id="91" name="Reshape_745" type="Const" version="opset1">
1628
+ <data element_type="f32" shape="1, 256, 1, 1" offset="11121196" size="1024" />
1629
+ <output>
1630
+ <port id="0" precision="FP32">
1631
+ <dim>1</dim>
1632
+ <dim>256</dim>
1633
+ <dim>1</dim>
1634
+ <dim>1</dim>
1635
+ </port>
1636
+ </output>
1637
+ </layer>
1638
+ <layer id="92" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" type="Add" version="opset1">
1639
+ <data auto_broadcast="numpy" />
1640
+ <input>
1641
+ <port id="0" precision="FP32">
1642
+ <dim>1</dim>
1643
+ <dim>256</dim>
1644
+ <dim>16</dim>
1645
+ <dim>16</dim>
1646
+ </port>
1647
+ <port id="1" precision="FP32">
1648
+ <dim>1</dim>
1649
+ <dim>256</dim>
1650
+ <dim>1</dim>
1651
+ <dim>1</dim>
1652
+ </port>
1653
+ </input>
1654
+ <output>
1655
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0">
1656
+ <dim>1</dim>
1657
+ <dim>256</dim>
1658
+ <dim>16</dim>
1659
+ <dim>16</dim>
1660
+ </port>
1661
+ </output>
1662
+ </layer>
1663
+ <layer id="93" name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" type="Add" version="opset1">
1664
+ <data auto_broadcast="numpy" />
1665
+ <input>
1666
+ <port id="0" precision="FP32">
1667
+ <dim>1</dim>
1668
+ <dim>256</dim>
1669
+ <dim>16</dim>
1670
+ <dim>16</dim>
1671
+ </port>
1672
+ <port id="1" precision="FP32">
1673
+ <dim>1</dim>
1674
+ <dim>256</dim>
1675
+ <dim>16</dim>
1676
+ <dim>16</dim>
1677
+ </port>
1678
+ </input>
1679
+ <output>
1680
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0">
1681
+ <dim>1</dim>
1682
+ <dim>256</dim>
1683
+ <dim>16</dim>
1684
+ <dim>16</dim>
1685
+ </port>
1686
+ </output>
1687
+ </layer>
1688
+ <layer id="94" name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" type="ReLU" version="opset1">
1689
+ <input>
1690
+ <port id="0" precision="FP32">
1691
+ <dim>1</dim>
1692
+ <dim>256</dim>
1693
+ <dim>16</dim>
1694
+ <dim>16</dim>
1695
+ </port>
1696
+ </input>
1697
+ <output>
1698
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0">
1699
+ <dim>1</dim>
1700
+ <dim>256</dim>
1701
+ <dim>16</dim>
1702
+ <dim>16</dim>
1703
+ </port>
1704
+ </output>
1705
+ </layer>
1706
+ <layer id="95" name="/Shape_3" type="ShapeOf" version="opset3">
1707
+ <data output_type="i64" />
1708
+ <input>
1709
+ <port id="0" precision="FP32">
1710
+ <dim>1</dim>
1711
+ <dim>256</dim>
1712
+ <dim>16</dim>
1713
+ <dim>16</dim>
1714
+ </port>
1715
+ </input>
1716
+ <output>
1717
+ <port id="1" precision="I64" names="/Shape_3_output_0">
1718
+ <dim>4</dim>
1719
+ </port>
1720
+ </output>
1721
+ </layer>
1722
+ <layer id="96" name="/Constant_9" type="Const" version="opset1">
1723
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1724
+ <output>
1725
+ <port id="0" precision="I64" names="/Constant_9_output_0">
1726
+ <dim>1</dim>
1727
+ </port>
1728
+ </output>
1729
+ </layer>
1730
+ <layer id="97" name="/Constant_10" type="Const" version="opset1">
1731
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1732
+ <output>
1733
+ <port id="0" precision="I64" names="/Constant_10_output_0">
1734
+ <dim>1</dim>
1735
+ </port>
1736
+ </output>
1737
+ </layer>
1738
+ <layer id="98" name="Broadcast_861" type="Const" version="opset1">
1739
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1740
+ <output>
1741
+ <port id="0" precision="I64">
1742
+ <dim>1</dim>
1743
+ </port>
1744
+ </output>
1745
+ </layer>
1746
+ <layer id="99" name="/Slice_1" type="StridedSlice" version="opset1">
1747
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1748
+ <input>
1749
+ <port id="0" precision="I64">
1750
+ <dim>4</dim>
1751
+ </port>
1752
+ <port id="1" precision="I64">
1753
+ <dim>1</dim>
1754
+ </port>
1755
+ <port id="2" precision="I64">
1756
+ <dim>1</dim>
1757
+ </port>
1758
+ <port id="3" precision="I64">
1759
+ <dim>1</dim>
1760
+ </port>
1761
+ </input>
1762
+ <output>
1763
+ <port id="4" precision="I64" names="/Slice_1_output_0">
1764
+ <dim>2</dim>
1765
+ </port>
1766
+ </output>
1767
+ </layer>
1768
+ <layer id="100" name="/Shape_1" type="ShapeOf" version="opset3">
1769
+ <data output_type="i64" />
1770
+ <input>
1771
+ <port id="0" precision="FP32">
1772
+ <dim>1</dim>
1773
+ <dim>192</dim>
1774
+ <dim>64</dim>
1775
+ <dim>64</dim>
1776
+ </port>
1777
+ </input>
1778
+ <output>
1779
+ <port id="1" precision="I64" names="/Shape_1_output_0,/Shape_2_output_0">
1780
+ <dim>4</dim>
1781
+ </port>
1782
+ </output>
1783
+ </layer>
1784
+ <layer id="101" name="Constant_6090" type="Const" version="opset1">
1785
+ <data element_type="i64" shape="2" offset="11122220" size="16" />
1786
+ <output>
1787
+ <port id="0" precision="I64">
1788
+ <dim>2</dim>
1789
+ </port>
1790
+ </output>
1791
+ </layer>
1792
+ <layer id="102" name="Constant_6091" type="Const" version="opset1">
1793
+ <data element_type="i64" shape="" offset="2728448" size="8" />
1794
+ <output>
1795
+ <port id="0" precision="I64" />
1796
+ </output>
1797
+ </layer>
1798
+ <layer id="103" name="Gather_6092" type="Gather" version="opset8">
1799
+ <data batch_dims="0" />
1800
+ <input>
1801
+ <port id="0" precision="I64">
1802
+ <dim>4</dim>
1803
+ </port>
1804
+ <port id="1" precision="I64">
1805
+ <dim>2</dim>
1806
+ </port>
1807
+ <port id="2" precision="I64" />
1808
+ </input>
1809
+ <output>
1810
+ <port id="3" precision="I64" names="/Cast_output_0,/Concat_2_output_0">
1811
+ <dim>2</dim>
1812
+ </port>
1813
+ </output>
1814
+ </layer>
1815
+ <layer id="104" name="/Concat_3" type="Concat" version="opset1">
1816
+ <data axis="0" />
1817
+ <input>
1818
+ <port id="0" precision="I64">
1819
+ <dim>2</dim>
1820
+ </port>
1821
+ <port id="1" precision="I64">
1822
+ <dim>2</dim>
1823
+ </port>
1824
+ </input>
1825
+ <output>
1826
+ <port id="2" precision="I64" names="/Concat_3_output_0">
1827
+ <dim>4</dim>
1828
+ </port>
1829
+ </output>
1830
+ </layer>
1831
+ <layer id="105" name="Convert_884" type="Convert" version="opset1">
1832
+ <data destination_type="f32" />
1833
+ <input>
1834
+ <port id="0" precision="I64">
1835
+ <dim>4</dim>
1836
+ </port>
1837
+ </input>
1838
+ <output>
1839
+ <port id="1" precision="FP32">
1840
+ <dim>4</dim>
1841
+ </port>
1842
+ </output>
1843
+ </layer>
1844
+ <layer id="106" name="Convert_883" type="Convert" version="opset1">
1845
+ <data destination_type="f32" />
1846
+ <input>
1847
+ <port id="0" precision="I64">
1848
+ <dim>4</dim>
1849
+ </port>
1850
+ </input>
1851
+ <output>
1852
+ <port id="1" precision="FP32">
1853
+ <dim>4</dim>
1854
+ </port>
1855
+ </output>
1856
+ </layer>
1857
+ <layer id="107" name="Divide_885" type="Divide" version="opset1">
1858
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1859
+ <input>
1860
+ <port id="0" precision="FP32">
1861
+ <dim>4</dim>
1862
+ </port>
1863
+ <port id="1" precision="FP32">
1864
+ <dim>4</dim>
1865
+ </port>
1866
+ </input>
1867
+ <output>
1868
+ <port id="2" precision="FP32">
1869
+ <dim>4</dim>
1870
+ </port>
1871
+ </output>
1872
+ </layer>
1873
+ <layer id="108" name="Constant_6583" type="Const" version="opset1">
1874
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1875
+ <output>
1876
+ <port id="0" precision="FP32">
1877
+ <dim>1</dim>
1878
+ </port>
1879
+ </output>
1880
+ </layer>
1881
+ <layer id="109" name="Add_887" type="Add" version="opset1">
1882
+ <data auto_broadcast="numpy" />
1883
+ <input>
1884
+ <port id="0" precision="FP32">
1885
+ <dim>4</dim>
1886
+ </port>
1887
+ <port id="1" precision="FP32">
1888
+ <dim>1</dim>
1889
+ </port>
1890
+ </input>
1891
+ <output>
1892
+ <port id="2" precision="FP32">
1893
+ <dim>4</dim>
1894
+ </port>
1895
+ </output>
1896
+ </layer>
1897
+ <layer id="110" name="/Resize_1" type="Interpolate" version="opset4">
1898
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1899
+ <input>
1900
+ <port id="0" precision="FP32">
1901
+ <dim>1</dim>
1902
+ <dim>256</dim>
1903
+ <dim>16</dim>
1904
+ <dim>16</dim>
1905
+ </port>
1906
+ <port id="1" precision="I64">
1907
+ <dim>4</dim>
1908
+ </port>
1909
+ <port id="2" precision="FP32">
1910
+ <dim>4</dim>
1911
+ </port>
1912
+ </input>
1913
+ <output>
1914
+ <port id="3" precision="FP32" names="/Resize_1_output_0">
1915
+ <dim>1</dim>
1916
+ <dim>256</dim>
1917
+ <dim>64</dim>
1918
+ <dim>64</dim>
1919
+ </port>
1920
+ </output>
1921
+ </layer>
1922
+ <layer id="111" name="/Concat_4" type="Concat" version="opset1">
1923
+ <data axis="1" />
1924
+ <input>
1925
+ <port id="0" precision="FP32">
1926
+ <dim>1</dim>
1927
+ <dim>192</dim>
1928
+ <dim>64</dim>
1929
+ <dim>64</dim>
1930
+ </port>
1931
+ <port id="1" precision="FP32">
1932
+ <dim>1</dim>
1933
+ <dim>256</dim>
1934
+ <dim>64</dim>
1935
+ <dim>64</dim>
1936
+ </port>
1937
+ </input>
1938
+ <output>
1939
+ <port id="2" precision="FP32" names="/Concat_4_output_0">
1940
+ <dim>1</dim>
1941
+ <dim>448</dim>
1942
+ <dim>64</dim>
1943
+ <dim>64</dim>
1944
+ </port>
1945
+ </output>
1946
+ </layer>
1947
+ <layer id="112" name="onnx::Gather_308" type="Const" version="opset1">
1948
+ <data element_type="i64" shape="100" offset="11122236" size="800" />
1949
+ <output>
1950
+ <port id="0" precision="I64" names="onnx::Gather_308">
1951
+ <dim>100</dim>
1952
+ </port>
1953
+ </output>
1954
+ </layer>
1955
+ <layer id="113" name="Constant_965" type="Const" version="opset1">
1956
+ <data element_type="i64" shape="" offset="2728464" size="8" />
1957
+ <output>
1958
+ <port id="0" precision="I64" />
1959
+ </output>
1960
+ </layer>
1961
+ <layer id="114" name="/Gather_2" type="Gather" version="opset8">
1962
+ <data batch_dims="0" />
1963
+ <input>
1964
+ <port id="0" precision="FP32">
1965
+ <dim>1</dim>
1966
+ <dim>448</dim>
1967
+ <dim>64</dim>
1968
+ <dim>64</dim>
1969
+ </port>
1970
+ <port id="1" precision="I64">
1971
+ <dim>100</dim>
1972
+ </port>
1973
+ <port id="2" precision="I64" />
1974
+ </input>
1975
+ <output>
1976
+ <port id="3" precision="FP32" names="/Gather_2_output_0">
1977
+ <dim>1</dim>
1978
+ <dim>100</dim>
1979
+ <dim>64</dim>
1980
+ <dim>64</dim>
1981
+ </port>
1982
+ </output>
1983
+ </layer>
1984
+ <layer id="115" name="/anomaly_map_generator/Constant" type="Const" version="opset1">
1985
+ <data element_type="i64" shape="3" offset="11123036" size="24" />
1986
+ <output>
1987
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_output_0">
1988
+ <dim>3</dim>
1989
+ </port>
1990
+ </output>
1991
+ </layer>
1992
+ <layer id="116" name="/anomaly_map_generator/Reshape" type="Reshape" version="opset1">
1993
+ <data special_zero="true" />
1994
+ <input>
1995
+ <port id="0" precision="FP32">
1996
+ <dim>1</dim>
1997
+ <dim>100</dim>
1998
+ <dim>64</dim>
1999
+ <dim>64</dim>
2000
+ </port>
2001
+ <port id="1" precision="I64">
2002
+ <dim>3</dim>
2003
+ </port>
2004
+ </input>
2005
+ <output>
2006
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_output_0">
2007
+ <dim>1</dim>
2008
+ <dim>100</dim>
2009
+ <dim>4096</dim>
2010
+ </port>
2011
+ </output>
2012
+ </layer>
2013
+ <layer id="117" name="Constant_6584" type="Const" version="opset1">
2014
+ <data element_type="f32" shape="1, 100, 4096" offset="11123060" size="1638400" />
2015
+ <output>
2016
+ <port id="0" precision="FP32">
2017
+ <dim>1</dim>
2018
+ <dim>100</dim>
2019
+ <dim>4096</dim>
2020
+ </port>
2021
+ </output>
2022
+ </layer>
2023
+ <layer id="118" name="/anomaly_map_generator/Sub" type="Subtract" version="opset1">
2024
+ <data auto_broadcast="numpy" />
2025
+ <input>
2026
+ <port id="0" precision="FP32">
2027
+ <dim>1</dim>
2028
+ <dim>100</dim>
2029
+ <dim>4096</dim>
2030
+ </port>
2031
+ <port id="1" precision="FP32">
2032
+ <dim>1</dim>
2033
+ <dim>100</dim>
2034
+ <dim>4096</dim>
2035
+ </port>
2036
+ </input>
2037
+ <output>
2038
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Sub_output_0">
2039
+ <dim>1</dim>
2040
+ <dim>100</dim>
2041
+ <dim>4096</dim>
2042
+ </port>
2043
+ </output>
2044
+ </layer>
2045
+ <layer id="119" name="Constant_975" type="Const" version="opset1">
2046
+ <data element_type="i64" shape="3" offset="12761460" size="24" />
2047
+ <output>
2048
+ <port id="0" precision="I64">
2049
+ <dim>3</dim>
2050
+ </port>
2051
+ </output>
2052
+ </layer>
2053
+ <layer id="120" name="/anomaly_map_generator/Transpose" type="Transpose" version="opset1">
2054
+ <input>
2055
+ <port id="0" precision="FP32">
2056
+ <dim>1</dim>
2057
+ <dim>100</dim>
2058
+ <dim>4096</dim>
2059
+ </port>
2060
+ <port id="1" precision="I64">
2061
+ <dim>3</dim>
2062
+ </port>
2063
+ </input>
2064
+ <output>
2065
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Transpose_output_0">
2066
+ <dim>4096</dim>
2067
+ <dim>1</dim>
2068
+ <dim>100</dim>
2069
+ </port>
2070
+ </output>
2071
+ </layer>
2072
+ <layer id="121" name="onnx::MatMul_310" type="Const" version="opset1">
2073
+ <data element_type="f32" shape="4096, 100, 100" offset="12761484" size="163840000" />
2074
+ <output>
2075
+ <port id="0" precision="FP32" names="onnx::MatMul_310">
2076
+ <dim>4096</dim>
2077
+ <dim>100</dim>
2078
+ <dim>100</dim>
2079
+ </port>
2080
+ </output>
2081
+ </layer>
2082
+ <layer id="122" name="/anomaly_map_generator/MatMul" type="MatMul" version="opset1">
2083
+ <data transpose_a="false" transpose_b="false" />
2084
+ <input>
2085
+ <port id="0" precision="FP32">
2086
+ <dim>4096</dim>
2087
+ <dim>1</dim>
2088
+ <dim>100</dim>
2089
+ </port>
2090
+ <port id="1" precision="FP32">
2091
+ <dim>4096</dim>
2092
+ <dim>100</dim>
2093
+ <dim>100</dim>
2094
+ </port>
2095
+ </input>
2096
+ <output>
2097
+ <port id="2" precision="FP32" names="/anomaly_map_generator/MatMul_output_0">
2098
+ <dim>4096</dim>
2099
+ <dim>1</dim>
2100
+ <dim>100</dim>
2101
+ </port>
2102
+ </output>
2103
+ </layer>
2104
+ <layer id="123" name="/anomaly_map_generator/Mul" type="Multiply" version="opset1">
2105
+ <data auto_broadcast="numpy" />
2106
+ <input>
2107
+ <port id="0" precision="FP32">
2108
+ <dim>4096</dim>
2109
+ <dim>1</dim>
2110
+ <dim>100</dim>
2111
+ </port>
2112
+ <port id="1" precision="FP32">
2113
+ <dim>4096</dim>
2114
+ <dim>1</dim>
2115
+ <dim>100</dim>
2116
+ </port>
2117
+ </input>
2118
+ <output>
2119
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Mul_output_0">
2120
+ <dim>4096</dim>
2121
+ <dim>1</dim>
2122
+ <dim>100</dim>
2123
+ </port>
2124
+ </output>
2125
+ </layer>
2126
+ <layer id="124" name="Constant_979" type="Const" version="opset1">
2127
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2128
+ <output>
2129
+ <port id="0" precision="I64">
2130
+ <dim>1</dim>
2131
+ </port>
2132
+ </output>
2133
+ </layer>
2134
+ <layer id="125" name="/anomaly_map_generator/ReduceSum" type="ReduceSum" version="opset1">
2135
+ <data keep_dims="false" />
2136
+ <input>
2137
+ <port id="0" precision="FP32">
2138
+ <dim>4096</dim>
2139
+ <dim>1</dim>
2140
+ <dim>100</dim>
2141
+ </port>
2142
+ <port id="1" precision="I64">
2143
+ <dim>1</dim>
2144
+ </port>
2145
+ </input>
2146
+ <output>
2147
+ <port id="2" precision="FP32" names="/anomaly_map_generator/ReduceSum_output_0">
2148
+ <dim>4096</dim>
2149
+ <dim>1</dim>
2150
+ </port>
2151
+ </output>
2152
+ </layer>
2153
+ <layer id="126" name="/anomaly_map_generator/Constant_1" type="Const" version="opset1">
2154
+ <data element_type="i64" shape="4" offset="176601484" size="32" />
2155
+ <output>
2156
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_1_output_0">
2157
+ <dim>4</dim>
2158
+ </port>
2159
+ </output>
2160
+ </layer>
2161
+ <layer id="127" name="/anomaly_map_generator/Reshape_1" type="Reshape" version="opset1">
2162
+ <data special_zero="true" />
2163
+ <input>
2164
+ <port id="0" precision="FP32">
2165
+ <dim>4096</dim>
2166
+ <dim>1</dim>
2167
+ </port>
2168
+ <port id="1" precision="I64">
2169
+ <dim>4</dim>
2170
+ </port>
2171
+ </input>
2172
+ <output>
2173
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_1_output_0">
2174
+ <dim>1</dim>
2175
+ <dim>1</dim>
2176
+ <dim>64</dim>
2177
+ <dim>64</dim>
2178
+ </port>
2179
+ </output>
2180
+ </layer>
2181
+ <layer id="128" name="/anomaly_map_generator/Clip" type="Clamp" version="opset1">
2182
+ <data min="0" max="3.4028234663852886e+38" />
2183
+ <input>
2184
+ <port id="0" precision="FP32">
2185
+ <dim>1</dim>
2186
+ <dim>1</dim>
2187
+ <dim>64</dim>
2188
+ <dim>64</dim>
2189
+ </port>
2190
+ </input>
2191
+ <output>
2192
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Clip_output_0">
2193
+ <dim>1</dim>
2194
+ <dim>1</dim>
2195
+ <dim>64</dim>
2196
+ <dim>64</dim>
2197
+ </port>
2198
+ </output>
2199
+ </layer>
2200
+ <layer id="129" name="/anomaly_map_generator/Sqrt" type="Sqrt" version="opset1">
2201
+ <input>
2202
+ <port id="0" precision="FP32">
2203
+ <dim>1</dim>
2204
+ <dim>1</dim>
2205
+ <dim>64</dim>
2206
+ <dim>64</dim>
2207
+ </port>
2208
+ </input>
2209
+ <output>
2210
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Sqrt_output_0">
2211
+ <dim>1</dim>
2212
+ <dim>1</dim>
2213
+ <dim>64</dim>
2214
+ <dim>64</dim>
2215
+ </port>
2216
+ </output>
2217
+ </layer>
2218
+ <layer id="130" name="/anomaly_map_generator/Shape" type="ShapeOf" version="opset3">
2219
+ <data output_type="i64" />
2220
+ <input>
2221
+ <port id="0" precision="FP32">
2222
+ <dim>1</dim>
2223
+ <dim>1</dim>
2224
+ <dim>64</dim>
2225
+ <dim>64</dim>
2226
+ </port>
2227
+ </input>
2228
+ <output>
2229
+ <port id="1" precision="I64" names="/anomaly_map_generator/Shape_output_0">
2230
+ <dim>4</dim>
2231
+ </port>
2232
+ </output>
2233
+ </layer>
2234
+ <layer id="131" name="/anomaly_map_generator/Constant_4" type="Const" version="opset1">
2235
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
2236
+ <output>
2237
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_4_output_0">
2238
+ <dim>1</dim>
2239
+ </port>
2240
+ </output>
2241
+ </layer>
2242
+ <layer id="132" name="/anomaly_map_generator/Constant_5" type="Const" version="opset1">
2243
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2244
+ <output>
2245
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_5_output_0">
2246
+ <dim>1</dim>
2247
+ </port>
2248
+ </output>
2249
+ </layer>
2250
+ <layer id="133" name="Broadcast_1002" type="Const" version="opset1">
2251
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
2252
+ <output>
2253
+ <port id="0" precision="I64">
2254
+ <dim>1</dim>
2255
+ </port>
2256
+ </output>
2257
+ </layer>
2258
+ <layer id="134" name="/anomaly_map_generator/Slice" type="StridedSlice" version="opset1">
2259
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
2260
+ <input>
2261
+ <port id="0" precision="I64">
2262
+ <dim>4</dim>
2263
+ </port>
2264
+ <port id="1" precision="I64">
2265
+ <dim>1</dim>
2266
+ </port>
2267
+ <port id="2" precision="I64">
2268
+ <dim>1</dim>
2269
+ </port>
2270
+ <port id="3" precision="I64">
2271
+ <dim>1</dim>
2272
+ </port>
2273
+ </input>
2274
+ <output>
2275
+ <port id="4" precision="I64" names="/anomaly_map_generator/Slice_output_0">
2276
+ <dim>2</dim>
2277
+ </port>
2278
+ </output>
2279
+ </layer>
2280
+ <layer id="135" name="/anomaly_map_generator/Constant_6" type="Const" version="opset1">
2281
+ <data element_type="i64" shape="2" offset="176601516" size="16" />
2282
+ <output>
2283
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_6_output_0">
2284
+ <dim>2</dim>
2285
+ </port>
2286
+ </output>
2287
+ </layer>
2288
+ <layer id="136" name="/anomaly_map_generator/Concat" type="Concat" version="opset1">
2289
+ <data axis="0" />
2290
+ <input>
2291
+ <port id="0" precision="I64">
2292
+ <dim>2</dim>
2293
+ </port>
2294
+ <port id="1" precision="I64">
2295
+ <dim>2</dim>
2296
+ </port>
2297
+ </input>
2298
+ <output>
2299
+ <port id="2" precision="I64" names="/anomaly_map_generator/Concat_output_0">
2300
+ <dim>4</dim>
2301
+ </port>
2302
+ </output>
2303
+ </layer>
2304
+ <layer id="137" name="Convert_1025" type="Convert" version="opset1">
2305
+ <data destination_type="f32" />
2306
+ <input>
2307
+ <port id="0" precision="I64">
2308
+ <dim>4</dim>
2309
+ </port>
2310
+ </input>
2311
+ <output>
2312
+ <port id="1" precision="FP32">
2313
+ <dim>4</dim>
2314
+ </port>
2315
+ </output>
2316
+ </layer>
2317
+ <layer id="138" name="Convert_1024" type="Convert" version="opset1">
2318
+ <data destination_type="f32" />
2319
+ <input>
2320
+ <port id="0" precision="I64">
2321
+ <dim>4</dim>
2322
+ </port>
2323
+ </input>
2324
+ <output>
2325
+ <port id="1" precision="FP32">
2326
+ <dim>4</dim>
2327
+ </port>
2328
+ </output>
2329
+ </layer>
2330
+ <layer id="139" name="Divide_1026" type="Divide" version="opset1">
2331
+ <data auto_broadcast="numpy" m_pythondiv="true" />
2332
+ <input>
2333
+ <port id="0" precision="FP32">
2334
+ <dim>4</dim>
2335
+ </port>
2336
+ <port id="1" precision="FP32">
2337
+ <dim>4</dim>
2338
+ </port>
2339
+ </input>
2340
+ <output>
2341
+ <port id="2" precision="FP32">
2342
+ <dim>4</dim>
2343
+ </port>
2344
+ </output>
2345
+ </layer>
2346
+ <layer id="140" name="Constant_6585" type="Const" version="opset1">
2347
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
2348
+ <output>
2349
+ <port id="0" precision="FP32">
2350
+ <dim>1</dim>
2351
+ </port>
2352
+ </output>
2353
+ </layer>
2354
+ <layer id="141" name="Add_1028" type="Add" version="opset1">
2355
+ <data auto_broadcast="numpy" />
2356
+ <input>
2357
+ <port id="0" precision="FP32">
2358
+ <dim>4</dim>
2359
+ </port>
2360
+ <port id="1" precision="FP32">
2361
+ <dim>1</dim>
2362
+ </port>
2363
+ </input>
2364
+ <output>
2365
+ <port id="2" precision="FP32">
2366
+ <dim>4</dim>
2367
+ </port>
2368
+ </output>
2369
+ </layer>
2370
+ <layer id="142" name="/anomaly_map_generator/Resize" type="Interpolate" version="opset4">
2371
+ <data mode="linear_onnx" shape_calculation_mode="sizes" coordinate_transformation_mode="half_pixel" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
2372
+ <input>
2373
+ <port id="0" precision="FP32">
2374
+ <dim>1</dim>
2375
+ <dim>1</dim>
2376
+ <dim>64</dim>
2377
+ <dim>64</dim>
2378
+ </port>
2379
+ <port id="1" precision="I64">
2380
+ <dim>4</dim>
2381
+ </port>
2382
+ <port id="2" precision="FP32">
2383
+ <dim>4</dim>
2384
+ </port>
2385
+ </input>
2386
+ <output>
2387
+ <port id="3" precision="FP32" names="/anomaly_map_generator/Resize_output_0">
2388
+ <dim>1</dim>
2389
+ <dim>1</dim>
2390
+ <dim>256</dim>
2391
+ <dim>256</dim>
2392
+ </port>
2393
+ </output>
2394
+ </layer>
2395
+ <layer id="143" name="Split_1097.0" type="Const" version="opset1">
2396
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2397
+ <output>
2398
+ <port id="0" precision="I64">
2399
+ <dim>4</dim>
2400
+ </port>
2401
+ </output>
2402
+ </layer>
2403
+ <layer id="144" name="Split_1097.1" type="Const" version="opset1">
2404
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2405
+ <output>
2406
+ <port id="0" precision="I64">
2407
+ <dim>4</dim>
2408
+ </port>
2409
+ </output>
2410
+ </layer>
2411
+ <layer id="145" name="Constant_1095" type="Const" version="opset1">
2412
+ <data element_type="f32" shape="" offset="176601564" size="4" />
2413
+ <output>
2414
+ <port id="0" precision="FP32" />
2415
+ </output>
2416
+ </layer>
2417
+ <layer id="146" name="/anomaly_map_generator/blur/Pad" type="Pad" version="opset1">
2418
+ <data pad_mode="reflect" />
2419
+ <input>
2420
+ <port id="0" precision="FP32">
2421
+ <dim>1</dim>
2422
+ <dim>1</dim>
2423
+ <dim>256</dim>
2424
+ <dim>256</dim>
2425
+ </port>
2426
+ <port id="1" precision="I64">
2427
+ <dim>4</dim>
2428
+ </port>
2429
+ <port id="2" precision="I64">
2430
+ <dim>4</dim>
2431
+ </port>
2432
+ <port id="3" precision="FP32" />
2433
+ </input>
2434
+ <output>
2435
+ <port id="4" precision="FP32" names="/anomaly_map_generator/blur/Pad_output_0">
2436
+ <dim>1</dim>
2437
+ <dim>1</dim>
2438
+ <dim>288</dim>
2439
+ <dim>288</dim>
2440
+ </port>
2441
+ </output>
2442
+ </layer>
2443
+ <layer id="147" name="anomaly_map_generator.blur.kernel" type="Const" version="opset1">
2444
+ <data element_type="f32" shape="1, 1, 33, 33" offset="176601568" size="4356" />
2445
+ <output>
2446
+ <port id="0" precision="FP32" names="anomaly_map_generator.blur.kernel">
2447
+ <dim>1</dim>
2448
+ <dim>1</dim>
2449
+ <dim>33</dim>
2450
+ <dim>33</dim>
2451
+ </port>
2452
+ </output>
2453
+ </layer>
2454
+ <layer id="148" name="/anomaly_map_generator/blur/Conv" type="Convolution" version="opset1">
2455
+ <data strides="1, 1" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
2456
+ <input>
2457
+ <port id="0" precision="FP32">
2458
+ <dim>1</dim>
2459
+ <dim>1</dim>
2460
+ <dim>288</dim>
2461
+ <dim>288</dim>
2462
+ </port>
2463
+ <port id="1" precision="FP32">
2464
+ <dim>1</dim>
2465
+ <dim>1</dim>
2466
+ <dim>33</dim>
2467
+ <dim>33</dim>
2468
+ </port>
2469
+ </input>
2470
+ <output>
2471
+ <port id="2" precision="FP32" names="/anomaly_map_generator/blur/Conv_output_0">
2472
+ <dim>1</dim>
2473
+ <dim>1</dim>
2474
+ <dim>256</dim>
2475
+ <dim>256</dim>
2476
+ </port>
2477
+ </output>
2478
+ </layer>
2479
+ <layer id="149" name="/anomaly_map_generator/blur/Constant_8" type="Const" version="opset1">
2480
+ <data element_type="i64" shape="4" offset="176605924" size="32" />
2481
+ <output>
2482
+ <port id="0" precision="I64" names="/anomaly_map_generator/blur/Constant_8_output_0">
2483
+ <dim>4</dim>
2484
+ </port>
2485
+ </output>
2486
+ </layer>
2487
+ <layer id="150" name="output" type="Reshape" version="opset1">
2488
+ <data special_zero="true" />
2489
+ <input>
2490
+ <port id="0" precision="FP32">
2491
+ <dim>1</dim>
2492
+ <dim>1</dim>
2493
+ <dim>256</dim>
2494
+ <dim>256</dim>
2495
+ </port>
2496
+ <port id="1" precision="I64">
2497
+ <dim>4</dim>
2498
+ </port>
2499
+ </input>
2500
+ <output>
2501
+ <port id="2" precision="FP32" names="output">
2502
+ <dim>1</dim>
2503
+ <dim>1</dim>
2504
+ <dim>256</dim>
2505
+ <dim>256</dim>
2506
+ </port>
2507
+ </output>
2508
+ </layer>
2509
+ <layer id="151" name="output/sink_port_0" type="Result" version="opset1">
2510
+ <input>
2511
+ <port id="0" precision="FP32">
2512
+ <dim>1</dim>
2513
+ <dim>1</dim>
2514
+ <dim>256</dim>
2515
+ <dim>256</dim>
2516
+ </port>
2517
+ </input>
2518
+ </layer>
2519
+ </layers>
2520
+ <edges>
2521
+ <edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
2522
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="1" />
2523
+ <edge from-layer="2" from-port="2" to-layer="4" to-port="0" />
2524
+ <edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
2525
+ <edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
2526
+ <edge from-layer="5" from-port="1" to-layer="6" to-port="0" />
2527
+ <edge from-layer="6" from-port="1" to-layer="8" to-port="0" />
2528
+ <edge from-layer="6" from-port="1" to-layer="16" to-port="1" />
2529
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="1" />
2530
+ <edge from-layer="8" from-port="2" to-layer="10" to-port="0" />
2531
+ <edge from-layer="9" from-port="0" to-layer="10" to-port="1" />
2532
+ <edge from-layer="10" from-port="2" to-layer="11" to-port="0" />
2533
+ <edge from-layer="11" from-port="1" to-layer="13" to-port="0" />
2534
+ <edge from-layer="12" from-port="0" to-layer="13" to-port="1" />
2535
+ <edge from-layer="13" from-port="2" to-layer="15" to-port="0" />
2536
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="1" />
2537
+ <edge from-layer="15" from-port="2" to-layer="16" to-port="0" />
2538
+ <edge from-layer="16" from-port="2" to-layer="17" to-port="0" />
2539
+ <edge from-layer="17" from-port="1" to-layer="19" to-port="0" />
2540
+ <edge from-layer="17" from-port="1" to-layer="27" to-port="1" />
2541
+ <edge from-layer="18" from-port="0" to-layer="19" to-port="1" />
2542
+ <edge from-layer="19" from-port="2" to-layer="21" to-port="0" />
2543
+ <edge from-layer="20" from-port="0" to-layer="21" to-port="1" />
2544
+ <edge from-layer="21" from-port="2" to-layer="22" to-port="0" />
2545
+ <edge from-layer="22" from-port="1" to-layer="24" to-port="0" />
2546
+ <edge from-layer="23" from-port="0" to-layer="24" to-port="1" />
2547
+ <edge from-layer="24" from-port="2" to-layer="26" to-port="0" />
2548
+ <edge from-layer="25" from-port="0" to-layer="26" to-port="1" />
2549
+ <edge from-layer="26" from-port="2" to-layer="27" to-port="0" />
2550
+ <edge from-layer="27" from-port="2" to-layer="28" to-port="0" />
2551
+ <edge from-layer="28" from-port="1" to-layer="30" to-port="0" />
2552
+ <edge from-layer="28" from-port="1" to-layer="39" to-port="0" />
2553
+ <edge from-layer="28" from-port="1" to-layer="68" to-port="0" />
2554
+ <edge from-layer="29" from-port="0" to-layer="30" to-port="1" />
2555
+ <edge from-layer="30" from-port="2" to-layer="32" to-port="0" />
2556
+ <edge from-layer="31" from-port="0" to-layer="32" to-port="1" />
2557
+ <edge from-layer="32" from-port="2" to-layer="33" to-port="0" />
2558
+ <edge from-layer="33" from-port="1" to-layer="35" to-port="0" />
2559
+ <edge from-layer="34" from-port="0" to-layer="35" to-port="1" />
2560
+ <edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
2561
+ <edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
2562
+ <edge from-layer="37" from-port="2" to-layer="42" to-port="0" />
2563
+ <edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
2564
+ <edge from-layer="39" from-port="2" to-layer="41" to-port="0" />
2565
+ <edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
2566
+ <edge from-layer="41" from-port="2" to-layer="42" to-port="1" />
2567
+ <edge from-layer="42" from-port="2" to-layer="43" to-port="0" />
2568
+ <edge from-layer="43" from-port="1" to-layer="45" to-port="0" />
2569
+ <edge from-layer="43" from-port="1" to-layer="53" to-port="1" />
2570
+ <edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
2571
+ <edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
2572
+ <edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
2573
+ <edge from-layer="47" from-port="2" to-layer="48" to-port="0" />
2574
+ <edge from-layer="48" from-port="1" to-layer="50" to-port="0" />
2575
+ <edge from-layer="49" from-port="0" to-layer="50" to-port="1" />
2576
+ <edge from-layer="50" from-port="2" to-layer="52" to-port="0" />
2577
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="1" />
2578
+ <edge from-layer="52" from-port="2" to-layer="53" to-port="0" />
2579
+ <edge from-layer="53" from-port="2" to-layer="54" to-port="0" />
2580
+ <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
2581
+ <edge from-layer="54" from-port="1" to-layer="67" to-port="0" />
2582
+ <edge from-layer="54" from-port="1" to-layer="79" to-port="0" />
2583
+ <edge from-layer="54" from-port="1" to-layer="70" to-port="0" />
2584
+ <edge from-layer="55" from-port="1" to-layer="59" to-port="0" />
2585
+ <edge from-layer="55" from-port="1" to-layer="63" to-port="0" />
2586
+ <edge from-layer="56" from-port="0" to-layer="59" to-port="1" />
2587
+ <edge from-layer="57" from-port="0" to-layer="59" to-port="2" />
2588
+ <edge from-layer="58" from-port="0" to-layer="59" to-port="3" />
2589
+ <edge from-layer="59" from-port="4" to-layer="61" to-port="0" />
2590
+ <edge from-layer="60" from-port="0" to-layer="61" to-port="1" />
2591
+ <edge from-layer="61" from-port="2" to-layer="62" to-port="0" />
2592
+ <edge from-layer="61" from-port="2" to-layer="67" to-port="1" />
2593
+ <edge from-layer="62" from-port="1" to-layer="64" to-port="0" />
2594
+ <edge from-layer="63" from-port="1" to-layer="64" to-port="1" />
2595
+ <edge from-layer="64" from-port="2" to-layer="66" to-port="0" />
2596
+ <edge from-layer="65" from-port="0" to-layer="66" to-port="1" />
2597
+ <edge from-layer="66" from-port="2" to-layer="67" to-port="2" />
2598
+ <edge from-layer="67" from-port="3" to-layer="68" to-port="1" />
2599
+ <edge from-layer="68" from-port="2" to-layer="111" to-port="0" />
2600
+ <edge from-layer="68" from-port="2" to-layer="100" to-port="0" />
2601
+ <edge from-layer="69" from-port="0" to-layer="70" to-port="1" />
2602
+ <edge from-layer="70" from-port="2" to-layer="72" to-port="0" />
2603
+ <edge from-layer="71" from-port="0" to-layer="72" to-port="1" />
2604
+ <edge from-layer="72" from-port="2" to-layer="73" to-port="0" />
2605
+ <edge from-layer="73" from-port="1" to-layer="75" to-port="0" />
2606
+ <edge from-layer="74" from-port="0" to-layer="75" to-port="1" />
2607
+ <edge from-layer="75" from-port="2" to-layer="77" to-port="0" />
2608
+ <edge from-layer="76" from-port="0" to-layer="77" to-port="1" />
2609
+ <edge from-layer="77" from-port="2" to-layer="82" to-port="0" />
2610
+ <edge from-layer="78" from-port="0" to-layer="79" to-port="1" />
2611
+ <edge from-layer="79" from-port="2" to-layer="81" to-port="0" />
2612
+ <edge from-layer="80" from-port="0" to-layer="81" to-port="1" />
2613
+ <edge from-layer="81" from-port="2" to-layer="82" to-port="1" />
2614
+ <edge from-layer="82" from-port="2" to-layer="83" to-port="0" />
2615
+ <edge from-layer="83" from-port="1" to-layer="85" to-port="0" />
2616
+ <edge from-layer="83" from-port="1" to-layer="93" to-port="1" />
2617
+ <edge from-layer="84" from-port="0" to-layer="85" to-port="1" />
2618
+ <edge from-layer="85" from-port="2" to-layer="87" to-port="0" />
2619
+ <edge from-layer="86" from-port="0" to-layer="87" to-port="1" />
2620
+ <edge from-layer="87" from-port="2" to-layer="88" to-port="0" />
2621
+ <edge from-layer="88" from-port="1" to-layer="90" to-port="0" />
2622
+ <edge from-layer="89" from-port="0" to-layer="90" to-port="1" />
2623
+ <edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
2624
+ <edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
2625
+ <edge from-layer="92" from-port="2" to-layer="93" to-port="0" />
2626
+ <edge from-layer="93" from-port="2" to-layer="94" to-port="0" />
2627
+ <edge from-layer="94" from-port="1" to-layer="95" to-port="0" />
2628
+ <edge from-layer="94" from-port="1" to-layer="110" to-port="0" />
2629
+ <edge from-layer="95" from-port="1" to-layer="99" to-port="0" />
2630
+ <edge from-layer="95" from-port="1" to-layer="106" to-port="0" />
2631
+ <edge from-layer="96" from-port="0" to-layer="99" to-port="1" />
2632
+ <edge from-layer="97" from-port="0" to-layer="99" to-port="2" />
2633
+ <edge from-layer="98" from-port="0" to-layer="99" to-port="3" />
2634
+ <edge from-layer="99" from-port="4" to-layer="104" to-port="0" />
2635
+ <edge from-layer="100" from-port="1" to-layer="103" to-port="0" />
2636
+ <edge from-layer="101" from-port="0" to-layer="103" to-port="1" />
2637
+ <edge from-layer="102" from-port="0" to-layer="103" to-port="2" />
2638
+ <edge from-layer="103" from-port="3" to-layer="104" to-port="1" />
2639
+ <edge from-layer="104" from-port="2" to-layer="105" to-port="0" />
2640
+ <edge from-layer="104" from-port="2" to-layer="110" to-port="1" />
2641
+ <edge from-layer="105" from-port="1" to-layer="107" to-port="0" />
2642
+ <edge from-layer="106" from-port="1" to-layer="107" to-port="1" />
2643
+ <edge from-layer="107" from-port="2" to-layer="109" to-port="0" />
2644
+ <edge from-layer="108" from-port="0" to-layer="109" to-port="1" />
2645
+ <edge from-layer="109" from-port="2" to-layer="110" to-port="2" />
2646
+ <edge from-layer="110" from-port="3" to-layer="111" to-port="1" />
2647
+ <edge from-layer="111" from-port="2" to-layer="114" to-port="0" />
2648
+ <edge from-layer="112" from-port="0" to-layer="114" to-port="1" />
2649
+ <edge from-layer="113" from-port="0" to-layer="114" to-port="2" />
2650
+ <edge from-layer="114" from-port="3" to-layer="116" to-port="0" />
2651
+ <edge from-layer="115" from-port="0" to-layer="116" to-port="1" />
2652
+ <edge from-layer="116" from-port="2" to-layer="118" to-port="0" />
2653
+ <edge from-layer="117" from-port="0" to-layer="118" to-port="1" />
2654
+ <edge from-layer="118" from-port="2" to-layer="120" to-port="0" />
2655
+ <edge from-layer="119" from-port="0" to-layer="120" to-port="1" />
2656
+ <edge from-layer="120" from-port="2" to-layer="122" to-port="0" />
2657
+ <edge from-layer="120" from-port="2" to-layer="123" to-port="1" />
2658
+ <edge from-layer="121" from-port="0" to-layer="122" to-port="1" />
2659
+ <edge from-layer="122" from-port="2" to-layer="123" to-port="0" />
2660
+ <edge from-layer="123" from-port="2" to-layer="125" to-port="0" />
2661
+ <edge from-layer="124" from-port="0" to-layer="125" to-port="1" />
2662
+ <edge from-layer="125" from-port="2" to-layer="127" to-port="0" />
2663
+ <edge from-layer="126" from-port="0" to-layer="127" to-port="1" />
2664
+ <edge from-layer="127" from-port="2" to-layer="128" to-port="0" />
2665
+ <edge from-layer="128" from-port="1" to-layer="129" to-port="0" />
2666
+ <edge from-layer="129" from-port="1" to-layer="130" to-port="0" />
2667
+ <edge from-layer="129" from-port="1" to-layer="142" to-port="0" />
2668
+ <edge from-layer="130" from-port="1" to-layer="134" to-port="0" />
2669
+ <edge from-layer="130" from-port="1" to-layer="138" to-port="0" />
2670
+ <edge from-layer="131" from-port="0" to-layer="134" to-port="1" />
2671
+ <edge from-layer="132" from-port="0" to-layer="134" to-port="2" />
2672
+ <edge from-layer="133" from-port="0" to-layer="134" to-port="3" />
2673
+ <edge from-layer="134" from-port="4" to-layer="136" to-port="0" />
2674
+ <edge from-layer="135" from-port="0" to-layer="136" to-port="1" />
2675
+ <edge from-layer="136" from-port="2" to-layer="137" to-port="0" />
2676
+ <edge from-layer="136" from-port="2" to-layer="142" to-port="1" />
2677
+ <edge from-layer="137" from-port="1" to-layer="139" to-port="0" />
2678
+ <edge from-layer="138" from-port="1" to-layer="139" to-port="1" />
2679
+ <edge from-layer="139" from-port="2" to-layer="141" to-port="0" />
2680
+ <edge from-layer="140" from-port="0" to-layer="141" to-port="1" />
2681
+ <edge from-layer="141" from-port="2" to-layer="142" to-port="2" />
2682
+ <edge from-layer="142" from-port="3" to-layer="146" to-port="0" />
2683
+ <edge from-layer="143" from-port="0" to-layer="146" to-port="1" />
2684
+ <edge from-layer="144" from-port="0" to-layer="146" to-port="2" />
2685
+ <edge from-layer="145" from-port="0" to-layer="146" to-port="3" />
2686
+ <edge from-layer="146" from-port="4" to-layer="148" to-port="0" />
2687
+ <edge from-layer="147" from-port="0" to-layer="148" to-port="1" />
2688
+ <edge from-layer="148" from-port="2" to-layer="150" to-port="0" />
2689
+ <edge from-layer="149" from-port="0" to-layer="150" to-port="1" />
2690
+ <edge from-layer="150" from-port="2" to-layer="151" to-port="0" />
2691
+ </edges>
2692
+ <rt_info>
2693
+ <MO_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2694
+ <Runtime_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2695
+ <conversion_parameters>
2696
+ <framework value="onnx" />
2697
+ <input_model value="DIR\model.onnx" />
2698
+ <model_name value="model" />
2699
+ <output_dir value="DIR" />
2700
+ </conversion_parameters>
2701
+ <legacy_frontend value="False" />
2702
+ </rt_info>
2703
+ </net>
cubes/examples/001.jpg ADDED
cubes/examples/002.jpg ADDED
cubes/examples/003.jpg ADDED
cubes/examples/004.jpg ADDED
cubes/examples/005.jpg ADDED
cubes/examples/006.jpg ADDED
cubes/examples/007.jpg ADDED
cubes/examples/008.jpg ADDED
cubes/run/weights/openvino/metadata.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "task": "classification",
3
+ "transform": {
4
+ "__version__": "1.3.0",
5
+ "transform": {
6
+ "__class_fullname__": "Compose",
7
+ "p": 1.0,
8
+ "transforms": [
9
+ {
10
+ "__class_fullname__": "Resize",
11
+ "always_apply": true,
12
+ "p": 1,
13
+ "height": 256,
14
+ "width": 256,
15
+ "interpolation": 1
16
+ },
17
+ {
18
+ "__class_fullname__": "Normalize",
19
+ "always_apply": false,
20
+ "p": 1.0,
21
+ "mean": [
22
+ 0.485,
23
+ 0.456,
24
+ 0.406
25
+ ],
26
+ "std": [
27
+ 0.229,
28
+ 0.224,
29
+ 0.225
30
+ ],
31
+ "max_pixel_value": 255.0
32
+ },
33
+ {
34
+ "__class_fullname__": "ToTensorV2",
35
+ "always_apply": true,
36
+ "p": 1.0,
37
+ "transpose_mask": false
38
+ }
39
+ ],
40
+ "bbox_params": null,
41
+ "keypoint_params": null,
42
+ "additional_targets": {
43
+ "image": "image",
44
+ "depth_image": "image"
45
+ }
46
+ }
47
+ },
48
+ "image_threshold": 37.23301696777344,
49
+ "pixel_threshold": 37.23301696777344,
50
+ "min": 1.7443453073501587,
51
+ "max": 57.53754806518555
52
+ }
cubes/run/weights/openvino/model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a722ea3da4f616de7463711d41794e9a3251d1fa22393ab694a0796e68c1945
3
+ size 176605956
cubes/run/weights/openvino/model.mapping ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <mapping>
3
+ <map>
4
+ <framework name="input" output_port_id="input" />
5
+ <IR name="input" output_port_id="0" />
6
+ </map>
7
+ <map>
8
+ <framework name="onnx::Conv_260" output_port_id="onnx::Conv_260" />
9
+ <IR name="onnx::Conv_260" output_port_id="0" />
10
+ </map>
11
+ <map>
12
+ <framework name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/conv1/Conv_output_0" />
13
+ <IR name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="2" />
14
+ </map>
15
+ <map>
16
+ <framework name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="/feature_extractor/feature_extractor/act1/Relu_output_0" />
17
+ <IR name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="1" />
18
+ </map>
19
+ <map>
20
+ <framework name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0" />
21
+ <IR name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="1" />
22
+ </map>
23
+ <map>
24
+ <framework name="onnx::Conv_263" output_port_id="onnx::Conv_263" />
25
+ <IR name="onnx::Conv_263" output_port_id="0" />
26
+ </map>
27
+ <map>
28
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0" />
29
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="2" />
30
+ </map>
31
+ <map>
32
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0" />
33
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="1" />
34
+ </map>
35
+ <map>
36
+ <framework name="onnx::Conv_266" output_port_id="onnx::Conv_266" />
37
+ <IR name="onnx::Conv_266" output_port_id="0" />
38
+ </map>
39
+ <map>
40
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0" />
41
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="2" />
42
+ </map>
43
+ <map>
44
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0" />
45
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="2" />
46
+ </map>
47
+ <map>
48
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0" />
49
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="1" />
50
+ </map>
51
+ <map>
52
+ <framework name="onnx::Conv_269" output_port_id="onnx::Conv_269" />
53
+ <IR name="onnx::Conv_269" output_port_id="0" />
54
+ </map>
55
+ <map>
56
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0" />
57
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="2" />
58
+ </map>
59
+ <map>
60
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0" />
61
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="1" />
62
+ </map>
63
+ <map>
64
+ <framework name="onnx::Conv_272" output_port_id="onnx::Conv_272" />
65
+ <IR name="onnx::Conv_272" output_port_id="0" />
66
+ </map>
67
+ <map>
68
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0" />
69
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="2" />
70
+ </map>
71
+ <map>
72
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0" />
73
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="2" />
74
+ </map>
75
+ <map>
76
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0" />
77
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="1" />
78
+ </map>
79
+ <map>
80
+ <framework name="onnx::Conv_275" output_port_id="onnx::Conv_275" />
81
+ <IR name="onnx::Conv_275" output_port_id="0" />
82
+ </map>
83
+ <map>
84
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0" />
85
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="2" />
86
+ </map>
87
+ <map>
88
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0" />
89
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="1" />
90
+ </map>
91
+ <map>
92
+ <framework name="onnx::Conv_278" output_port_id="onnx::Conv_278" />
93
+ <IR name="onnx::Conv_278" output_port_id="0" />
94
+ </map>
95
+ <map>
96
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0" />
97
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="2" />
98
+ </map>
99
+ <map>
100
+ <framework name="onnx::Conv_281" output_port_id="onnx::Conv_281" />
101
+ <IR name="onnx::Conv_281" output_port_id="0" />
102
+ </map>
103
+ <map>
104
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0" />
105
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="2" />
106
+ </map>
107
+ <map>
108
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0" />
109
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="2" />
110
+ </map>
111
+ <map>
112
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0" />
113
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="1" />
114
+ </map>
115
+ <map>
116
+ <framework name="onnx::Conv_284" output_port_id="onnx::Conv_284" />
117
+ <IR name="onnx::Conv_284" output_port_id="0" />
118
+ </map>
119
+ <map>
120
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0" />
121
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="2" />
122
+ </map>
123
+ <map>
124
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0" />
125
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="1" />
126
+ </map>
127
+ <map>
128
+ <framework name="onnx::Conv_287" output_port_id="onnx::Conv_287" />
129
+ <IR name="onnx::Conv_287" output_port_id="0" />
130
+ </map>
131
+ <map>
132
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0" />
133
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="2" />
134
+ </map>
135
+ <map>
136
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0" />
137
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="2" />
138
+ </map>
139
+ <map>
140
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0" />
141
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="1" />
142
+ </map>
143
+ <map>
144
+ <framework name="/Shape" output_port_id="/Shape_output_0" />
145
+ <IR name="/Shape" output_port_id="1" />
146
+ </map>
147
+ <map>
148
+ <framework name="/Constant_1" output_port_id="/Constant_1_output_0" />
149
+ <IR name="/Constant_1" output_port_id="0" />
150
+ </map>
151
+ <map>
152
+ <framework name="/Constant_2" output_port_id="/Constant_2_output_0" />
153
+ <IR name="/Constant_2" output_port_id="0" />
154
+ </map>
155
+ <map>
156
+ <framework name="/Slice" output_port_id="/Slice_output_0" />
157
+ <IR name="/Slice" output_port_id="4" />
158
+ </map>
159
+ <map>
160
+ <framework name="/Constant_3" output_port_id="/Constant_3_output_0" />
161
+ <IR name="/Constant_3" output_port_id="0" />
162
+ </map>
163
+ <map>
164
+ <framework name="/Concat" output_port_id="/Concat_output_0" />
165
+ <IR name="/Concat" output_port_id="2" />
166
+ </map>
167
+ <map>
168
+ <framework name="/Resize" output_port_id="/Resize_output_0" />
169
+ <IR name="/Resize" output_port_id="3" />
170
+ </map>
171
+ <map>
172
+ <framework name="/Concat_1" output_port_id="/Concat_1_output_0" />
173
+ <IR name="/Concat_1" output_port_id="2" />
174
+ </map>
175
+ <map>
176
+ <framework name="onnx::Conv_290" output_port_id="onnx::Conv_290" />
177
+ <IR name="onnx::Conv_290" output_port_id="0" />
178
+ </map>
179
+ <map>
180
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0" />
181
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="2" />
182
+ </map>
183
+ <map>
184
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0" />
185
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="1" />
186
+ </map>
187
+ <map>
188
+ <framework name="onnx::Conv_293" output_port_id="onnx::Conv_293" />
189
+ <IR name="onnx::Conv_293" output_port_id="0" />
190
+ </map>
191
+ <map>
192
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0" />
193
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="2" />
194
+ </map>
195
+ <map>
196
+ <framework name="onnx::Conv_296" output_port_id="onnx::Conv_296" />
197
+ <IR name="onnx::Conv_296" output_port_id="0" />
198
+ </map>
199
+ <map>
200
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0" />
201
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="2" />
202
+ </map>
203
+ <map>
204
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0" />
205
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="2" />
206
+ </map>
207
+ <map>
208
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0" />
209
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="1" />
210
+ </map>
211
+ <map>
212
+ <framework name="onnx::Conv_299" output_port_id="onnx::Conv_299" />
213
+ <IR name="onnx::Conv_299" output_port_id="0" />
214
+ </map>
215
+ <map>
216
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0" />
217
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="2" />
218
+ </map>
219
+ <map>
220
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0" />
221
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="1" />
222
+ </map>
223
+ <map>
224
+ <framework name="onnx::Conv_302" output_port_id="onnx::Conv_302" />
225
+ <IR name="onnx::Conv_302" output_port_id="0" />
226
+ </map>
227
+ <map>
228
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0" />
229
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="2" />
230
+ </map>
231
+ <map>
232
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0" />
233
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="2" />
234
+ </map>
235
+ <map>
236
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0" />
237
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="1" />
238
+ </map>
239
+ <map>
240
+ <framework name="/Shape_3" output_port_id="/Shape_3_output_0" />
241
+ <IR name="/Shape_3" output_port_id="1" />
242
+ </map>
243
+ <map>
244
+ <framework name="/Constant_9" output_port_id="/Constant_9_output_0" />
245
+ <IR name="/Constant_9" output_port_id="0" />
246
+ </map>
247
+ <map>
248
+ <framework name="/Constant_10" output_port_id="/Constant_10_output_0" />
249
+ <IR name="/Constant_10" output_port_id="0" />
250
+ </map>
251
+ <map>
252
+ <framework name="/Slice_1" output_port_id="/Slice_1_output_0" />
253
+ <IR name="/Slice_1" output_port_id="4" />
254
+ </map>
255
+ <map>
256
+ <framework name="/Shape_1" output_port_id="/Shape_1_output_0" />
257
+ <IR name="/Shape_1" output_port_id="1" />
258
+ </map>
259
+ <map>
260
+ <framework name="/Shape_1" output_port_id="/Shape_2_output_0" />
261
+ <IR name="/Shape_1" output_port_id="1" />
262
+ </map>
263
+ <map>
264
+ <framework name="Gather_6092" output_port_id="/Concat_2_output_0" />
265
+ <IR name="Gather_6092" output_port_id="3" />
266
+ </map>
267
+ <map>
268
+ <framework name="Gather_6092" output_port_id="/Cast_output_0" />
269
+ <IR name="Gather_6092" output_port_id="3" />
270
+ </map>
271
+ <map>
272
+ <framework name="/Concat_3" output_port_id="/Concat_3_output_0" />
273
+ <IR name="/Concat_3" output_port_id="2" />
274
+ </map>
275
+ <map>
276
+ <framework name="/Resize_1" output_port_id="/Resize_1_output_0" />
277
+ <IR name="/Resize_1" output_port_id="3" />
278
+ </map>
279
+ <map>
280
+ <framework name="/Concat_4" output_port_id="/Concat_4_output_0" />
281
+ <IR name="/Concat_4" output_port_id="2" />
282
+ </map>
283
+ <map>
284
+ <framework name="onnx::Gather_308" output_port_id="onnx::Gather_308" />
285
+ <IR name="onnx::Gather_308" output_port_id="0" />
286
+ </map>
287
+ <map>
288
+ <framework name="/Gather_2" output_port_id="/Gather_2_output_0" />
289
+ <IR name="/Gather_2" output_port_id="3" />
290
+ </map>
291
+ <map>
292
+ <framework name="/anomaly_map_generator/Constant" output_port_id="/anomaly_map_generator/Constant_output_0" />
293
+ <IR name="/anomaly_map_generator/Constant" output_port_id="0" />
294
+ </map>
295
+ <map>
296
+ <framework name="/anomaly_map_generator/Reshape" output_port_id="/anomaly_map_generator/Reshape_output_0" />
297
+ <IR name="/anomaly_map_generator/Reshape" output_port_id="2" />
298
+ </map>
299
+ <map>
300
+ <framework name="/anomaly_map_generator/Sub" output_port_id="/anomaly_map_generator/Sub_output_0" />
301
+ <IR name="/anomaly_map_generator/Sub" output_port_id="2" />
302
+ </map>
303
+ <map>
304
+ <framework name="/anomaly_map_generator/Transpose" output_port_id="/anomaly_map_generator/Transpose_output_0" />
305
+ <IR name="/anomaly_map_generator/Transpose" output_port_id="2" />
306
+ </map>
307
+ <map>
308
+ <framework name="onnx::MatMul_310" output_port_id="onnx::MatMul_310" />
309
+ <IR name="onnx::MatMul_310" output_port_id="0" />
310
+ </map>
311
+ <map>
312
+ <framework name="/anomaly_map_generator/MatMul" output_port_id="/anomaly_map_generator/MatMul_output_0" />
313
+ <IR name="/anomaly_map_generator/MatMul" output_port_id="2" />
314
+ </map>
315
+ <map>
316
+ <framework name="/anomaly_map_generator/Mul" output_port_id="/anomaly_map_generator/Mul_output_0" />
317
+ <IR name="/anomaly_map_generator/Mul" output_port_id="2" />
318
+ </map>
319
+ <map>
320
+ <framework name="/anomaly_map_generator/ReduceSum" output_port_id="/anomaly_map_generator/ReduceSum_output_0" />
321
+ <IR name="/anomaly_map_generator/ReduceSum" output_port_id="2" />
322
+ </map>
323
+ <map>
324
+ <framework name="/anomaly_map_generator/Constant_1" output_port_id="/anomaly_map_generator/Constant_1_output_0" />
325
+ <IR name="/anomaly_map_generator/Constant_1" output_port_id="0" />
326
+ </map>
327
+ <map>
328
+ <framework name="/anomaly_map_generator/Reshape_1" output_port_id="/anomaly_map_generator/Reshape_1_output_0" />
329
+ <IR name="/anomaly_map_generator/Reshape_1" output_port_id="2" />
330
+ </map>
331
+ <map>
332
+ <framework name="/anomaly_map_generator/Clip" output_port_id="/anomaly_map_generator/Clip_output_0" />
333
+ <IR name="/anomaly_map_generator/Clip" output_port_id="1" />
334
+ </map>
335
+ <map>
336
+ <framework name="/anomaly_map_generator/Sqrt" output_port_id="/anomaly_map_generator/Sqrt_output_0" />
337
+ <IR name="/anomaly_map_generator/Sqrt" output_port_id="1" />
338
+ </map>
339
+ <map>
340
+ <framework name="/anomaly_map_generator/Shape" output_port_id="/anomaly_map_generator/Shape_output_0" />
341
+ <IR name="/anomaly_map_generator/Shape" output_port_id="1" />
342
+ </map>
343
+ <map>
344
+ <framework name="/anomaly_map_generator/Constant_4" output_port_id="/anomaly_map_generator/Constant_4_output_0" />
345
+ <IR name="/anomaly_map_generator/Constant_4" output_port_id="0" />
346
+ </map>
347
+ <map>
348
+ <framework name="/anomaly_map_generator/Constant_5" output_port_id="/anomaly_map_generator/Constant_5_output_0" />
349
+ <IR name="/anomaly_map_generator/Constant_5" output_port_id="0" />
350
+ </map>
351
+ <map>
352
+ <framework name="/anomaly_map_generator/Slice" output_port_id="/anomaly_map_generator/Slice_output_0" />
353
+ <IR name="/anomaly_map_generator/Slice" output_port_id="4" />
354
+ </map>
355
+ <map>
356
+ <framework name="/anomaly_map_generator/Constant_6" output_port_id="/anomaly_map_generator/Constant_6_output_0" />
357
+ <IR name="/anomaly_map_generator/Constant_6" output_port_id="0" />
358
+ </map>
359
+ <map>
360
+ <framework name="/anomaly_map_generator/Concat" output_port_id="/anomaly_map_generator/Concat_output_0" />
361
+ <IR name="/anomaly_map_generator/Concat" output_port_id="2" />
362
+ </map>
363
+ <map>
364
+ <framework name="/anomaly_map_generator/Resize" output_port_id="/anomaly_map_generator/Resize_output_0" />
365
+ <IR name="/anomaly_map_generator/Resize" output_port_id="3" />
366
+ </map>
367
+ <map>
368
+ <framework name="/anomaly_map_generator/blur/Pad" output_port_id="/anomaly_map_generator/blur/Pad_output_0" />
369
+ <IR name="/anomaly_map_generator/blur/Pad" output_port_id="4" />
370
+ </map>
371
+ <map>
372
+ <framework name="anomaly_map_generator.blur.kernel" output_port_id="anomaly_map_generator.blur.kernel" />
373
+ <IR name="anomaly_map_generator.blur.kernel" output_port_id="0" />
374
+ </map>
375
+ <map>
376
+ <framework name="/anomaly_map_generator/blur/Conv" output_port_id="/anomaly_map_generator/blur/Conv_output_0" />
377
+ <IR name="/anomaly_map_generator/blur/Conv" output_port_id="2" />
378
+ </map>
379
+ <map>
380
+ <framework name="/anomaly_map_generator/blur/Constant_8" output_port_id="/anomaly_map_generator/blur/Constant_8_output_0" />
381
+ <IR name="/anomaly_map_generator/blur/Constant_8" output_port_id="0" />
382
+ </map>
383
+ <map>
384
+ <framework name="output" output_port_id="output" />
385
+ <IR name="output" output_port_id="2" />
386
+ </map>
387
+ </mapping>
cubes/run/weights/openvino/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:00bd7e7262e71aed2bc320412e13957efba5e40d380dd215267ef52cf73e73a8
3
+ size 176626689
cubes/run/weights/openvino/model.xml ADDED
@@ -0,0 +1,2703 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="torch_jit" version="11">
3
+ <layers>
4
+ <layer id="0" name="input" type="Parameter" version="opset1">
5
+ <data shape="1,3,256,256" element_type="f32" />
6
+ <output>
7
+ <port id="0" precision="FP32" names="input">
8
+ <dim>1</dim>
9
+ <dim>3</dim>
10
+ <dim>256</dim>
11
+ <dim>256</dim>
12
+ </port>
13
+ </output>
14
+ </layer>
15
+ <layer id="1" name="onnx::Conv_260" type="Const" version="opset1">
16
+ <data element_type="f32" shape="64, 3, 7, 7" offset="0" size="37632" />
17
+ <output>
18
+ <port id="0" precision="FP32" names="onnx::Conv_260">
19
+ <dim>64</dim>
20
+ <dim>3</dim>
21
+ <dim>7</dim>
22
+ <dim>7</dim>
23
+ </port>
24
+ </output>
25
+ </layer>
26
+ <layer id="2" name="/feature_extractor/feature_extractor/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
27
+ <data strides="2, 2" dilations="1, 1" pads_begin="3, 3" pads_end="3, 3" auto_pad="explicit" />
28
+ <input>
29
+ <port id="0" precision="FP32">
30
+ <dim>1</dim>
31
+ <dim>3</dim>
32
+ <dim>256</dim>
33
+ <dim>256</dim>
34
+ </port>
35
+ <port id="1" precision="FP32">
36
+ <dim>64</dim>
37
+ <dim>3</dim>
38
+ <dim>7</dim>
39
+ <dim>7</dim>
40
+ </port>
41
+ </input>
42
+ <output>
43
+ <port id="2" precision="FP32">
44
+ <dim>1</dim>
45
+ <dim>64</dim>
46
+ <dim>128</dim>
47
+ <dim>128</dim>
48
+ </port>
49
+ </output>
50
+ </layer>
51
+ <layer id="3" name="Reshape_55" type="Const" version="opset1">
52
+ <data element_type="f32" shape="1, 64, 1, 1" offset="37632" size="256" />
53
+ <output>
54
+ <port id="0" precision="FP32">
55
+ <dim>1</dim>
56
+ <dim>64</dim>
57
+ <dim>1</dim>
58
+ <dim>1</dim>
59
+ </port>
60
+ </output>
61
+ </layer>
62
+ <layer id="4" name="/feature_extractor/feature_extractor/conv1/Conv" type="Add" version="opset1">
63
+ <data auto_broadcast="numpy" />
64
+ <input>
65
+ <port id="0" precision="FP32">
66
+ <dim>1</dim>
67
+ <dim>64</dim>
68
+ <dim>128</dim>
69
+ <dim>128</dim>
70
+ </port>
71
+ <port id="1" precision="FP32">
72
+ <dim>1</dim>
73
+ <dim>64</dim>
74
+ <dim>1</dim>
75
+ <dim>1</dim>
76
+ </port>
77
+ </input>
78
+ <output>
79
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/conv1/Conv_output_0">
80
+ <dim>1</dim>
81
+ <dim>64</dim>
82
+ <dim>128</dim>
83
+ <dim>128</dim>
84
+ </port>
85
+ </output>
86
+ </layer>
87
+ <layer id="5" name="/feature_extractor/feature_extractor/act1/Relu" type="ReLU" version="opset1">
88
+ <input>
89
+ <port id="0" precision="FP32">
90
+ <dim>1</dim>
91
+ <dim>64</dim>
92
+ <dim>128</dim>
93
+ <dim>128</dim>
94
+ </port>
95
+ </input>
96
+ <output>
97
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/act1/Relu_output_0">
98
+ <dim>1</dim>
99
+ <dim>64</dim>
100
+ <dim>128</dim>
101
+ <dim>128</dim>
102
+ </port>
103
+ </output>
104
+ </layer>
105
+ <layer id="6" name="/feature_extractor/feature_extractor/maxpool/MaxPool" type="MaxPool" version="opset8">
106
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" kernel="3, 3" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
107
+ <input>
108
+ <port id="0" precision="FP32">
109
+ <dim>1</dim>
110
+ <dim>64</dim>
111
+ <dim>128</dim>
112
+ <dim>128</dim>
113
+ </port>
114
+ </input>
115
+ <output>
116
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0">
117
+ <dim>1</dim>
118
+ <dim>64</dim>
119
+ <dim>64</dim>
120
+ <dim>64</dim>
121
+ </port>
122
+ <port id="2" precision="I64">
123
+ <dim>1</dim>
124
+ <dim>64</dim>
125
+ <dim>64</dim>
126
+ <dim>64</dim>
127
+ </port>
128
+ </output>
129
+ </layer>
130
+ <layer id="7" name="onnx::Conv_263" type="Const" version="opset1">
131
+ <data element_type="f32" shape="64, 64, 3, 3" offset="37888" size="147456" />
132
+ <output>
133
+ <port id="0" precision="FP32" names="onnx::Conv_263">
134
+ <dim>64</dim>
135
+ <dim>64</dim>
136
+ <dim>3</dim>
137
+ <dim>3</dim>
138
+ </port>
139
+ </output>
140
+ </layer>
141
+ <layer id="8" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
142
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
143
+ <input>
144
+ <port id="0" precision="FP32">
145
+ <dim>1</dim>
146
+ <dim>64</dim>
147
+ <dim>64</dim>
148
+ <dim>64</dim>
149
+ </port>
150
+ <port id="1" precision="FP32">
151
+ <dim>64</dim>
152
+ <dim>64</dim>
153
+ <dim>3</dim>
154
+ <dim>3</dim>
155
+ </port>
156
+ </input>
157
+ <output>
158
+ <port id="2" precision="FP32">
159
+ <dim>1</dim>
160
+ <dim>64</dim>
161
+ <dim>64</dim>
162
+ <dim>64</dim>
163
+ </port>
164
+ </output>
165
+ </layer>
166
+ <layer id="9" name="Reshape_105" type="Const" version="opset1">
167
+ <data element_type="f32" shape="1, 64, 1, 1" offset="185344" size="256" />
168
+ <output>
169
+ <port id="0" precision="FP32">
170
+ <dim>1</dim>
171
+ <dim>64</dim>
172
+ <dim>1</dim>
173
+ <dim>1</dim>
174
+ </port>
175
+ </output>
176
+ </layer>
177
+ <layer id="10" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" type="Add" version="opset1">
178
+ <data auto_broadcast="numpy" />
179
+ <input>
180
+ <port id="0" precision="FP32">
181
+ <dim>1</dim>
182
+ <dim>64</dim>
183
+ <dim>64</dim>
184
+ <dim>64</dim>
185
+ </port>
186
+ <port id="1" precision="FP32">
187
+ <dim>1</dim>
188
+ <dim>64</dim>
189
+ <dim>1</dim>
190
+ <dim>1</dim>
191
+ </port>
192
+ </input>
193
+ <output>
194
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0">
195
+ <dim>1</dim>
196
+ <dim>64</dim>
197
+ <dim>64</dim>
198
+ <dim>64</dim>
199
+ </port>
200
+ </output>
201
+ </layer>
202
+ <layer id="11" name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" type="ReLU" version="opset1">
203
+ <input>
204
+ <port id="0" precision="FP32">
205
+ <dim>1</dim>
206
+ <dim>64</dim>
207
+ <dim>64</dim>
208
+ <dim>64</dim>
209
+ </port>
210
+ </input>
211
+ <output>
212
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0">
213
+ <dim>1</dim>
214
+ <dim>64</dim>
215
+ <dim>64</dim>
216
+ <dim>64</dim>
217
+ </port>
218
+ </output>
219
+ </layer>
220
+ <layer id="12" name="onnx::Conv_266" type="Const" version="opset1">
221
+ <data element_type="f32" shape="64, 64, 3, 3" offset="185600" size="147456" />
222
+ <output>
223
+ <port id="0" precision="FP32" names="onnx::Conv_266">
224
+ <dim>64</dim>
225
+ <dim>64</dim>
226
+ <dim>3</dim>
227
+ <dim>3</dim>
228
+ </port>
229
+ </output>
230
+ </layer>
231
+ <layer id="13" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
232
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
233
+ <input>
234
+ <port id="0" precision="FP32">
235
+ <dim>1</dim>
236
+ <dim>64</dim>
237
+ <dim>64</dim>
238
+ <dim>64</dim>
239
+ </port>
240
+ <port id="1" precision="FP32">
241
+ <dim>64</dim>
242
+ <dim>64</dim>
243
+ <dim>3</dim>
244
+ <dim>3</dim>
245
+ </port>
246
+ </input>
247
+ <output>
248
+ <port id="2" precision="FP32">
249
+ <dim>1</dim>
250
+ <dim>64</dim>
251
+ <dim>64</dim>
252
+ <dim>64</dim>
253
+ </port>
254
+ </output>
255
+ </layer>
256
+ <layer id="14" name="Reshape_154" type="Const" version="opset1">
257
+ <data element_type="f32" shape="1, 64, 1, 1" offset="333056" size="256" />
258
+ <output>
259
+ <port id="0" precision="FP32">
260
+ <dim>1</dim>
261
+ <dim>64</dim>
262
+ <dim>1</dim>
263
+ <dim>1</dim>
264
+ </port>
265
+ </output>
266
+ </layer>
267
+ <layer id="15" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" type="Add" version="opset1">
268
+ <data auto_broadcast="numpy" />
269
+ <input>
270
+ <port id="0" precision="FP32">
271
+ <dim>1</dim>
272
+ <dim>64</dim>
273
+ <dim>64</dim>
274
+ <dim>64</dim>
275
+ </port>
276
+ <port id="1" precision="FP32">
277
+ <dim>1</dim>
278
+ <dim>64</dim>
279
+ <dim>1</dim>
280
+ <dim>1</dim>
281
+ </port>
282
+ </input>
283
+ <output>
284
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0">
285
+ <dim>1</dim>
286
+ <dim>64</dim>
287
+ <dim>64</dim>
288
+ <dim>64</dim>
289
+ </port>
290
+ </output>
291
+ </layer>
292
+ <layer id="16" name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" type="Add" version="opset1">
293
+ <data auto_broadcast="numpy" />
294
+ <input>
295
+ <port id="0" precision="FP32">
296
+ <dim>1</dim>
297
+ <dim>64</dim>
298
+ <dim>64</dim>
299
+ <dim>64</dim>
300
+ </port>
301
+ <port id="1" precision="FP32">
302
+ <dim>1</dim>
303
+ <dim>64</dim>
304
+ <dim>64</dim>
305
+ <dim>64</dim>
306
+ </port>
307
+ </input>
308
+ <output>
309
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0">
310
+ <dim>1</dim>
311
+ <dim>64</dim>
312
+ <dim>64</dim>
313
+ <dim>64</dim>
314
+ </port>
315
+ </output>
316
+ </layer>
317
+ <layer id="17" name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" type="ReLU" version="opset1">
318
+ <input>
319
+ <port id="0" precision="FP32">
320
+ <dim>1</dim>
321
+ <dim>64</dim>
322
+ <dim>64</dim>
323
+ <dim>64</dim>
324
+ </port>
325
+ </input>
326
+ <output>
327
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0">
328
+ <dim>1</dim>
329
+ <dim>64</dim>
330
+ <dim>64</dim>
331
+ <dim>64</dim>
332
+ </port>
333
+ </output>
334
+ </layer>
335
+ <layer id="18" name="onnx::Conv_269" type="Const" version="opset1">
336
+ <data element_type="f32" shape="64, 64, 3, 3" offset="333312" size="147456" />
337
+ <output>
338
+ <port id="0" precision="FP32" names="onnx::Conv_269">
339
+ <dim>64</dim>
340
+ <dim>64</dim>
341
+ <dim>3</dim>
342
+ <dim>3</dim>
343
+ </port>
344
+ </output>
345
+ </layer>
346
+ <layer id="19" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
347
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
348
+ <input>
349
+ <port id="0" precision="FP32">
350
+ <dim>1</dim>
351
+ <dim>64</dim>
352
+ <dim>64</dim>
353
+ <dim>64</dim>
354
+ </port>
355
+ <port id="1" precision="FP32">
356
+ <dim>64</dim>
357
+ <dim>64</dim>
358
+ <dim>3</dim>
359
+ <dim>3</dim>
360
+ </port>
361
+ </input>
362
+ <output>
363
+ <port id="2" precision="FP32">
364
+ <dim>1</dim>
365
+ <dim>64</dim>
366
+ <dim>64</dim>
367
+ <dim>64</dim>
368
+ </port>
369
+ </output>
370
+ </layer>
371
+ <layer id="20" name="Reshape_204" type="Const" version="opset1">
372
+ <data element_type="f32" shape="1, 64, 1, 1" offset="480768" size="256" />
373
+ <output>
374
+ <port id="0" precision="FP32">
375
+ <dim>1</dim>
376
+ <dim>64</dim>
377
+ <dim>1</dim>
378
+ <dim>1</dim>
379
+ </port>
380
+ </output>
381
+ </layer>
382
+ <layer id="21" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" type="Add" version="opset1">
383
+ <data auto_broadcast="numpy" />
384
+ <input>
385
+ <port id="0" precision="FP32">
386
+ <dim>1</dim>
387
+ <dim>64</dim>
388
+ <dim>64</dim>
389
+ <dim>64</dim>
390
+ </port>
391
+ <port id="1" precision="FP32">
392
+ <dim>1</dim>
393
+ <dim>64</dim>
394
+ <dim>1</dim>
395
+ <dim>1</dim>
396
+ </port>
397
+ </input>
398
+ <output>
399
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0">
400
+ <dim>1</dim>
401
+ <dim>64</dim>
402
+ <dim>64</dim>
403
+ <dim>64</dim>
404
+ </port>
405
+ </output>
406
+ </layer>
407
+ <layer id="22" name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" type="ReLU" version="opset1">
408
+ <input>
409
+ <port id="0" precision="FP32">
410
+ <dim>1</dim>
411
+ <dim>64</dim>
412
+ <dim>64</dim>
413
+ <dim>64</dim>
414
+ </port>
415
+ </input>
416
+ <output>
417
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0">
418
+ <dim>1</dim>
419
+ <dim>64</dim>
420
+ <dim>64</dim>
421
+ <dim>64</dim>
422
+ </port>
423
+ </output>
424
+ </layer>
425
+ <layer id="23" name="onnx::Conv_272" type="Const" version="opset1">
426
+ <data element_type="f32" shape="64, 64, 3, 3" offset="481024" size="147456" />
427
+ <output>
428
+ <port id="0" precision="FP32" names="onnx::Conv_272">
429
+ <dim>64</dim>
430
+ <dim>64</dim>
431
+ <dim>3</dim>
432
+ <dim>3</dim>
433
+ </port>
434
+ </output>
435
+ </layer>
436
+ <layer id="24" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
437
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
438
+ <input>
439
+ <port id="0" precision="FP32">
440
+ <dim>1</dim>
441
+ <dim>64</dim>
442
+ <dim>64</dim>
443
+ <dim>64</dim>
444
+ </port>
445
+ <port id="1" precision="FP32">
446
+ <dim>64</dim>
447
+ <dim>64</dim>
448
+ <dim>3</dim>
449
+ <dim>3</dim>
450
+ </port>
451
+ </input>
452
+ <output>
453
+ <port id="2" precision="FP32">
454
+ <dim>1</dim>
455
+ <dim>64</dim>
456
+ <dim>64</dim>
457
+ <dim>64</dim>
458
+ </port>
459
+ </output>
460
+ </layer>
461
+ <layer id="25" name="Reshape_253" type="Const" version="opset1">
462
+ <data element_type="f32" shape="1, 64, 1, 1" offset="628480" size="256" />
463
+ <output>
464
+ <port id="0" precision="FP32">
465
+ <dim>1</dim>
466
+ <dim>64</dim>
467
+ <dim>1</dim>
468
+ <dim>1</dim>
469
+ </port>
470
+ </output>
471
+ </layer>
472
+ <layer id="26" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" type="Add" version="opset1">
473
+ <data auto_broadcast="numpy" />
474
+ <input>
475
+ <port id="0" precision="FP32">
476
+ <dim>1</dim>
477
+ <dim>64</dim>
478
+ <dim>64</dim>
479
+ <dim>64</dim>
480
+ </port>
481
+ <port id="1" precision="FP32">
482
+ <dim>1</dim>
483
+ <dim>64</dim>
484
+ <dim>1</dim>
485
+ <dim>1</dim>
486
+ </port>
487
+ </input>
488
+ <output>
489
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0">
490
+ <dim>1</dim>
491
+ <dim>64</dim>
492
+ <dim>64</dim>
493
+ <dim>64</dim>
494
+ </port>
495
+ </output>
496
+ </layer>
497
+ <layer id="27" name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" type="Add" version="opset1">
498
+ <data auto_broadcast="numpy" />
499
+ <input>
500
+ <port id="0" precision="FP32">
501
+ <dim>1</dim>
502
+ <dim>64</dim>
503
+ <dim>64</dim>
504
+ <dim>64</dim>
505
+ </port>
506
+ <port id="1" precision="FP32">
507
+ <dim>1</dim>
508
+ <dim>64</dim>
509
+ <dim>64</dim>
510
+ <dim>64</dim>
511
+ </port>
512
+ </input>
513
+ <output>
514
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0">
515
+ <dim>1</dim>
516
+ <dim>64</dim>
517
+ <dim>64</dim>
518
+ <dim>64</dim>
519
+ </port>
520
+ </output>
521
+ </layer>
522
+ <layer id="28" name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" type="ReLU" version="opset1">
523
+ <input>
524
+ <port id="0" precision="FP32">
525
+ <dim>1</dim>
526
+ <dim>64</dim>
527
+ <dim>64</dim>
528
+ <dim>64</dim>
529
+ </port>
530
+ </input>
531
+ <output>
532
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0">
533
+ <dim>1</dim>
534
+ <dim>64</dim>
535
+ <dim>64</dim>
536
+ <dim>64</dim>
537
+ </port>
538
+ </output>
539
+ </layer>
540
+ <layer id="29" name="onnx::Conv_275" type="Const" version="opset1">
541
+ <data element_type="f32" shape="128, 64, 3, 3" offset="628736" size="294912" />
542
+ <output>
543
+ <port id="0" precision="FP32" names="onnx::Conv_275">
544
+ <dim>128</dim>
545
+ <dim>64</dim>
546
+ <dim>3</dim>
547
+ <dim>3</dim>
548
+ </port>
549
+ </output>
550
+ </layer>
551
+ <layer id="30" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
552
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
553
+ <input>
554
+ <port id="0" precision="FP32">
555
+ <dim>1</dim>
556
+ <dim>64</dim>
557
+ <dim>64</dim>
558
+ <dim>64</dim>
559
+ </port>
560
+ <port id="1" precision="FP32">
561
+ <dim>128</dim>
562
+ <dim>64</dim>
563
+ <dim>3</dim>
564
+ <dim>3</dim>
565
+ </port>
566
+ </input>
567
+ <output>
568
+ <port id="2" precision="FP32">
569
+ <dim>1</dim>
570
+ <dim>128</dim>
571
+ <dim>32</dim>
572
+ <dim>32</dim>
573
+ </port>
574
+ </output>
575
+ </layer>
576
+ <layer id="31" name="Reshape_303" type="Const" version="opset1">
577
+ <data element_type="f32" shape="1, 128, 1, 1" offset="923648" size="512" />
578
+ <output>
579
+ <port id="0" precision="FP32">
580
+ <dim>1</dim>
581
+ <dim>128</dim>
582
+ <dim>1</dim>
583
+ <dim>1</dim>
584
+ </port>
585
+ </output>
586
+ </layer>
587
+ <layer id="32" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" type="Add" version="opset1">
588
+ <data auto_broadcast="numpy" />
589
+ <input>
590
+ <port id="0" precision="FP32">
591
+ <dim>1</dim>
592
+ <dim>128</dim>
593
+ <dim>32</dim>
594
+ <dim>32</dim>
595
+ </port>
596
+ <port id="1" precision="FP32">
597
+ <dim>1</dim>
598
+ <dim>128</dim>
599
+ <dim>1</dim>
600
+ <dim>1</dim>
601
+ </port>
602
+ </input>
603
+ <output>
604
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0">
605
+ <dim>1</dim>
606
+ <dim>128</dim>
607
+ <dim>32</dim>
608
+ <dim>32</dim>
609
+ </port>
610
+ </output>
611
+ </layer>
612
+ <layer id="33" name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" type="ReLU" version="opset1">
613
+ <input>
614
+ <port id="0" precision="FP32">
615
+ <dim>1</dim>
616
+ <dim>128</dim>
617
+ <dim>32</dim>
618
+ <dim>32</dim>
619
+ </port>
620
+ </input>
621
+ <output>
622
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0">
623
+ <dim>1</dim>
624
+ <dim>128</dim>
625
+ <dim>32</dim>
626
+ <dim>32</dim>
627
+ </port>
628
+ </output>
629
+ </layer>
630
+ <layer id="34" name="onnx::Conv_278" type="Const" version="opset1">
631
+ <data element_type="f32" shape="128, 128, 3, 3" offset="924160" size="589824" />
632
+ <output>
633
+ <port id="0" precision="FP32" names="onnx::Conv_278">
634
+ <dim>128</dim>
635
+ <dim>128</dim>
636
+ <dim>3</dim>
637
+ <dim>3</dim>
638
+ </port>
639
+ </output>
640
+ </layer>
641
+ <layer id="35" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
642
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
643
+ <input>
644
+ <port id="0" precision="FP32">
645
+ <dim>1</dim>
646
+ <dim>128</dim>
647
+ <dim>32</dim>
648
+ <dim>32</dim>
649
+ </port>
650
+ <port id="1" precision="FP32">
651
+ <dim>128</dim>
652
+ <dim>128</dim>
653
+ <dim>3</dim>
654
+ <dim>3</dim>
655
+ </port>
656
+ </input>
657
+ <output>
658
+ <port id="2" precision="FP32">
659
+ <dim>1</dim>
660
+ <dim>128</dim>
661
+ <dim>32</dim>
662
+ <dim>32</dim>
663
+ </port>
664
+ </output>
665
+ </layer>
666
+ <layer id="36" name="Reshape_352" type="Const" version="opset1">
667
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1513984" size="512" />
668
+ <output>
669
+ <port id="0" precision="FP32">
670
+ <dim>1</dim>
671
+ <dim>128</dim>
672
+ <dim>1</dim>
673
+ <dim>1</dim>
674
+ </port>
675
+ </output>
676
+ </layer>
677
+ <layer id="37" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" type="Add" version="opset1">
678
+ <data auto_broadcast="numpy" />
679
+ <input>
680
+ <port id="0" precision="FP32">
681
+ <dim>1</dim>
682
+ <dim>128</dim>
683
+ <dim>32</dim>
684
+ <dim>32</dim>
685
+ </port>
686
+ <port id="1" precision="FP32">
687
+ <dim>1</dim>
688
+ <dim>128</dim>
689
+ <dim>1</dim>
690
+ <dim>1</dim>
691
+ </port>
692
+ </input>
693
+ <output>
694
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0">
695
+ <dim>1</dim>
696
+ <dim>128</dim>
697
+ <dim>32</dim>
698
+ <dim>32</dim>
699
+ </port>
700
+ </output>
701
+ </layer>
702
+ <layer id="38" name="onnx::Conv_281" type="Const" version="opset1">
703
+ <data element_type="f32" shape="128, 64, 1, 1" offset="1514496" size="32768" />
704
+ <output>
705
+ <port id="0" precision="FP32" names="onnx::Conv_281">
706
+ <dim>128</dim>
707
+ <dim>64</dim>
708
+ <dim>1</dim>
709
+ <dim>1</dim>
710
+ </port>
711
+ </output>
712
+ </layer>
713
+ <layer id="39" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
714
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
715
+ <input>
716
+ <port id="0" precision="FP32">
717
+ <dim>1</dim>
718
+ <dim>64</dim>
719
+ <dim>64</dim>
720
+ <dim>64</dim>
721
+ </port>
722
+ <port id="1" precision="FP32">
723
+ <dim>128</dim>
724
+ <dim>64</dim>
725
+ <dim>1</dim>
726
+ <dim>1</dim>
727
+ </port>
728
+ </input>
729
+ <output>
730
+ <port id="2" precision="FP32">
731
+ <dim>1</dim>
732
+ <dim>128</dim>
733
+ <dim>32</dim>
734
+ <dim>32</dim>
735
+ </port>
736
+ </output>
737
+ </layer>
738
+ <layer id="40" name="Reshape_400" type="Const" version="opset1">
739
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1547264" size="512" />
740
+ <output>
741
+ <port id="0" precision="FP32">
742
+ <dim>1</dim>
743
+ <dim>128</dim>
744
+ <dim>1</dim>
745
+ <dim>1</dim>
746
+ </port>
747
+ </output>
748
+ </layer>
749
+ <layer id="41" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" type="Add" version="opset1">
750
+ <data auto_broadcast="numpy" />
751
+ <input>
752
+ <port id="0" precision="FP32">
753
+ <dim>1</dim>
754
+ <dim>128</dim>
755
+ <dim>32</dim>
756
+ <dim>32</dim>
757
+ </port>
758
+ <port id="1" precision="FP32">
759
+ <dim>1</dim>
760
+ <dim>128</dim>
761
+ <dim>1</dim>
762
+ <dim>1</dim>
763
+ </port>
764
+ </input>
765
+ <output>
766
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0">
767
+ <dim>1</dim>
768
+ <dim>128</dim>
769
+ <dim>32</dim>
770
+ <dim>32</dim>
771
+ </port>
772
+ </output>
773
+ </layer>
774
+ <layer id="42" name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" type="Add" version="opset1">
775
+ <data auto_broadcast="numpy" />
776
+ <input>
777
+ <port id="0" precision="FP32">
778
+ <dim>1</dim>
779
+ <dim>128</dim>
780
+ <dim>32</dim>
781
+ <dim>32</dim>
782
+ </port>
783
+ <port id="1" precision="FP32">
784
+ <dim>1</dim>
785
+ <dim>128</dim>
786
+ <dim>32</dim>
787
+ <dim>32</dim>
788
+ </port>
789
+ </input>
790
+ <output>
791
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0">
792
+ <dim>1</dim>
793
+ <dim>128</dim>
794
+ <dim>32</dim>
795
+ <dim>32</dim>
796
+ </port>
797
+ </output>
798
+ </layer>
799
+ <layer id="43" name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" type="ReLU" version="opset1">
800
+ <input>
801
+ <port id="0" precision="FP32">
802
+ <dim>1</dim>
803
+ <dim>128</dim>
804
+ <dim>32</dim>
805
+ <dim>32</dim>
806
+ </port>
807
+ </input>
808
+ <output>
809
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0">
810
+ <dim>1</dim>
811
+ <dim>128</dim>
812
+ <dim>32</dim>
813
+ <dim>32</dim>
814
+ </port>
815
+ </output>
816
+ </layer>
817
+ <layer id="44" name="onnx::Conv_284" type="Const" version="opset1">
818
+ <data element_type="f32" shape="128, 128, 3, 3" offset="1547776" size="589824" />
819
+ <output>
820
+ <port id="0" precision="FP32" names="onnx::Conv_284">
821
+ <dim>128</dim>
822
+ <dim>128</dim>
823
+ <dim>3</dim>
824
+ <dim>3</dim>
825
+ </port>
826
+ </output>
827
+ </layer>
828
+ <layer id="45" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
829
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
830
+ <input>
831
+ <port id="0" precision="FP32">
832
+ <dim>1</dim>
833
+ <dim>128</dim>
834
+ <dim>32</dim>
835
+ <dim>32</dim>
836
+ </port>
837
+ <port id="1" precision="FP32">
838
+ <dim>128</dim>
839
+ <dim>128</dim>
840
+ <dim>3</dim>
841
+ <dim>3</dim>
842
+ </port>
843
+ </input>
844
+ <output>
845
+ <port id="2" precision="FP32">
846
+ <dim>1</dim>
847
+ <dim>128</dim>
848
+ <dim>32</dim>
849
+ <dim>32</dim>
850
+ </port>
851
+ </output>
852
+ </layer>
853
+ <layer id="46" name="Reshape_450" type="Const" version="opset1">
854
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2137600" size="512" />
855
+ <output>
856
+ <port id="0" precision="FP32">
857
+ <dim>1</dim>
858
+ <dim>128</dim>
859
+ <dim>1</dim>
860
+ <dim>1</dim>
861
+ </port>
862
+ </output>
863
+ </layer>
864
+ <layer id="47" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" type="Add" version="opset1">
865
+ <data auto_broadcast="numpy" />
866
+ <input>
867
+ <port id="0" precision="FP32">
868
+ <dim>1</dim>
869
+ <dim>128</dim>
870
+ <dim>32</dim>
871
+ <dim>32</dim>
872
+ </port>
873
+ <port id="1" precision="FP32">
874
+ <dim>1</dim>
875
+ <dim>128</dim>
876
+ <dim>1</dim>
877
+ <dim>1</dim>
878
+ </port>
879
+ </input>
880
+ <output>
881
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0">
882
+ <dim>1</dim>
883
+ <dim>128</dim>
884
+ <dim>32</dim>
885
+ <dim>32</dim>
886
+ </port>
887
+ </output>
888
+ </layer>
889
+ <layer id="48" name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" type="ReLU" version="opset1">
890
+ <input>
891
+ <port id="0" precision="FP32">
892
+ <dim>1</dim>
893
+ <dim>128</dim>
894
+ <dim>32</dim>
895
+ <dim>32</dim>
896
+ </port>
897
+ </input>
898
+ <output>
899
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0">
900
+ <dim>1</dim>
901
+ <dim>128</dim>
902
+ <dim>32</dim>
903
+ <dim>32</dim>
904
+ </port>
905
+ </output>
906
+ </layer>
907
+ <layer id="49" name="onnx::Conv_287" type="Const" version="opset1">
908
+ <data element_type="f32" shape="128, 128, 3, 3" offset="2138112" size="589824" />
909
+ <output>
910
+ <port id="0" precision="FP32" names="onnx::Conv_287">
911
+ <dim>128</dim>
912
+ <dim>128</dim>
913
+ <dim>3</dim>
914
+ <dim>3</dim>
915
+ </port>
916
+ </output>
917
+ </layer>
918
+ <layer id="50" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
919
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
920
+ <input>
921
+ <port id="0" precision="FP32">
922
+ <dim>1</dim>
923
+ <dim>128</dim>
924
+ <dim>32</dim>
925
+ <dim>32</dim>
926
+ </port>
927
+ <port id="1" precision="FP32">
928
+ <dim>128</dim>
929
+ <dim>128</dim>
930
+ <dim>3</dim>
931
+ <dim>3</dim>
932
+ </port>
933
+ </input>
934
+ <output>
935
+ <port id="2" precision="FP32">
936
+ <dim>1</dim>
937
+ <dim>128</dim>
938
+ <dim>32</dim>
939
+ <dim>32</dim>
940
+ </port>
941
+ </output>
942
+ </layer>
943
+ <layer id="51" name="Reshape_499" type="Const" version="opset1">
944
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2727936" size="512" />
945
+ <output>
946
+ <port id="0" precision="FP32">
947
+ <dim>1</dim>
948
+ <dim>128</dim>
949
+ <dim>1</dim>
950
+ <dim>1</dim>
951
+ </port>
952
+ </output>
953
+ </layer>
954
+ <layer id="52" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" type="Add" version="opset1">
955
+ <data auto_broadcast="numpy" />
956
+ <input>
957
+ <port id="0" precision="FP32">
958
+ <dim>1</dim>
959
+ <dim>128</dim>
960
+ <dim>32</dim>
961
+ <dim>32</dim>
962
+ </port>
963
+ <port id="1" precision="FP32">
964
+ <dim>1</dim>
965
+ <dim>128</dim>
966
+ <dim>1</dim>
967
+ <dim>1</dim>
968
+ </port>
969
+ </input>
970
+ <output>
971
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0">
972
+ <dim>1</dim>
973
+ <dim>128</dim>
974
+ <dim>32</dim>
975
+ <dim>32</dim>
976
+ </port>
977
+ </output>
978
+ </layer>
979
+ <layer id="53" name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" type="Add" version="opset1">
980
+ <data auto_broadcast="numpy" />
981
+ <input>
982
+ <port id="0" precision="FP32">
983
+ <dim>1</dim>
984
+ <dim>128</dim>
985
+ <dim>32</dim>
986
+ <dim>32</dim>
987
+ </port>
988
+ <port id="1" precision="FP32">
989
+ <dim>1</dim>
990
+ <dim>128</dim>
991
+ <dim>32</dim>
992
+ <dim>32</dim>
993
+ </port>
994
+ </input>
995
+ <output>
996
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0">
997
+ <dim>1</dim>
998
+ <dim>128</dim>
999
+ <dim>32</dim>
1000
+ <dim>32</dim>
1001
+ </port>
1002
+ </output>
1003
+ </layer>
1004
+ <layer id="54" name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" type="ReLU" version="opset1">
1005
+ <input>
1006
+ <port id="0" precision="FP32">
1007
+ <dim>1</dim>
1008
+ <dim>128</dim>
1009
+ <dim>32</dim>
1010
+ <dim>32</dim>
1011
+ </port>
1012
+ </input>
1013
+ <output>
1014
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0">
1015
+ <dim>1</dim>
1016
+ <dim>128</dim>
1017
+ <dim>32</dim>
1018
+ <dim>32</dim>
1019
+ </port>
1020
+ </output>
1021
+ </layer>
1022
+ <layer id="55" name="/Shape" type="ShapeOf" version="opset3">
1023
+ <data output_type="i64" />
1024
+ <input>
1025
+ <port id="0" precision="FP32">
1026
+ <dim>1</dim>
1027
+ <dim>128</dim>
1028
+ <dim>32</dim>
1029
+ <dim>32</dim>
1030
+ </port>
1031
+ </input>
1032
+ <output>
1033
+ <port id="1" precision="I64" names="/Shape_output_0">
1034
+ <dim>4</dim>
1035
+ </port>
1036
+ </output>
1037
+ </layer>
1038
+ <layer id="56" name="/Constant_1" type="Const" version="opset1">
1039
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1040
+ <output>
1041
+ <port id="0" precision="I64" names="/Constant_1_output_0">
1042
+ <dim>1</dim>
1043
+ </port>
1044
+ </output>
1045
+ </layer>
1046
+ <layer id="57" name="/Constant_2" type="Const" version="opset1">
1047
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1048
+ <output>
1049
+ <port id="0" precision="I64" names="/Constant_2_output_0">
1050
+ <dim>1</dim>
1051
+ </port>
1052
+ </output>
1053
+ </layer>
1054
+ <layer id="58" name="Broadcast_781" type="Const" version="opset1">
1055
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1056
+ <output>
1057
+ <port id="0" precision="I64">
1058
+ <dim>1</dim>
1059
+ </port>
1060
+ </output>
1061
+ </layer>
1062
+ <layer id="59" name="/Slice" type="StridedSlice" version="opset1">
1063
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1064
+ <input>
1065
+ <port id="0" precision="I64">
1066
+ <dim>4</dim>
1067
+ </port>
1068
+ <port id="1" precision="I64">
1069
+ <dim>1</dim>
1070
+ </port>
1071
+ <port id="2" precision="I64">
1072
+ <dim>1</dim>
1073
+ </port>
1074
+ <port id="3" precision="I64">
1075
+ <dim>1</dim>
1076
+ </port>
1077
+ </input>
1078
+ <output>
1079
+ <port id="4" precision="I64" names="/Slice_output_0">
1080
+ <dim>2</dim>
1081
+ </port>
1082
+ </output>
1083
+ </layer>
1084
+ <layer id="60" name="/Constant_3" type="Const" version="opset1">
1085
+ <data element_type="i64" shape="2" offset="2728472" size="16" />
1086
+ <output>
1087
+ <port id="0" precision="I64" names="/Constant_3_output_0">
1088
+ <dim>2</dim>
1089
+ </port>
1090
+ </output>
1091
+ </layer>
1092
+ <layer id="61" name="/Concat" type="Concat" version="opset1">
1093
+ <data axis="0" />
1094
+ <input>
1095
+ <port id="0" precision="I64">
1096
+ <dim>2</dim>
1097
+ </port>
1098
+ <port id="1" precision="I64">
1099
+ <dim>2</dim>
1100
+ </port>
1101
+ </input>
1102
+ <output>
1103
+ <port id="2" precision="I64" names="/Concat_output_0">
1104
+ <dim>4</dim>
1105
+ </port>
1106
+ </output>
1107
+ </layer>
1108
+ <layer id="62" name="Convert_804" type="Convert" version="opset1">
1109
+ <data destination_type="f32" />
1110
+ <input>
1111
+ <port id="0" precision="I64">
1112
+ <dim>4</dim>
1113
+ </port>
1114
+ </input>
1115
+ <output>
1116
+ <port id="1" precision="FP32">
1117
+ <dim>4</dim>
1118
+ </port>
1119
+ </output>
1120
+ </layer>
1121
+ <layer id="63" name="Convert_803" type="Convert" version="opset1">
1122
+ <data destination_type="f32" />
1123
+ <input>
1124
+ <port id="0" precision="I64">
1125
+ <dim>4</dim>
1126
+ </port>
1127
+ </input>
1128
+ <output>
1129
+ <port id="1" precision="FP32">
1130
+ <dim>4</dim>
1131
+ </port>
1132
+ </output>
1133
+ </layer>
1134
+ <layer id="64" name="Divide_805" type="Divide" version="opset1">
1135
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1136
+ <input>
1137
+ <port id="0" precision="FP32">
1138
+ <dim>4</dim>
1139
+ </port>
1140
+ <port id="1" precision="FP32">
1141
+ <dim>4</dim>
1142
+ </port>
1143
+ </input>
1144
+ <output>
1145
+ <port id="2" precision="FP32">
1146
+ <dim>4</dim>
1147
+ </port>
1148
+ </output>
1149
+ </layer>
1150
+ <layer id="65" name="Constant_6582" type="Const" version="opset1">
1151
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1152
+ <output>
1153
+ <port id="0" precision="FP32">
1154
+ <dim>1</dim>
1155
+ </port>
1156
+ </output>
1157
+ </layer>
1158
+ <layer id="66" name="Add_807" type="Add" version="opset1">
1159
+ <data auto_broadcast="numpy" />
1160
+ <input>
1161
+ <port id="0" precision="FP32">
1162
+ <dim>4</dim>
1163
+ </port>
1164
+ <port id="1" precision="FP32">
1165
+ <dim>1</dim>
1166
+ </port>
1167
+ </input>
1168
+ <output>
1169
+ <port id="2" precision="FP32">
1170
+ <dim>4</dim>
1171
+ </port>
1172
+ </output>
1173
+ </layer>
1174
+ <layer id="67" name="/Resize" type="Interpolate" version="opset4">
1175
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1176
+ <input>
1177
+ <port id="0" precision="FP32">
1178
+ <dim>1</dim>
1179
+ <dim>128</dim>
1180
+ <dim>32</dim>
1181
+ <dim>32</dim>
1182
+ </port>
1183
+ <port id="1" precision="I64">
1184
+ <dim>4</dim>
1185
+ </port>
1186
+ <port id="2" precision="FP32">
1187
+ <dim>4</dim>
1188
+ </port>
1189
+ </input>
1190
+ <output>
1191
+ <port id="3" precision="FP32" names="/Resize_output_0">
1192
+ <dim>1</dim>
1193
+ <dim>128</dim>
1194
+ <dim>64</dim>
1195
+ <dim>64</dim>
1196
+ </port>
1197
+ </output>
1198
+ </layer>
1199
+ <layer id="68" name="/Concat_1" type="Concat" version="opset1">
1200
+ <data axis="1" />
1201
+ <input>
1202
+ <port id="0" precision="FP32">
1203
+ <dim>1</dim>
1204
+ <dim>64</dim>
1205
+ <dim>64</dim>
1206
+ <dim>64</dim>
1207
+ </port>
1208
+ <port id="1" precision="FP32">
1209
+ <dim>1</dim>
1210
+ <dim>128</dim>
1211
+ <dim>64</dim>
1212
+ <dim>64</dim>
1213
+ </port>
1214
+ </input>
1215
+ <output>
1216
+ <port id="2" precision="FP32" names="/Concat_1_output_0">
1217
+ <dim>1</dim>
1218
+ <dim>192</dim>
1219
+ <dim>64</dim>
1220
+ <dim>64</dim>
1221
+ </port>
1222
+ </output>
1223
+ </layer>
1224
+ <layer id="69" name="onnx::Conv_290" type="Const" version="opset1">
1225
+ <data element_type="f32" shape="256, 128, 3, 3" offset="2728492" size="1179648" />
1226
+ <output>
1227
+ <port id="0" precision="FP32" names="onnx::Conv_290">
1228
+ <dim>256</dim>
1229
+ <dim>128</dim>
1230
+ <dim>3</dim>
1231
+ <dim>3</dim>
1232
+ </port>
1233
+ </output>
1234
+ </layer>
1235
+ <layer id="70" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1236
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1237
+ <input>
1238
+ <port id="0" precision="FP32">
1239
+ <dim>1</dim>
1240
+ <dim>128</dim>
1241
+ <dim>32</dim>
1242
+ <dim>32</dim>
1243
+ </port>
1244
+ <port id="1" precision="FP32">
1245
+ <dim>256</dim>
1246
+ <dim>128</dim>
1247
+ <dim>3</dim>
1248
+ <dim>3</dim>
1249
+ </port>
1250
+ </input>
1251
+ <output>
1252
+ <port id="2" precision="FP32">
1253
+ <dim>1</dim>
1254
+ <dim>256</dim>
1255
+ <dim>16</dim>
1256
+ <dim>16</dim>
1257
+ </port>
1258
+ </output>
1259
+ </layer>
1260
+ <layer id="71" name="Reshape_549" type="Const" version="opset1">
1261
+ <data element_type="f32" shape="1, 256, 1, 1" offset="3908140" size="1024" />
1262
+ <output>
1263
+ <port id="0" precision="FP32">
1264
+ <dim>1</dim>
1265
+ <dim>256</dim>
1266
+ <dim>1</dim>
1267
+ <dim>1</dim>
1268
+ </port>
1269
+ </output>
1270
+ </layer>
1271
+ <layer id="72" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" type="Add" version="opset1">
1272
+ <data auto_broadcast="numpy" />
1273
+ <input>
1274
+ <port id="0" precision="FP32">
1275
+ <dim>1</dim>
1276
+ <dim>256</dim>
1277
+ <dim>16</dim>
1278
+ <dim>16</dim>
1279
+ </port>
1280
+ <port id="1" precision="FP32">
1281
+ <dim>1</dim>
1282
+ <dim>256</dim>
1283
+ <dim>1</dim>
1284
+ <dim>1</dim>
1285
+ </port>
1286
+ </input>
1287
+ <output>
1288
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0">
1289
+ <dim>1</dim>
1290
+ <dim>256</dim>
1291
+ <dim>16</dim>
1292
+ <dim>16</dim>
1293
+ </port>
1294
+ </output>
1295
+ </layer>
1296
+ <layer id="73" name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" type="ReLU" version="opset1">
1297
+ <input>
1298
+ <port id="0" precision="FP32">
1299
+ <dim>1</dim>
1300
+ <dim>256</dim>
1301
+ <dim>16</dim>
1302
+ <dim>16</dim>
1303
+ </port>
1304
+ </input>
1305
+ <output>
1306
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0">
1307
+ <dim>1</dim>
1308
+ <dim>256</dim>
1309
+ <dim>16</dim>
1310
+ <dim>16</dim>
1311
+ </port>
1312
+ </output>
1313
+ </layer>
1314
+ <layer id="74" name="onnx::Conv_293" type="Const" version="opset1">
1315
+ <data element_type="f32" shape="256, 256, 3, 3" offset="3909164" size="2359296" />
1316
+ <output>
1317
+ <port id="0" precision="FP32" names="onnx::Conv_293">
1318
+ <dim>256</dim>
1319
+ <dim>256</dim>
1320
+ <dim>3</dim>
1321
+ <dim>3</dim>
1322
+ </port>
1323
+ </output>
1324
+ </layer>
1325
+ <layer id="75" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1326
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1327
+ <input>
1328
+ <port id="0" precision="FP32">
1329
+ <dim>1</dim>
1330
+ <dim>256</dim>
1331
+ <dim>16</dim>
1332
+ <dim>16</dim>
1333
+ </port>
1334
+ <port id="1" precision="FP32">
1335
+ <dim>256</dim>
1336
+ <dim>256</dim>
1337
+ <dim>3</dim>
1338
+ <dim>3</dim>
1339
+ </port>
1340
+ </input>
1341
+ <output>
1342
+ <port id="2" precision="FP32">
1343
+ <dim>1</dim>
1344
+ <dim>256</dim>
1345
+ <dim>16</dim>
1346
+ <dim>16</dim>
1347
+ </port>
1348
+ </output>
1349
+ </layer>
1350
+ <layer id="76" name="Reshape_598" type="Const" version="opset1">
1351
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6268460" size="1024" />
1352
+ <output>
1353
+ <port id="0" precision="FP32">
1354
+ <dim>1</dim>
1355
+ <dim>256</dim>
1356
+ <dim>1</dim>
1357
+ <dim>1</dim>
1358
+ </port>
1359
+ </output>
1360
+ </layer>
1361
+ <layer id="77" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" type="Add" version="opset1">
1362
+ <data auto_broadcast="numpy" />
1363
+ <input>
1364
+ <port id="0" precision="FP32">
1365
+ <dim>1</dim>
1366
+ <dim>256</dim>
1367
+ <dim>16</dim>
1368
+ <dim>16</dim>
1369
+ </port>
1370
+ <port id="1" precision="FP32">
1371
+ <dim>1</dim>
1372
+ <dim>256</dim>
1373
+ <dim>1</dim>
1374
+ <dim>1</dim>
1375
+ </port>
1376
+ </input>
1377
+ <output>
1378
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0">
1379
+ <dim>1</dim>
1380
+ <dim>256</dim>
1381
+ <dim>16</dim>
1382
+ <dim>16</dim>
1383
+ </port>
1384
+ </output>
1385
+ </layer>
1386
+ <layer id="78" name="onnx::Conv_296" type="Const" version="opset1">
1387
+ <data element_type="f32" shape="256, 128, 1, 1" offset="6269484" size="131072" />
1388
+ <output>
1389
+ <port id="0" precision="FP32" names="onnx::Conv_296">
1390
+ <dim>256</dim>
1391
+ <dim>128</dim>
1392
+ <dim>1</dim>
1393
+ <dim>1</dim>
1394
+ </port>
1395
+ </output>
1396
+ </layer>
1397
+ <layer id="79" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
1398
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
1399
+ <input>
1400
+ <port id="0" precision="FP32">
1401
+ <dim>1</dim>
1402
+ <dim>128</dim>
1403
+ <dim>32</dim>
1404
+ <dim>32</dim>
1405
+ </port>
1406
+ <port id="1" precision="FP32">
1407
+ <dim>256</dim>
1408
+ <dim>128</dim>
1409
+ <dim>1</dim>
1410
+ <dim>1</dim>
1411
+ </port>
1412
+ </input>
1413
+ <output>
1414
+ <port id="2" precision="FP32">
1415
+ <dim>1</dim>
1416
+ <dim>256</dim>
1417
+ <dim>16</dim>
1418
+ <dim>16</dim>
1419
+ </port>
1420
+ </output>
1421
+ </layer>
1422
+ <layer id="80" name="Reshape_646" type="Const" version="opset1">
1423
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6400556" size="1024" />
1424
+ <output>
1425
+ <port id="0" precision="FP32">
1426
+ <dim>1</dim>
1427
+ <dim>256</dim>
1428
+ <dim>1</dim>
1429
+ <dim>1</dim>
1430
+ </port>
1431
+ </output>
1432
+ </layer>
1433
+ <layer id="81" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" type="Add" version="opset1">
1434
+ <data auto_broadcast="numpy" />
1435
+ <input>
1436
+ <port id="0" precision="FP32">
1437
+ <dim>1</dim>
1438
+ <dim>256</dim>
1439
+ <dim>16</dim>
1440
+ <dim>16</dim>
1441
+ </port>
1442
+ <port id="1" precision="FP32">
1443
+ <dim>1</dim>
1444
+ <dim>256</dim>
1445
+ <dim>1</dim>
1446
+ <dim>1</dim>
1447
+ </port>
1448
+ </input>
1449
+ <output>
1450
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0">
1451
+ <dim>1</dim>
1452
+ <dim>256</dim>
1453
+ <dim>16</dim>
1454
+ <dim>16</dim>
1455
+ </port>
1456
+ </output>
1457
+ </layer>
1458
+ <layer id="82" name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" type="Add" version="opset1">
1459
+ <data auto_broadcast="numpy" />
1460
+ <input>
1461
+ <port id="0" precision="FP32">
1462
+ <dim>1</dim>
1463
+ <dim>256</dim>
1464
+ <dim>16</dim>
1465
+ <dim>16</dim>
1466
+ </port>
1467
+ <port id="1" precision="FP32">
1468
+ <dim>1</dim>
1469
+ <dim>256</dim>
1470
+ <dim>16</dim>
1471
+ <dim>16</dim>
1472
+ </port>
1473
+ </input>
1474
+ <output>
1475
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0">
1476
+ <dim>1</dim>
1477
+ <dim>256</dim>
1478
+ <dim>16</dim>
1479
+ <dim>16</dim>
1480
+ </port>
1481
+ </output>
1482
+ </layer>
1483
+ <layer id="83" name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" type="ReLU" version="opset1">
1484
+ <input>
1485
+ <port id="0" precision="FP32">
1486
+ <dim>1</dim>
1487
+ <dim>256</dim>
1488
+ <dim>16</dim>
1489
+ <dim>16</dim>
1490
+ </port>
1491
+ </input>
1492
+ <output>
1493
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0">
1494
+ <dim>1</dim>
1495
+ <dim>256</dim>
1496
+ <dim>16</dim>
1497
+ <dim>16</dim>
1498
+ </port>
1499
+ </output>
1500
+ </layer>
1501
+ <layer id="84" name="onnx::Conv_299" type="Const" version="opset1">
1502
+ <data element_type="f32" shape="256, 256, 3, 3" offset="6401580" size="2359296" />
1503
+ <output>
1504
+ <port id="0" precision="FP32" names="onnx::Conv_299">
1505
+ <dim>256</dim>
1506
+ <dim>256</dim>
1507
+ <dim>3</dim>
1508
+ <dim>3</dim>
1509
+ </port>
1510
+ </output>
1511
+ </layer>
1512
+ <layer id="85" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1513
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1514
+ <input>
1515
+ <port id="0" precision="FP32">
1516
+ <dim>1</dim>
1517
+ <dim>256</dim>
1518
+ <dim>16</dim>
1519
+ <dim>16</dim>
1520
+ </port>
1521
+ <port id="1" precision="FP32">
1522
+ <dim>256</dim>
1523
+ <dim>256</dim>
1524
+ <dim>3</dim>
1525
+ <dim>3</dim>
1526
+ </port>
1527
+ </input>
1528
+ <output>
1529
+ <port id="2" precision="FP32">
1530
+ <dim>1</dim>
1531
+ <dim>256</dim>
1532
+ <dim>16</dim>
1533
+ <dim>16</dim>
1534
+ </port>
1535
+ </output>
1536
+ </layer>
1537
+ <layer id="86" name="Reshape_696" type="Const" version="opset1">
1538
+ <data element_type="f32" shape="1, 256, 1, 1" offset="8760876" size="1024" />
1539
+ <output>
1540
+ <port id="0" precision="FP32">
1541
+ <dim>1</dim>
1542
+ <dim>256</dim>
1543
+ <dim>1</dim>
1544
+ <dim>1</dim>
1545
+ </port>
1546
+ </output>
1547
+ </layer>
1548
+ <layer id="87" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" type="Add" version="opset1">
1549
+ <data auto_broadcast="numpy" />
1550
+ <input>
1551
+ <port id="0" precision="FP32">
1552
+ <dim>1</dim>
1553
+ <dim>256</dim>
1554
+ <dim>16</dim>
1555
+ <dim>16</dim>
1556
+ </port>
1557
+ <port id="1" precision="FP32">
1558
+ <dim>1</dim>
1559
+ <dim>256</dim>
1560
+ <dim>1</dim>
1561
+ <dim>1</dim>
1562
+ </port>
1563
+ </input>
1564
+ <output>
1565
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0">
1566
+ <dim>1</dim>
1567
+ <dim>256</dim>
1568
+ <dim>16</dim>
1569
+ <dim>16</dim>
1570
+ </port>
1571
+ </output>
1572
+ </layer>
1573
+ <layer id="88" name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" type="ReLU" version="opset1">
1574
+ <input>
1575
+ <port id="0" precision="FP32">
1576
+ <dim>1</dim>
1577
+ <dim>256</dim>
1578
+ <dim>16</dim>
1579
+ <dim>16</dim>
1580
+ </port>
1581
+ </input>
1582
+ <output>
1583
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0">
1584
+ <dim>1</dim>
1585
+ <dim>256</dim>
1586
+ <dim>16</dim>
1587
+ <dim>16</dim>
1588
+ </port>
1589
+ </output>
1590
+ </layer>
1591
+ <layer id="89" name="onnx::Conv_302" type="Const" version="opset1">
1592
+ <data element_type="f32" shape="256, 256, 3, 3" offset="8761900" size="2359296" />
1593
+ <output>
1594
+ <port id="0" precision="FP32" names="onnx::Conv_302">
1595
+ <dim>256</dim>
1596
+ <dim>256</dim>
1597
+ <dim>3</dim>
1598
+ <dim>3</dim>
1599
+ </port>
1600
+ </output>
1601
+ </layer>
1602
+ <layer id="90" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1603
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1604
+ <input>
1605
+ <port id="0" precision="FP32">
1606
+ <dim>1</dim>
1607
+ <dim>256</dim>
1608
+ <dim>16</dim>
1609
+ <dim>16</dim>
1610
+ </port>
1611
+ <port id="1" precision="FP32">
1612
+ <dim>256</dim>
1613
+ <dim>256</dim>
1614
+ <dim>3</dim>
1615
+ <dim>3</dim>
1616
+ </port>
1617
+ </input>
1618
+ <output>
1619
+ <port id="2" precision="FP32">
1620
+ <dim>1</dim>
1621
+ <dim>256</dim>
1622
+ <dim>16</dim>
1623
+ <dim>16</dim>
1624
+ </port>
1625
+ </output>
1626
+ </layer>
1627
+ <layer id="91" name="Reshape_745" type="Const" version="opset1">
1628
+ <data element_type="f32" shape="1, 256, 1, 1" offset="11121196" size="1024" />
1629
+ <output>
1630
+ <port id="0" precision="FP32">
1631
+ <dim>1</dim>
1632
+ <dim>256</dim>
1633
+ <dim>1</dim>
1634
+ <dim>1</dim>
1635
+ </port>
1636
+ </output>
1637
+ </layer>
1638
+ <layer id="92" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" type="Add" version="opset1">
1639
+ <data auto_broadcast="numpy" />
1640
+ <input>
1641
+ <port id="0" precision="FP32">
1642
+ <dim>1</dim>
1643
+ <dim>256</dim>
1644
+ <dim>16</dim>
1645
+ <dim>16</dim>
1646
+ </port>
1647
+ <port id="1" precision="FP32">
1648
+ <dim>1</dim>
1649
+ <dim>256</dim>
1650
+ <dim>1</dim>
1651
+ <dim>1</dim>
1652
+ </port>
1653
+ </input>
1654
+ <output>
1655
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0">
1656
+ <dim>1</dim>
1657
+ <dim>256</dim>
1658
+ <dim>16</dim>
1659
+ <dim>16</dim>
1660
+ </port>
1661
+ </output>
1662
+ </layer>
1663
+ <layer id="93" name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" type="Add" version="opset1">
1664
+ <data auto_broadcast="numpy" />
1665
+ <input>
1666
+ <port id="0" precision="FP32">
1667
+ <dim>1</dim>
1668
+ <dim>256</dim>
1669
+ <dim>16</dim>
1670
+ <dim>16</dim>
1671
+ </port>
1672
+ <port id="1" precision="FP32">
1673
+ <dim>1</dim>
1674
+ <dim>256</dim>
1675
+ <dim>16</dim>
1676
+ <dim>16</dim>
1677
+ </port>
1678
+ </input>
1679
+ <output>
1680
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0">
1681
+ <dim>1</dim>
1682
+ <dim>256</dim>
1683
+ <dim>16</dim>
1684
+ <dim>16</dim>
1685
+ </port>
1686
+ </output>
1687
+ </layer>
1688
+ <layer id="94" name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" type="ReLU" version="opset1">
1689
+ <input>
1690
+ <port id="0" precision="FP32">
1691
+ <dim>1</dim>
1692
+ <dim>256</dim>
1693
+ <dim>16</dim>
1694
+ <dim>16</dim>
1695
+ </port>
1696
+ </input>
1697
+ <output>
1698
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0">
1699
+ <dim>1</dim>
1700
+ <dim>256</dim>
1701
+ <dim>16</dim>
1702
+ <dim>16</dim>
1703
+ </port>
1704
+ </output>
1705
+ </layer>
1706
+ <layer id="95" name="/Shape_3" type="ShapeOf" version="opset3">
1707
+ <data output_type="i64" />
1708
+ <input>
1709
+ <port id="0" precision="FP32">
1710
+ <dim>1</dim>
1711
+ <dim>256</dim>
1712
+ <dim>16</dim>
1713
+ <dim>16</dim>
1714
+ </port>
1715
+ </input>
1716
+ <output>
1717
+ <port id="1" precision="I64" names="/Shape_3_output_0">
1718
+ <dim>4</dim>
1719
+ </port>
1720
+ </output>
1721
+ </layer>
1722
+ <layer id="96" name="/Constant_9" type="Const" version="opset1">
1723
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1724
+ <output>
1725
+ <port id="0" precision="I64" names="/Constant_9_output_0">
1726
+ <dim>1</dim>
1727
+ </port>
1728
+ </output>
1729
+ </layer>
1730
+ <layer id="97" name="/Constant_10" type="Const" version="opset1">
1731
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1732
+ <output>
1733
+ <port id="0" precision="I64" names="/Constant_10_output_0">
1734
+ <dim>1</dim>
1735
+ </port>
1736
+ </output>
1737
+ </layer>
1738
+ <layer id="98" name="Broadcast_861" type="Const" version="opset1">
1739
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1740
+ <output>
1741
+ <port id="0" precision="I64">
1742
+ <dim>1</dim>
1743
+ </port>
1744
+ </output>
1745
+ </layer>
1746
+ <layer id="99" name="/Slice_1" type="StridedSlice" version="opset1">
1747
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1748
+ <input>
1749
+ <port id="0" precision="I64">
1750
+ <dim>4</dim>
1751
+ </port>
1752
+ <port id="1" precision="I64">
1753
+ <dim>1</dim>
1754
+ </port>
1755
+ <port id="2" precision="I64">
1756
+ <dim>1</dim>
1757
+ </port>
1758
+ <port id="3" precision="I64">
1759
+ <dim>1</dim>
1760
+ </port>
1761
+ </input>
1762
+ <output>
1763
+ <port id="4" precision="I64" names="/Slice_1_output_0">
1764
+ <dim>2</dim>
1765
+ </port>
1766
+ </output>
1767
+ </layer>
1768
+ <layer id="100" name="/Shape_1" type="ShapeOf" version="opset3">
1769
+ <data output_type="i64" />
1770
+ <input>
1771
+ <port id="0" precision="FP32">
1772
+ <dim>1</dim>
1773
+ <dim>192</dim>
1774
+ <dim>64</dim>
1775
+ <dim>64</dim>
1776
+ </port>
1777
+ </input>
1778
+ <output>
1779
+ <port id="1" precision="I64" names="/Shape_1_output_0,/Shape_2_output_0">
1780
+ <dim>4</dim>
1781
+ </port>
1782
+ </output>
1783
+ </layer>
1784
+ <layer id="101" name="Constant_6090" type="Const" version="opset1">
1785
+ <data element_type="i64" shape="2" offset="11122220" size="16" />
1786
+ <output>
1787
+ <port id="0" precision="I64">
1788
+ <dim>2</dim>
1789
+ </port>
1790
+ </output>
1791
+ </layer>
1792
+ <layer id="102" name="Constant_6091" type="Const" version="opset1">
1793
+ <data element_type="i64" shape="" offset="2728448" size="8" />
1794
+ <output>
1795
+ <port id="0" precision="I64" />
1796
+ </output>
1797
+ </layer>
1798
+ <layer id="103" name="Gather_6092" type="Gather" version="opset8">
1799
+ <data batch_dims="0" />
1800
+ <input>
1801
+ <port id="0" precision="I64">
1802
+ <dim>4</dim>
1803
+ </port>
1804
+ <port id="1" precision="I64">
1805
+ <dim>2</dim>
1806
+ </port>
1807
+ <port id="2" precision="I64" />
1808
+ </input>
1809
+ <output>
1810
+ <port id="3" precision="I64" names="/Cast_output_0,/Concat_2_output_0">
1811
+ <dim>2</dim>
1812
+ </port>
1813
+ </output>
1814
+ </layer>
1815
+ <layer id="104" name="/Concat_3" type="Concat" version="opset1">
1816
+ <data axis="0" />
1817
+ <input>
1818
+ <port id="0" precision="I64">
1819
+ <dim>2</dim>
1820
+ </port>
1821
+ <port id="1" precision="I64">
1822
+ <dim>2</dim>
1823
+ </port>
1824
+ </input>
1825
+ <output>
1826
+ <port id="2" precision="I64" names="/Concat_3_output_0">
1827
+ <dim>4</dim>
1828
+ </port>
1829
+ </output>
1830
+ </layer>
1831
+ <layer id="105" name="Convert_884" type="Convert" version="opset1">
1832
+ <data destination_type="f32" />
1833
+ <input>
1834
+ <port id="0" precision="I64">
1835
+ <dim>4</dim>
1836
+ </port>
1837
+ </input>
1838
+ <output>
1839
+ <port id="1" precision="FP32">
1840
+ <dim>4</dim>
1841
+ </port>
1842
+ </output>
1843
+ </layer>
1844
+ <layer id="106" name="Convert_883" type="Convert" version="opset1">
1845
+ <data destination_type="f32" />
1846
+ <input>
1847
+ <port id="0" precision="I64">
1848
+ <dim>4</dim>
1849
+ </port>
1850
+ </input>
1851
+ <output>
1852
+ <port id="1" precision="FP32">
1853
+ <dim>4</dim>
1854
+ </port>
1855
+ </output>
1856
+ </layer>
1857
+ <layer id="107" name="Divide_885" type="Divide" version="opset1">
1858
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1859
+ <input>
1860
+ <port id="0" precision="FP32">
1861
+ <dim>4</dim>
1862
+ </port>
1863
+ <port id="1" precision="FP32">
1864
+ <dim>4</dim>
1865
+ </port>
1866
+ </input>
1867
+ <output>
1868
+ <port id="2" precision="FP32">
1869
+ <dim>4</dim>
1870
+ </port>
1871
+ </output>
1872
+ </layer>
1873
+ <layer id="108" name="Constant_6583" type="Const" version="opset1">
1874
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1875
+ <output>
1876
+ <port id="0" precision="FP32">
1877
+ <dim>1</dim>
1878
+ </port>
1879
+ </output>
1880
+ </layer>
1881
+ <layer id="109" name="Add_887" type="Add" version="opset1">
1882
+ <data auto_broadcast="numpy" />
1883
+ <input>
1884
+ <port id="0" precision="FP32">
1885
+ <dim>4</dim>
1886
+ </port>
1887
+ <port id="1" precision="FP32">
1888
+ <dim>1</dim>
1889
+ </port>
1890
+ </input>
1891
+ <output>
1892
+ <port id="2" precision="FP32">
1893
+ <dim>4</dim>
1894
+ </port>
1895
+ </output>
1896
+ </layer>
1897
+ <layer id="110" name="/Resize_1" type="Interpolate" version="opset4">
1898
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1899
+ <input>
1900
+ <port id="0" precision="FP32">
1901
+ <dim>1</dim>
1902
+ <dim>256</dim>
1903
+ <dim>16</dim>
1904
+ <dim>16</dim>
1905
+ </port>
1906
+ <port id="1" precision="I64">
1907
+ <dim>4</dim>
1908
+ </port>
1909
+ <port id="2" precision="FP32">
1910
+ <dim>4</dim>
1911
+ </port>
1912
+ </input>
1913
+ <output>
1914
+ <port id="3" precision="FP32" names="/Resize_1_output_0">
1915
+ <dim>1</dim>
1916
+ <dim>256</dim>
1917
+ <dim>64</dim>
1918
+ <dim>64</dim>
1919
+ </port>
1920
+ </output>
1921
+ </layer>
1922
+ <layer id="111" name="/Concat_4" type="Concat" version="opset1">
1923
+ <data axis="1" />
1924
+ <input>
1925
+ <port id="0" precision="FP32">
1926
+ <dim>1</dim>
1927
+ <dim>192</dim>
1928
+ <dim>64</dim>
1929
+ <dim>64</dim>
1930
+ </port>
1931
+ <port id="1" precision="FP32">
1932
+ <dim>1</dim>
1933
+ <dim>256</dim>
1934
+ <dim>64</dim>
1935
+ <dim>64</dim>
1936
+ </port>
1937
+ </input>
1938
+ <output>
1939
+ <port id="2" precision="FP32" names="/Concat_4_output_0">
1940
+ <dim>1</dim>
1941
+ <dim>448</dim>
1942
+ <dim>64</dim>
1943
+ <dim>64</dim>
1944
+ </port>
1945
+ </output>
1946
+ </layer>
1947
+ <layer id="112" name="onnx::Gather_308" type="Const" version="opset1">
1948
+ <data element_type="i64" shape="100" offset="11122236" size="800" />
1949
+ <output>
1950
+ <port id="0" precision="I64" names="onnx::Gather_308">
1951
+ <dim>100</dim>
1952
+ </port>
1953
+ </output>
1954
+ </layer>
1955
+ <layer id="113" name="Constant_965" type="Const" version="opset1">
1956
+ <data element_type="i64" shape="" offset="2728464" size="8" />
1957
+ <output>
1958
+ <port id="0" precision="I64" />
1959
+ </output>
1960
+ </layer>
1961
+ <layer id="114" name="/Gather_2" type="Gather" version="opset8">
1962
+ <data batch_dims="0" />
1963
+ <input>
1964
+ <port id="0" precision="FP32">
1965
+ <dim>1</dim>
1966
+ <dim>448</dim>
1967
+ <dim>64</dim>
1968
+ <dim>64</dim>
1969
+ </port>
1970
+ <port id="1" precision="I64">
1971
+ <dim>100</dim>
1972
+ </port>
1973
+ <port id="2" precision="I64" />
1974
+ </input>
1975
+ <output>
1976
+ <port id="3" precision="FP32" names="/Gather_2_output_0">
1977
+ <dim>1</dim>
1978
+ <dim>100</dim>
1979
+ <dim>64</dim>
1980
+ <dim>64</dim>
1981
+ </port>
1982
+ </output>
1983
+ </layer>
1984
+ <layer id="115" name="/anomaly_map_generator/Constant" type="Const" version="opset1">
1985
+ <data element_type="i64" shape="3" offset="11123036" size="24" />
1986
+ <output>
1987
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_output_0">
1988
+ <dim>3</dim>
1989
+ </port>
1990
+ </output>
1991
+ </layer>
1992
+ <layer id="116" name="/anomaly_map_generator/Reshape" type="Reshape" version="opset1">
1993
+ <data special_zero="true" />
1994
+ <input>
1995
+ <port id="0" precision="FP32">
1996
+ <dim>1</dim>
1997
+ <dim>100</dim>
1998
+ <dim>64</dim>
1999
+ <dim>64</dim>
2000
+ </port>
2001
+ <port id="1" precision="I64">
2002
+ <dim>3</dim>
2003
+ </port>
2004
+ </input>
2005
+ <output>
2006
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_output_0">
2007
+ <dim>1</dim>
2008
+ <dim>100</dim>
2009
+ <dim>4096</dim>
2010
+ </port>
2011
+ </output>
2012
+ </layer>
2013
+ <layer id="117" name="Constant_6584" type="Const" version="opset1">
2014
+ <data element_type="f32" shape="1, 100, 4096" offset="11123060" size="1638400" />
2015
+ <output>
2016
+ <port id="0" precision="FP32">
2017
+ <dim>1</dim>
2018
+ <dim>100</dim>
2019
+ <dim>4096</dim>
2020
+ </port>
2021
+ </output>
2022
+ </layer>
2023
+ <layer id="118" name="/anomaly_map_generator/Sub" type="Subtract" version="opset1">
2024
+ <data auto_broadcast="numpy" />
2025
+ <input>
2026
+ <port id="0" precision="FP32">
2027
+ <dim>1</dim>
2028
+ <dim>100</dim>
2029
+ <dim>4096</dim>
2030
+ </port>
2031
+ <port id="1" precision="FP32">
2032
+ <dim>1</dim>
2033
+ <dim>100</dim>
2034
+ <dim>4096</dim>
2035
+ </port>
2036
+ </input>
2037
+ <output>
2038
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Sub_output_0">
2039
+ <dim>1</dim>
2040
+ <dim>100</dim>
2041
+ <dim>4096</dim>
2042
+ </port>
2043
+ </output>
2044
+ </layer>
2045
+ <layer id="119" name="Constant_975" type="Const" version="opset1">
2046
+ <data element_type="i64" shape="3" offset="12761460" size="24" />
2047
+ <output>
2048
+ <port id="0" precision="I64">
2049
+ <dim>3</dim>
2050
+ </port>
2051
+ </output>
2052
+ </layer>
2053
+ <layer id="120" name="/anomaly_map_generator/Transpose" type="Transpose" version="opset1">
2054
+ <input>
2055
+ <port id="0" precision="FP32">
2056
+ <dim>1</dim>
2057
+ <dim>100</dim>
2058
+ <dim>4096</dim>
2059
+ </port>
2060
+ <port id="1" precision="I64">
2061
+ <dim>3</dim>
2062
+ </port>
2063
+ </input>
2064
+ <output>
2065
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Transpose_output_0">
2066
+ <dim>4096</dim>
2067
+ <dim>1</dim>
2068
+ <dim>100</dim>
2069
+ </port>
2070
+ </output>
2071
+ </layer>
2072
+ <layer id="121" name="onnx::MatMul_310" type="Const" version="opset1">
2073
+ <data element_type="f32" shape="4096, 100, 100" offset="12761484" size="163840000" />
2074
+ <output>
2075
+ <port id="0" precision="FP32" names="onnx::MatMul_310">
2076
+ <dim>4096</dim>
2077
+ <dim>100</dim>
2078
+ <dim>100</dim>
2079
+ </port>
2080
+ </output>
2081
+ </layer>
2082
+ <layer id="122" name="/anomaly_map_generator/MatMul" type="MatMul" version="opset1">
2083
+ <data transpose_a="false" transpose_b="false" />
2084
+ <input>
2085
+ <port id="0" precision="FP32">
2086
+ <dim>4096</dim>
2087
+ <dim>1</dim>
2088
+ <dim>100</dim>
2089
+ </port>
2090
+ <port id="1" precision="FP32">
2091
+ <dim>4096</dim>
2092
+ <dim>100</dim>
2093
+ <dim>100</dim>
2094
+ </port>
2095
+ </input>
2096
+ <output>
2097
+ <port id="2" precision="FP32" names="/anomaly_map_generator/MatMul_output_0">
2098
+ <dim>4096</dim>
2099
+ <dim>1</dim>
2100
+ <dim>100</dim>
2101
+ </port>
2102
+ </output>
2103
+ </layer>
2104
+ <layer id="123" name="/anomaly_map_generator/Mul" type="Multiply" version="opset1">
2105
+ <data auto_broadcast="numpy" />
2106
+ <input>
2107
+ <port id="0" precision="FP32">
2108
+ <dim>4096</dim>
2109
+ <dim>1</dim>
2110
+ <dim>100</dim>
2111
+ </port>
2112
+ <port id="1" precision="FP32">
2113
+ <dim>4096</dim>
2114
+ <dim>1</dim>
2115
+ <dim>100</dim>
2116
+ </port>
2117
+ </input>
2118
+ <output>
2119
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Mul_output_0">
2120
+ <dim>4096</dim>
2121
+ <dim>1</dim>
2122
+ <dim>100</dim>
2123
+ </port>
2124
+ </output>
2125
+ </layer>
2126
+ <layer id="124" name="Constant_979" type="Const" version="opset1">
2127
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2128
+ <output>
2129
+ <port id="0" precision="I64">
2130
+ <dim>1</dim>
2131
+ </port>
2132
+ </output>
2133
+ </layer>
2134
+ <layer id="125" name="/anomaly_map_generator/ReduceSum" type="ReduceSum" version="opset1">
2135
+ <data keep_dims="false" />
2136
+ <input>
2137
+ <port id="0" precision="FP32">
2138
+ <dim>4096</dim>
2139
+ <dim>1</dim>
2140
+ <dim>100</dim>
2141
+ </port>
2142
+ <port id="1" precision="I64">
2143
+ <dim>1</dim>
2144
+ </port>
2145
+ </input>
2146
+ <output>
2147
+ <port id="2" precision="FP32" names="/anomaly_map_generator/ReduceSum_output_0">
2148
+ <dim>4096</dim>
2149
+ <dim>1</dim>
2150
+ </port>
2151
+ </output>
2152
+ </layer>
2153
+ <layer id="126" name="/anomaly_map_generator/Constant_1" type="Const" version="opset1">
2154
+ <data element_type="i64" shape="4" offset="176601484" size="32" />
2155
+ <output>
2156
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_1_output_0">
2157
+ <dim>4</dim>
2158
+ </port>
2159
+ </output>
2160
+ </layer>
2161
+ <layer id="127" name="/anomaly_map_generator/Reshape_1" type="Reshape" version="opset1">
2162
+ <data special_zero="true" />
2163
+ <input>
2164
+ <port id="0" precision="FP32">
2165
+ <dim>4096</dim>
2166
+ <dim>1</dim>
2167
+ </port>
2168
+ <port id="1" precision="I64">
2169
+ <dim>4</dim>
2170
+ </port>
2171
+ </input>
2172
+ <output>
2173
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_1_output_0">
2174
+ <dim>1</dim>
2175
+ <dim>1</dim>
2176
+ <dim>64</dim>
2177
+ <dim>64</dim>
2178
+ </port>
2179
+ </output>
2180
+ </layer>
2181
+ <layer id="128" name="/anomaly_map_generator/Clip" type="Clamp" version="opset1">
2182
+ <data min="0" max="3.4028234663852886e+38" />
2183
+ <input>
2184
+ <port id="0" precision="FP32">
2185
+ <dim>1</dim>
2186
+ <dim>1</dim>
2187
+ <dim>64</dim>
2188
+ <dim>64</dim>
2189
+ </port>
2190
+ </input>
2191
+ <output>
2192
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Clip_output_0">
2193
+ <dim>1</dim>
2194
+ <dim>1</dim>
2195
+ <dim>64</dim>
2196
+ <dim>64</dim>
2197
+ </port>
2198
+ </output>
2199
+ </layer>
2200
+ <layer id="129" name="/anomaly_map_generator/Sqrt" type="Sqrt" version="opset1">
2201
+ <input>
2202
+ <port id="0" precision="FP32">
2203
+ <dim>1</dim>
2204
+ <dim>1</dim>
2205
+ <dim>64</dim>
2206
+ <dim>64</dim>
2207
+ </port>
2208
+ </input>
2209
+ <output>
2210
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Sqrt_output_0">
2211
+ <dim>1</dim>
2212
+ <dim>1</dim>
2213
+ <dim>64</dim>
2214
+ <dim>64</dim>
2215
+ </port>
2216
+ </output>
2217
+ </layer>
2218
+ <layer id="130" name="/anomaly_map_generator/Shape" type="ShapeOf" version="opset3">
2219
+ <data output_type="i64" />
2220
+ <input>
2221
+ <port id="0" precision="FP32">
2222
+ <dim>1</dim>
2223
+ <dim>1</dim>
2224
+ <dim>64</dim>
2225
+ <dim>64</dim>
2226
+ </port>
2227
+ </input>
2228
+ <output>
2229
+ <port id="1" precision="I64" names="/anomaly_map_generator/Shape_output_0">
2230
+ <dim>4</dim>
2231
+ </port>
2232
+ </output>
2233
+ </layer>
2234
+ <layer id="131" name="/anomaly_map_generator/Constant_4" type="Const" version="opset1">
2235
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
2236
+ <output>
2237
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_4_output_0">
2238
+ <dim>1</dim>
2239
+ </port>
2240
+ </output>
2241
+ </layer>
2242
+ <layer id="132" name="/anomaly_map_generator/Constant_5" type="Const" version="opset1">
2243
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2244
+ <output>
2245
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_5_output_0">
2246
+ <dim>1</dim>
2247
+ </port>
2248
+ </output>
2249
+ </layer>
2250
+ <layer id="133" name="Broadcast_1002" type="Const" version="opset1">
2251
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
2252
+ <output>
2253
+ <port id="0" precision="I64">
2254
+ <dim>1</dim>
2255
+ </port>
2256
+ </output>
2257
+ </layer>
2258
+ <layer id="134" name="/anomaly_map_generator/Slice" type="StridedSlice" version="opset1">
2259
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
2260
+ <input>
2261
+ <port id="0" precision="I64">
2262
+ <dim>4</dim>
2263
+ </port>
2264
+ <port id="1" precision="I64">
2265
+ <dim>1</dim>
2266
+ </port>
2267
+ <port id="2" precision="I64">
2268
+ <dim>1</dim>
2269
+ </port>
2270
+ <port id="3" precision="I64">
2271
+ <dim>1</dim>
2272
+ </port>
2273
+ </input>
2274
+ <output>
2275
+ <port id="4" precision="I64" names="/anomaly_map_generator/Slice_output_0">
2276
+ <dim>2</dim>
2277
+ </port>
2278
+ </output>
2279
+ </layer>
2280
+ <layer id="135" name="/anomaly_map_generator/Constant_6" type="Const" version="opset1">
2281
+ <data element_type="i64" shape="2" offset="176601516" size="16" />
2282
+ <output>
2283
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_6_output_0">
2284
+ <dim>2</dim>
2285
+ </port>
2286
+ </output>
2287
+ </layer>
2288
+ <layer id="136" name="/anomaly_map_generator/Concat" type="Concat" version="opset1">
2289
+ <data axis="0" />
2290
+ <input>
2291
+ <port id="0" precision="I64">
2292
+ <dim>2</dim>
2293
+ </port>
2294
+ <port id="1" precision="I64">
2295
+ <dim>2</dim>
2296
+ </port>
2297
+ </input>
2298
+ <output>
2299
+ <port id="2" precision="I64" names="/anomaly_map_generator/Concat_output_0">
2300
+ <dim>4</dim>
2301
+ </port>
2302
+ </output>
2303
+ </layer>
2304
+ <layer id="137" name="Convert_1025" type="Convert" version="opset1">
2305
+ <data destination_type="f32" />
2306
+ <input>
2307
+ <port id="0" precision="I64">
2308
+ <dim>4</dim>
2309
+ </port>
2310
+ </input>
2311
+ <output>
2312
+ <port id="1" precision="FP32">
2313
+ <dim>4</dim>
2314
+ </port>
2315
+ </output>
2316
+ </layer>
2317
+ <layer id="138" name="Convert_1024" type="Convert" version="opset1">
2318
+ <data destination_type="f32" />
2319
+ <input>
2320
+ <port id="0" precision="I64">
2321
+ <dim>4</dim>
2322
+ </port>
2323
+ </input>
2324
+ <output>
2325
+ <port id="1" precision="FP32">
2326
+ <dim>4</dim>
2327
+ </port>
2328
+ </output>
2329
+ </layer>
2330
+ <layer id="139" name="Divide_1026" type="Divide" version="opset1">
2331
+ <data auto_broadcast="numpy" m_pythondiv="true" />
2332
+ <input>
2333
+ <port id="0" precision="FP32">
2334
+ <dim>4</dim>
2335
+ </port>
2336
+ <port id="1" precision="FP32">
2337
+ <dim>4</dim>
2338
+ </port>
2339
+ </input>
2340
+ <output>
2341
+ <port id="2" precision="FP32">
2342
+ <dim>4</dim>
2343
+ </port>
2344
+ </output>
2345
+ </layer>
2346
+ <layer id="140" name="Constant_6585" type="Const" version="opset1">
2347
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
2348
+ <output>
2349
+ <port id="0" precision="FP32">
2350
+ <dim>1</dim>
2351
+ </port>
2352
+ </output>
2353
+ </layer>
2354
+ <layer id="141" name="Add_1028" type="Add" version="opset1">
2355
+ <data auto_broadcast="numpy" />
2356
+ <input>
2357
+ <port id="0" precision="FP32">
2358
+ <dim>4</dim>
2359
+ </port>
2360
+ <port id="1" precision="FP32">
2361
+ <dim>1</dim>
2362
+ </port>
2363
+ </input>
2364
+ <output>
2365
+ <port id="2" precision="FP32">
2366
+ <dim>4</dim>
2367
+ </port>
2368
+ </output>
2369
+ </layer>
2370
+ <layer id="142" name="/anomaly_map_generator/Resize" type="Interpolate" version="opset4">
2371
+ <data mode="linear_onnx" shape_calculation_mode="sizes" coordinate_transformation_mode="half_pixel" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
2372
+ <input>
2373
+ <port id="0" precision="FP32">
2374
+ <dim>1</dim>
2375
+ <dim>1</dim>
2376
+ <dim>64</dim>
2377
+ <dim>64</dim>
2378
+ </port>
2379
+ <port id="1" precision="I64">
2380
+ <dim>4</dim>
2381
+ </port>
2382
+ <port id="2" precision="FP32">
2383
+ <dim>4</dim>
2384
+ </port>
2385
+ </input>
2386
+ <output>
2387
+ <port id="3" precision="FP32" names="/anomaly_map_generator/Resize_output_0">
2388
+ <dim>1</dim>
2389
+ <dim>1</dim>
2390
+ <dim>256</dim>
2391
+ <dim>256</dim>
2392
+ </port>
2393
+ </output>
2394
+ </layer>
2395
+ <layer id="143" name="Split_1097.0" type="Const" version="opset1">
2396
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2397
+ <output>
2398
+ <port id="0" precision="I64">
2399
+ <dim>4</dim>
2400
+ </port>
2401
+ </output>
2402
+ </layer>
2403
+ <layer id="144" name="Split_1097.1" type="Const" version="opset1">
2404
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2405
+ <output>
2406
+ <port id="0" precision="I64">
2407
+ <dim>4</dim>
2408
+ </port>
2409
+ </output>
2410
+ </layer>
2411
+ <layer id="145" name="Constant_1095" type="Const" version="opset1">
2412
+ <data element_type="f32" shape="" offset="176601564" size="4" />
2413
+ <output>
2414
+ <port id="0" precision="FP32" />
2415
+ </output>
2416
+ </layer>
2417
+ <layer id="146" name="/anomaly_map_generator/blur/Pad" type="Pad" version="opset1">
2418
+ <data pad_mode="reflect" />
2419
+ <input>
2420
+ <port id="0" precision="FP32">
2421
+ <dim>1</dim>
2422
+ <dim>1</dim>
2423
+ <dim>256</dim>
2424
+ <dim>256</dim>
2425
+ </port>
2426
+ <port id="1" precision="I64">
2427
+ <dim>4</dim>
2428
+ </port>
2429
+ <port id="2" precision="I64">
2430
+ <dim>4</dim>
2431
+ </port>
2432
+ <port id="3" precision="FP32" />
2433
+ </input>
2434
+ <output>
2435
+ <port id="4" precision="FP32" names="/anomaly_map_generator/blur/Pad_output_0">
2436
+ <dim>1</dim>
2437
+ <dim>1</dim>
2438
+ <dim>288</dim>
2439
+ <dim>288</dim>
2440
+ </port>
2441
+ </output>
2442
+ </layer>
2443
+ <layer id="147" name="anomaly_map_generator.blur.kernel" type="Const" version="opset1">
2444
+ <data element_type="f32" shape="1, 1, 33, 33" offset="176601568" size="4356" />
2445
+ <output>
2446
+ <port id="0" precision="FP32" names="anomaly_map_generator.blur.kernel">
2447
+ <dim>1</dim>
2448
+ <dim>1</dim>
2449
+ <dim>33</dim>
2450
+ <dim>33</dim>
2451
+ </port>
2452
+ </output>
2453
+ </layer>
2454
+ <layer id="148" name="/anomaly_map_generator/blur/Conv" type="Convolution" version="opset1">
2455
+ <data strides="1, 1" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
2456
+ <input>
2457
+ <port id="0" precision="FP32">
2458
+ <dim>1</dim>
2459
+ <dim>1</dim>
2460
+ <dim>288</dim>
2461
+ <dim>288</dim>
2462
+ </port>
2463
+ <port id="1" precision="FP32">
2464
+ <dim>1</dim>
2465
+ <dim>1</dim>
2466
+ <dim>33</dim>
2467
+ <dim>33</dim>
2468
+ </port>
2469
+ </input>
2470
+ <output>
2471
+ <port id="2" precision="FP32" names="/anomaly_map_generator/blur/Conv_output_0">
2472
+ <dim>1</dim>
2473
+ <dim>1</dim>
2474
+ <dim>256</dim>
2475
+ <dim>256</dim>
2476
+ </port>
2477
+ </output>
2478
+ </layer>
2479
+ <layer id="149" name="/anomaly_map_generator/blur/Constant_8" type="Const" version="opset1">
2480
+ <data element_type="i64" shape="4" offset="176605924" size="32" />
2481
+ <output>
2482
+ <port id="0" precision="I64" names="/anomaly_map_generator/blur/Constant_8_output_0">
2483
+ <dim>4</dim>
2484
+ </port>
2485
+ </output>
2486
+ </layer>
2487
+ <layer id="150" name="output" type="Reshape" version="opset1">
2488
+ <data special_zero="true" />
2489
+ <input>
2490
+ <port id="0" precision="FP32">
2491
+ <dim>1</dim>
2492
+ <dim>1</dim>
2493
+ <dim>256</dim>
2494
+ <dim>256</dim>
2495
+ </port>
2496
+ <port id="1" precision="I64">
2497
+ <dim>4</dim>
2498
+ </port>
2499
+ </input>
2500
+ <output>
2501
+ <port id="2" precision="FP32" names="output">
2502
+ <dim>1</dim>
2503
+ <dim>1</dim>
2504
+ <dim>256</dim>
2505
+ <dim>256</dim>
2506
+ </port>
2507
+ </output>
2508
+ </layer>
2509
+ <layer id="151" name="output/sink_port_0" type="Result" version="opset1">
2510
+ <input>
2511
+ <port id="0" precision="FP32">
2512
+ <dim>1</dim>
2513
+ <dim>1</dim>
2514
+ <dim>256</dim>
2515
+ <dim>256</dim>
2516
+ </port>
2517
+ </input>
2518
+ </layer>
2519
+ </layers>
2520
+ <edges>
2521
+ <edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
2522
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="1" />
2523
+ <edge from-layer="2" from-port="2" to-layer="4" to-port="0" />
2524
+ <edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
2525
+ <edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
2526
+ <edge from-layer="5" from-port="1" to-layer="6" to-port="0" />
2527
+ <edge from-layer="6" from-port="1" to-layer="8" to-port="0" />
2528
+ <edge from-layer="6" from-port="1" to-layer="16" to-port="1" />
2529
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="1" />
2530
+ <edge from-layer="8" from-port="2" to-layer="10" to-port="0" />
2531
+ <edge from-layer="9" from-port="0" to-layer="10" to-port="1" />
2532
+ <edge from-layer="10" from-port="2" to-layer="11" to-port="0" />
2533
+ <edge from-layer="11" from-port="1" to-layer="13" to-port="0" />
2534
+ <edge from-layer="12" from-port="0" to-layer="13" to-port="1" />
2535
+ <edge from-layer="13" from-port="2" to-layer="15" to-port="0" />
2536
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="1" />
2537
+ <edge from-layer="15" from-port="2" to-layer="16" to-port="0" />
2538
+ <edge from-layer="16" from-port="2" to-layer="17" to-port="0" />
2539
+ <edge from-layer="17" from-port="1" to-layer="19" to-port="0" />
2540
+ <edge from-layer="17" from-port="1" to-layer="27" to-port="1" />
2541
+ <edge from-layer="18" from-port="0" to-layer="19" to-port="1" />
2542
+ <edge from-layer="19" from-port="2" to-layer="21" to-port="0" />
2543
+ <edge from-layer="20" from-port="0" to-layer="21" to-port="1" />
2544
+ <edge from-layer="21" from-port="2" to-layer="22" to-port="0" />
2545
+ <edge from-layer="22" from-port="1" to-layer="24" to-port="0" />
2546
+ <edge from-layer="23" from-port="0" to-layer="24" to-port="1" />
2547
+ <edge from-layer="24" from-port="2" to-layer="26" to-port="0" />
2548
+ <edge from-layer="25" from-port="0" to-layer="26" to-port="1" />
2549
+ <edge from-layer="26" from-port="2" to-layer="27" to-port="0" />
2550
+ <edge from-layer="27" from-port="2" to-layer="28" to-port="0" />
2551
+ <edge from-layer="28" from-port="1" to-layer="30" to-port="0" />
2552
+ <edge from-layer="28" from-port="1" to-layer="39" to-port="0" />
2553
+ <edge from-layer="28" from-port="1" to-layer="68" to-port="0" />
2554
+ <edge from-layer="29" from-port="0" to-layer="30" to-port="1" />
2555
+ <edge from-layer="30" from-port="2" to-layer="32" to-port="0" />
2556
+ <edge from-layer="31" from-port="0" to-layer="32" to-port="1" />
2557
+ <edge from-layer="32" from-port="2" to-layer="33" to-port="0" />
2558
+ <edge from-layer="33" from-port="1" to-layer="35" to-port="0" />
2559
+ <edge from-layer="34" from-port="0" to-layer="35" to-port="1" />
2560
+ <edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
2561
+ <edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
2562
+ <edge from-layer="37" from-port="2" to-layer="42" to-port="0" />
2563
+ <edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
2564
+ <edge from-layer="39" from-port="2" to-layer="41" to-port="0" />
2565
+ <edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
2566
+ <edge from-layer="41" from-port="2" to-layer="42" to-port="1" />
2567
+ <edge from-layer="42" from-port="2" to-layer="43" to-port="0" />
2568
+ <edge from-layer="43" from-port="1" to-layer="45" to-port="0" />
2569
+ <edge from-layer="43" from-port="1" to-layer="53" to-port="1" />
2570
+ <edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
2571
+ <edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
2572
+ <edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
2573
+ <edge from-layer="47" from-port="2" to-layer="48" to-port="0" />
2574
+ <edge from-layer="48" from-port="1" to-layer="50" to-port="0" />
2575
+ <edge from-layer="49" from-port="0" to-layer="50" to-port="1" />
2576
+ <edge from-layer="50" from-port="2" to-layer="52" to-port="0" />
2577
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="1" />
2578
+ <edge from-layer="52" from-port="2" to-layer="53" to-port="0" />
2579
+ <edge from-layer="53" from-port="2" to-layer="54" to-port="0" />
2580
+ <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
2581
+ <edge from-layer="54" from-port="1" to-layer="67" to-port="0" />
2582
+ <edge from-layer="54" from-port="1" to-layer="79" to-port="0" />
2583
+ <edge from-layer="54" from-port="1" to-layer="70" to-port="0" />
2584
+ <edge from-layer="55" from-port="1" to-layer="59" to-port="0" />
2585
+ <edge from-layer="55" from-port="1" to-layer="63" to-port="0" />
2586
+ <edge from-layer="56" from-port="0" to-layer="59" to-port="1" />
2587
+ <edge from-layer="57" from-port="0" to-layer="59" to-port="2" />
2588
+ <edge from-layer="58" from-port="0" to-layer="59" to-port="3" />
2589
+ <edge from-layer="59" from-port="4" to-layer="61" to-port="0" />
2590
+ <edge from-layer="60" from-port="0" to-layer="61" to-port="1" />
2591
+ <edge from-layer="61" from-port="2" to-layer="62" to-port="0" />
2592
+ <edge from-layer="61" from-port="2" to-layer="67" to-port="1" />
2593
+ <edge from-layer="62" from-port="1" to-layer="64" to-port="0" />
2594
+ <edge from-layer="63" from-port="1" to-layer="64" to-port="1" />
2595
+ <edge from-layer="64" from-port="2" to-layer="66" to-port="0" />
2596
+ <edge from-layer="65" from-port="0" to-layer="66" to-port="1" />
2597
+ <edge from-layer="66" from-port="2" to-layer="67" to-port="2" />
2598
+ <edge from-layer="67" from-port="3" to-layer="68" to-port="1" />
2599
+ <edge from-layer="68" from-port="2" to-layer="111" to-port="0" />
2600
+ <edge from-layer="68" from-port="2" to-layer="100" to-port="0" />
2601
+ <edge from-layer="69" from-port="0" to-layer="70" to-port="1" />
2602
+ <edge from-layer="70" from-port="2" to-layer="72" to-port="0" />
2603
+ <edge from-layer="71" from-port="0" to-layer="72" to-port="1" />
2604
+ <edge from-layer="72" from-port="2" to-layer="73" to-port="0" />
2605
+ <edge from-layer="73" from-port="1" to-layer="75" to-port="0" />
2606
+ <edge from-layer="74" from-port="0" to-layer="75" to-port="1" />
2607
+ <edge from-layer="75" from-port="2" to-layer="77" to-port="0" />
2608
+ <edge from-layer="76" from-port="0" to-layer="77" to-port="1" />
2609
+ <edge from-layer="77" from-port="2" to-layer="82" to-port="0" />
2610
+ <edge from-layer="78" from-port="0" to-layer="79" to-port="1" />
2611
+ <edge from-layer="79" from-port="2" to-layer="81" to-port="0" />
2612
+ <edge from-layer="80" from-port="0" to-layer="81" to-port="1" />
2613
+ <edge from-layer="81" from-port="2" to-layer="82" to-port="1" />
2614
+ <edge from-layer="82" from-port="2" to-layer="83" to-port="0" />
2615
+ <edge from-layer="83" from-port="1" to-layer="85" to-port="0" />
2616
+ <edge from-layer="83" from-port="1" to-layer="93" to-port="1" />
2617
+ <edge from-layer="84" from-port="0" to-layer="85" to-port="1" />
2618
+ <edge from-layer="85" from-port="2" to-layer="87" to-port="0" />
2619
+ <edge from-layer="86" from-port="0" to-layer="87" to-port="1" />
2620
+ <edge from-layer="87" from-port="2" to-layer="88" to-port="0" />
2621
+ <edge from-layer="88" from-port="1" to-layer="90" to-port="0" />
2622
+ <edge from-layer="89" from-port="0" to-layer="90" to-port="1" />
2623
+ <edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
2624
+ <edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
2625
+ <edge from-layer="92" from-port="2" to-layer="93" to-port="0" />
2626
+ <edge from-layer="93" from-port="2" to-layer="94" to-port="0" />
2627
+ <edge from-layer="94" from-port="1" to-layer="95" to-port="0" />
2628
+ <edge from-layer="94" from-port="1" to-layer="110" to-port="0" />
2629
+ <edge from-layer="95" from-port="1" to-layer="99" to-port="0" />
2630
+ <edge from-layer="95" from-port="1" to-layer="106" to-port="0" />
2631
+ <edge from-layer="96" from-port="0" to-layer="99" to-port="1" />
2632
+ <edge from-layer="97" from-port="0" to-layer="99" to-port="2" />
2633
+ <edge from-layer="98" from-port="0" to-layer="99" to-port="3" />
2634
+ <edge from-layer="99" from-port="4" to-layer="104" to-port="0" />
2635
+ <edge from-layer="100" from-port="1" to-layer="103" to-port="0" />
2636
+ <edge from-layer="101" from-port="0" to-layer="103" to-port="1" />
2637
+ <edge from-layer="102" from-port="0" to-layer="103" to-port="2" />
2638
+ <edge from-layer="103" from-port="3" to-layer="104" to-port="1" />
2639
+ <edge from-layer="104" from-port="2" to-layer="105" to-port="0" />
2640
+ <edge from-layer="104" from-port="2" to-layer="110" to-port="1" />
2641
+ <edge from-layer="105" from-port="1" to-layer="107" to-port="0" />
2642
+ <edge from-layer="106" from-port="1" to-layer="107" to-port="1" />
2643
+ <edge from-layer="107" from-port="2" to-layer="109" to-port="0" />
2644
+ <edge from-layer="108" from-port="0" to-layer="109" to-port="1" />
2645
+ <edge from-layer="109" from-port="2" to-layer="110" to-port="2" />
2646
+ <edge from-layer="110" from-port="3" to-layer="111" to-port="1" />
2647
+ <edge from-layer="111" from-port="2" to-layer="114" to-port="0" />
2648
+ <edge from-layer="112" from-port="0" to-layer="114" to-port="1" />
2649
+ <edge from-layer="113" from-port="0" to-layer="114" to-port="2" />
2650
+ <edge from-layer="114" from-port="3" to-layer="116" to-port="0" />
2651
+ <edge from-layer="115" from-port="0" to-layer="116" to-port="1" />
2652
+ <edge from-layer="116" from-port="2" to-layer="118" to-port="0" />
2653
+ <edge from-layer="117" from-port="0" to-layer="118" to-port="1" />
2654
+ <edge from-layer="118" from-port="2" to-layer="120" to-port="0" />
2655
+ <edge from-layer="119" from-port="0" to-layer="120" to-port="1" />
2656
+ <edge from-layer="120" from-port="2" to-layer="122" to-port="0" />
2657
+ <edge from-layer="120" from-port="2" to-layer="123" to-port="1" />
2658
+ <edge from-layer="121" from-port="0" to-layer="122" to-port="1" />
2659
+ <edge from-layer="122" from-port="2" to-layer="123" to-port="0" />
2660
+ <edge from-layer="123" from-port="2" to-layer="125" to-port="0" />
2661
+ <edge from-layer="124" from-port="0" to-layer="125" to-port="1" />
2662
+ <edge from-layer="125" from-port="2" to-layer="127" to-port="0" />
2663
+ <edge from-layer="126" from-port="0" to-layer="127" to-port="1" />
2664
+ <edge from-layer="127" from-port="2" to-layer="128" to-port="0" />
2665
+ <edge from-layer="128" from-port="1" to-layer="129" to-port="0" />
2666
+ <edge from-layer="129" from-port="1" to-layer="130" to-port="0" />
2667
+ <edge from-layer="129" from-port="1" to-layer="142" to-port="0" />
2668
+ <edge from-layer="130" from-port="1" to-layer="134" to-port="0" />
2669
+ <edge from-layer="130" from-port="1" to-layer="138" to-port="0" />
2670
+ <edge from-layer="131" from-port="0" to-layer="134" to-port="1" />
2671
+ <edge from-layer="132" from-port="0" to-layer="134" to-port="2" />
2672
+ <edge from-layer="133" from-port="0" to-layer="134" to-port="3" />
2673
+ <edge from-layer="134" from-port="4" to-layer="136" to-port="0" />
2674
+ <edge from-layer="135" from-port="0" to-layer="136" to-port="1" />
2675
+ <edge from-layer="136" from-port="2" to-layer="137" to-port="0" />
2676
+ <edge from-layer="136" from-port="2" to-layer="142" to-port="1" />
2677
+ <edge from-layer="137" from-port="1" to-layer="139" to-port="0" />
2678
+ <edge from-layer="138" from-port="1" to-layer="139" to-port="1" />
2679
+ <edge from-layer="139" from-port="2" to-layer="141" to-port="0" />
2680
+ <edge from-layer="140" from-port="0" to-layer="141" to-port="1" />
2681
+ <edge from-layer="141" from-port="2" to-layer="142" to-port="2" />
2682
+ <edge from-layer="142" from-port="3" to-layer="146" to-port="0" />
2683
+ <edge from-layer="143" from-port="0" to-layer="146" to-port="1" />
2684
+ <edge from-layer="144" from-port="0" to-layer="146" to-port="2" />
2685
+ <edge from-layer="145" from-port="0" to-layer="146" to-port="3" />
2686
+ <edge from-layer="146" from-port="4" to-layer="148" to-port="0" />
2687
+ <edge from-layer="147" from-port="0" to-layer="148" to-port="1" />
2688
+ <edge from-layer="148" from-port="2" to-layer="150" to-port="0" />
2689
+ <edge from-layer="149" from-port="0" to-layer="150" to-port="1" />
2690
+ <edge from-layer="150" from-port="2" to-layer="151" to-port="0" />
2691
+ </edges>
2692
+ <rt_info>
2693
+ <MO_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2694
+ <Runtime_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2695
+ <conversion_parameters>
2696
+ <framework value="onnx" />
2697
+ <input_model value="DIR\model.onnx" />
2698
+ <model_name value="model" />
2699
+ <output_dir value="DIR" />
2700
+ </conversion_parameters>
2701
+ <legacy_frontend value="False" />
2702
+ </rt_info>
2703
+ </net>
grid/examples/001.png ADDED
grid/examples/005.png ADDED
grid/examples/006.png ADDED
grid/examples/007.png ADDED
grid/examples/009.png ADDED
grid/examples/010.png ADDED
grid/run/weights/lightning/model.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3d201e840f685a2d077284582466865bb9caca2183207ab4e00d2e584a28802
3
+ size 189222268
grid/run/weights/openvino/metadata.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "task": "segmentation",
3
+ "transform": {
4
+ "__version__": "1.3.0",
5
+ "transform": {
6
+ "__class_fullname__": "Compose",
7
+ "p": 1.0,
8
+ "transforms": [
9
+ {
10
+ "__class_fullname__": "Resize",
11
+ "always_apply": true,
12
+ "p": 1,
13
+ "height": 256,
14
+ "width": 256,
15
+ "interpolation": 1
16
+ },
17
+ {
18
+ "__class_fullname__": "Normalize",
19
+ "always_apply": false,
20
+ "p": 1.0,
21
+ "mean": [
22
+ 0.485,
23
+ 0.456,
24
+ 0.406
25
+ ],
26
+ "std": [
27
+ 0.229,
28
+ 0.224,
29
+ 0.225
30
+ ],
31
+ "max_pixel_value": 255.0
32
+ },
33
+ {
34
+ "__class_fullname__": "ToTensorV2",
35
+ "always_apply": true,
36
+ "p": 1.0,
37
+ "transpose_mask": false
38
+ }
39
+ ],
40
+ "bbox_params": null,
41
+ "keypoint_params": null,
42
+ "additional_targets": {
43
+ "image": "image",
44
+ "depth_image": "image"
45
+ }
46
+ }
47
+ },
48
+ "image_threshold": 9.912899017333984,
49
+ "pixel_threshold": 9.94576358795166,
50
+ "min": 5.021185874938965,
51
+ "max": 17.323503494262695
52
+ }
grid/run/weights/openvino/model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04946b5123f6ead62260ec2fdd8396e3f4feba09a5d9e54aec2de5be1ad8c781
3
+ size 176605956
grid/run/weights/openvino/model.mapping ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <mapping>
3
+ <map>
4
+ <framework name="input" output_port_id="input" />
5
+ <IR name="input" output_port_id="0" />
6
+ </map>
7
+ <map>
8
+ <framework name="onnx::Conv_260" output_port_id="onnx::Conv_260" />
9
+ <IR name="onnx::Conv_260" output_port_id="0" />
10
+ </map>
11
+ <map>
12
+ <framework name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/conv1/Conv_output_0" />
13
+ <IR name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="2" />
14
+ </map>
15
+ <map>
16
+ <framework name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="/feature_extractor/feature_extractor/act1/Relu_output_0" />
17
+ <IR name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="1" />
18
+ </map>
19
+ <map>
20
+ <framework name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0" />
21
+ <IR name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="1" />
22
+ </map>
23
+ <map>
24
+ <framework name="onnx::Conv_263" output_port_id="onnx::Conv_263" />
25
+ <IR name="onnx::Conv_263" output_port_id="0" />
26
+ </map>
27
+ <map>
28
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0" />
29
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="2" />
30
+ </map>
31
+ <map>
32
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0" />
33
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="1" />
34
+ </map>
35
+ <map>
36
+ <framework name="onnx::Conv_266" output_port_id="onnx::Conv_266" />
37
+ <IR name="onnx::Conv_266" output_port_id="0" />
38
+ </map>
39
+ <map>
40
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0" />
41
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="2" />
42
+ </map>
43
+ <map>
44
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0" />
45
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="2" />
46
+ </map>
47
+ <map>
48
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0" />
49
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="1" />
50
+ </map>
51
+ <map>
52
+ <framework name="onnx::Conv_269" output_port_id="onnx::Conv_269" />
53
+ <IR name="onnx::Conv_269" output_port_id="0" />
54
+ </map>
55
+ <map>
56
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0" />
57
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="2" />
58
+ </map>
59
+ <map>
60
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0" />
61
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="1" />
62
+ </map>
63
+ <map>
64
+ <framework name="onnx::Conv_272" output_port_id="onnx::Conv_272" />
65
+ <IR name="onnx::Conv_272" output_port_id="0" />
66
+ </map>
67
+ <map>
68
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0" />
69
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="2" />
70
+ </map>
71
+ <map>
72
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0" />
73
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="2" />
74
+ </map>
75
+ <map>
76
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0" />
77
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="1" />
78
+ </map>
79
+ <map>
80
+ <framework name="onnx::Conv_275" output_port_id="onnx::Conv_275" />
81
+ <IR name="onnx::Conv_275" output_port_id="0" />
82
+ </map>
83
+ <map>
84
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0" />
85
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="2" />
86
+ </map>
87
+ <map>
88
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0" />
89
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="1" />
90
+ </map>
91
+ <map>
92
+ <framework name="onnx::Conv_278" output_port_id="onnx::Conv_278" />
93
+ <IR name="onnx::Conv_278" output_port_id="0" />
94
+ </map>
95
+ <map>
96
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0" />
97
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="2" />
98
+ </map>
99
+ <map>
100
+ <framework name="onnx::Conv_281" output_port_id="onnx::Conv_281" />
101
+ <IR name="onnx::Conv_281" output_port_id="0" />
102
+ </map>
103
+ <map>
104
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0" />
105
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="2" />
106
+ </map>
107
+ <map>
108
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0" />
109
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="2" />
110
+ </map>
111
+ <map>
112
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0" />
113
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="1" />
114
+ </map>
115
+ <map>
116
+ <framework name="onnx::Conv_284" output_port_id="onnx::Conv_284" />
117
+ <IR name="onnx::Conv_284" output_port_id="0" />
118
+ </map>
119
+ <map>
120
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0" />
121
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="2" />
122
+ </map>
123
+ <map>
124
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0" />
125
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="1" />
126
+ </map>
127
+ <map>
128
+ <framework name="onnx::Conv_287" output_port_id="onnx::Conv_287" />
129
+ <IR name="onnx::Conv_287" output_port_id="0" />
130
+ </map>
131
+ <map>
132
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0" />
133
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="2" />
134
+ </map>
135
+ <map>
136
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0" />
137
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="2" />
138
+ </map>
139
+ <map>
140
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0" />
141
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="1" />
142
+ </map>
143
+ <map>
144
+ <framework name="/Shape" output_port_id="/Shape_output_0" />
145
+ <IR name="/Shape" output_port_id="1" />
146
+ </map>
147
+ <map>
148
+ <framework name="/Constant_1" output_port_id="/Constant_1_output_0" />
149
+ <IR name="/Constant_1" output_port_id="0" />
150
+ </map>
151
+ <map>
152
+ <framework name="/Constant_2" output_port_id="/Constant_2_output_0" />
153
+ <IR name="/Constant_2" output_port_id="0" />
154
+ </map>
155
+ <map>
156
+ <framework name="/Slice" output_port_id="/Slice_output_0" />
157
+ <IR name="/Slice" output_port_id="4" />
158
+ </map>
159
+ <map>
160
+ <framework name="/Constant_3" output_port_id="/Constant_3_output_0" />
161
+ <IR name="/Constant_3" output_port_id="0" />
162
+ </map>
163
+ <map>
164
+ <framework name="/Concat" output_port_id="/Concat_output_0" />
165
+ <IR name="/Concat" output_port_id="2" />
166
+ </map>
167
+ <map>
168
+ <framework name="/Resize" output_port_id="/Resize_output_0" />
169
+ <IR name="/Resize" output_port_id="3" />
170
+ </map>
171
+ <map>
172
+ <framework name="/Concat_1" output_port_id="/Concat_1_output_0" />
173
+ <IR name="/Concat_1" output_port_id="2" />
174
+ </map>
175
+ <map>
176
+ <framework name="onnx::Conv_290" output_port_id="onnx::Conv_290" />
177
+ <IR name="onnx::Conv_290" output_port_id="0" />
178
+ </map>
179
+ <map>
180
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0" />
181
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="2" />
182
+ </map>
183
+ <map>
184
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0" />
185
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="1" />
186
+ </map>
187
+ <map>
188
+ <framework name="onnx::Conv_293" output_port_id="onnx::Conv_293" />
189
+ <IR name="onnx::Conv_293" output_port_id="0" />
190
+ </map>
191
+ <map>
192
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0" />
193
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="2" />
194
+ </map>
195
+ <map>
196
+ <framework name="onnx::Conv_296" output_port_id="onnx::Conv_296" />
197
+ <IR name="onnx::Conv_296" output_port_id="0" />
198
+ </map>
199
+ <map>
200
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0" />
201
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="2" />
202
+ </map>
203
+ <map>
204
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0" />
205
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="2" />
206
+ </map>
207
+ <map>
208
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0" />
209
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="1" />
210
+ </map>
211
+ <map>
212
+ <framework name="onnx::Conv_299" output_port_id="onnx::Conv_299" />
213
+ <IR name="onnx::Conv_299" output_port_id="0" />
214
+ </map>
215
+ <map>
216
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0" />
217
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="2" />
218
+ </map>
219
+ <map>
220
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0" />
221
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="1" />
222
+ </map>
223
+ <map>
224
+ <framework name="onnx::Conv_302" output_port_id="onnx::Conv_302" />
225
+ <IR name="onnx::Conv_302" output_port_id="0" />
226
+ </map>
227
+ <map>
228
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0" />
229
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="2" />
230
+ </map>
231
+ <map>
232
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0" />
233
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="2" />
234
+ </map>
235
+ <map>
236
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0" />
237
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="1" />
238
+ </map>
239
+ <map>
240
+ <framework name="/Shape_3" output_port_id="/Shape_3_output_0" />
241
+ <IR name="/Shape_3" output_port_id="1" />
242
+ </map>
243
+ <map>
244
+ <framework name="/Constant_9" output_port_id="/Constant_9_output_0" />
245
+ <IR name="/Constant_9" output_port_id="0" />
246
+ </map>
247
+ <map>
248
+ <framework name="/Constant_10" output_port_id="/Constant_10_output_0" />
249
+ <IR name="/Constant_10" output_port_id="0" />
250
+ </map>
251
+ <map>
252
+ <framework name="/Slice_1" output_port_id="/Slice_1_output_0" />
253
+ <IR name="/Slice_1" output_port_id="4" />
254
+ </map>
255
+ <map>
256
+ <framework name="/Shape_1" output_port_id="/Shape_1_output_0" />
257
+ <IR name="/Shape_1" output_port_id="1" />
258
+ </map>
259
+ <map>
260
+ <framework name="/Shape_1" output_port_id="/Shape_2_output_0" />
261
+ <IR name="/Shape_1" output_port_id="1" />
262
+ </map>
263
+ <map>
264
+ <framework name="Gather_6092" output_port_id="/Concat_2_output_0" />
265
+ <IR name="Gather_6092" output_port_id="3" />
266
+ </map>
267
+ <map>
268
+ <framework name="Gather_6092" output_port_id="/Cast_output_0" />
269
+ <IR name="Gather_6092" output_port_id="3" />
270
+ </map>
271
+ <map>
272
+ <framework name="/Concat_3" output_port_id="/Concat_3_output_0" />
273
+ <IR name="/Concat_3" output_port_id="2" />
274
+ </map>
275
+ <map>
276
+ <framework name="/Resize_1" output_port_id="/Resize_1_output_0" />
277
+ <IR name="/Resize_1" output_port_id="3" />
278
+ </map>
279
+ <map>
280
+ <framework name="/Concat_4" output_port_id="/Concat_4_output_0" />
281
+ <IR name="/Concat_4" output_port_id="2" />
282
+ </map>
283
+ <map>
284
+ <framework name="onnx::Gather_308" output_port_id="onnx::Gather_308" />
285
+ <IR name="onnx::Gather_308" output_port_id="0" />
286
+ </map>
287
+ <map>
288
+ <framework name="/Gather_2" output_port_id="/Gather_2_output_0" />
289
+ <IR name="/Gather_2" output_port_id="3" />
290
+ </map>
291
+ <map>
292
+ <framework name="/anomaly_map_generator/Constant" output_port_id="/anomaly_map_generator/Constant_output_0" />
293
+ <IR name="/anomaly_map_generator/Constant" output_port_id="0" />
294
+ </map>
295
+ <map>
296
+ <framework name="/anomaly_map_generator/Reshape" output_port_id="/anomaly_map_generator/Reshape_output_0" />
297
+ <IR name="/anomaly_map_generator/Reshape" output_port_id="2" />
298
+ </map>
299
+ <map>
300
+ <framework name="/anomaly_map_generator/Sub" output_port_id="/anomaly_map_generator/Sub_output_0" />
301
+ <IR name="/anomaly_map_generator/Sub" output_port_id="2" />
302
+ </map>
303
+ <map>
304
+ <framework name="/anomaly_map_generator/Transpose" output_port_id="/anomaly_map_generator/Transpose_output_0" />
305
+ <IR name="/anomaly_map_generator/Transpose" output_port_id="2" />
306
+ </map>
307
+ <map>
308
+ <framework name="onnx::MatMul_310" output_port_id="onnx::MatMul_310" />
309
+ <IR name="onnx::MatMul_310" output_port_id="0" />
310
+ </map>
311
+ <map>
312
+ <framework name="/anomaly_map_generator/MatMul" output_port_id="/anomaly_map_generator/MatMul_output_0" />
313
+ <IR name="/anomaly_map_generator/MatMul" output_port_id="2" />
314
+ </map>
315
+ <map>
316
+ <framework name="/anomaly_map_generator/Mul" output_port_id="/anomaly_map_generator/Mul_output_0" />
317
+ <IR name="/anomaly_map_generator/Mul" output_port_id="2" />
318
+ </map>
319
+ <map>
320
+ <framework name="/anomaly_map_generator/ReduceSum" output_port_id="/anomaly_map_generator/ReduceSum_output_0" />
321
+ <IR name="/anomaly_map_generator/ReduceSum" output_port_id="2" />
322
+ </map>
323
+ <map>
324
+ <framework name="/anomaly_map_generator/Constant_1" output_port_id="/anomaly_map_generator/Constant_1_output_0" />
325
+ <IR name="/anomaly_map_generator/Constant_1" output_port_id="0" />
326
+ </map>
327
+ <map>
328
+ <framework name="/anomaly_map_generator/Reshape_1" output_port_id="/anomaly_map_generator/Reshape_1_output_0" />
329
+ <IR name="/anomaly_map_generator/Reshape_1" output_port_id="2" />
330
+ </map>
331
+ <map>
332
+ <framework name="/anomaly_map_generator/Clip" output_port_id="/anomaly_map_generator/Clip_output_0" />
333
+ <IR name="/anomaly_map_generator/Clip" output_port_id="1" />
334
+ </map>
335
+ <map>
336
+ <framework name="/anomaly_map_generator/Sqrt" output_port_id="/anomaly_map_generator/Sqrt_output_0" />
337
+ <IR name="/anomaly_map_generator/Sqrt" output_port_id="1" />
338
+ </map>
339
+ <map>
340
+ <framework name="/anomaly_map_generator/Shape" output_port_id="/anomaly_map_generator/Shape_output_0" />
341
+ <IR name="/anomaly_map_generator/Shape" output_port_id="1" />
342
+ </map>
343
+ <map>
344
+ <framework name="/anomaly_map_generator/Constant_4" output_port_id="/anomaly_map_generator/Constant_4_output_0" />
345
+ <IR name="/anomaly_map_generator/Constant_4" output_port_id="0" />
346
+ </map>
347
+ <map>
348
+ <framework name="/anomaly_map_generator/Constant_5" output_port_id="/anomaly_map_generator/Constant_5_output_0" />
349
+ <IR name="/anomaly_map_generator/Constant_5" output_port_id="0" />
350
+ </map>
351
+ <map>
352
+ <framework name="/anomaly_map_generator/Slice" output_port_id="/anomaly_map_generator/Slice_output_0" />
353
+ <IR name="/anomaly_map_generator/Slice" output_port_id="4" />
354
+ </map>
355
+ <map>
356
+ <framework name="/anomaly_map_generator/Constant_6" output_port_id="/anomaly_map_generator/Constant_6_output_0" />
357
+ <IR name="/anomaly_map_generator/Constant_6" output_port_id="0" />
358
+ </map>
359
+ <map>
360
+ <framework name="/anomaly_map_generator/Concat" output_port_id="/anomaly_map_generator/Concat_output_0" />
361
+ <IR name="/anomaly_map_generator/Concat" output_port_id="2" />
362
+ </map>
363
+ <map>
364
+ <framework name="/anomaly_map_generator/Resize" output_port_id="/anomaly_map_generator/Resize_output_0" />
365
+ <IR name="/anomaly_map_generator/Resize" output_port_id="3" />
366
+ </map>
367
+ <map>
368
+ <framework name="/anomaly_map_generator/blur/Pad" output_port_id="/anomaly_map_generator/blur/Pad_output_0" />
369
+ <IR name="/anomaly_map_generator/blur/Pad" output_port_id="4" />
370
+ </map>
371
+ <map>
372
+ <framework name="anomaly_map_generator.blur.kernel" output_port_id="anomaly_map_generator.blur.kernel" />
373
+ <IR name="anomaly_map_generator.blur.kernel" output_port_id="0" />
374
+ </map>
375
+ <map>
376
+ <framework name="/anomaly_map_generator/blur/Conv" output_port_id="/anomaly_map_generator/blur/Conv_output_0" />
377
+ <IR name="/anomaly_map_generator/blur/Conv" output_port_id="2" />
378
+ </map>
379
+ <map>
380
+ <framework name="/anomaly_map_generator/blur/Constant_8" output_port_id="/anomaly_map_generator/blur/Constant_8_output_0" />
381
+ <IR name="/anomaly_map_generator/blur/Constant_8" output_port_id="0" />
382
+ </map>
383
+ <map>
384
+ <framework name="output" output_port_id="output" />
385
+ <IR name="output" output_port_id="2" />
386
+ </map>
387
+ </mapping>
grid/run/weights/openvino/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a5c998a95ec492be0d9981a09e7afa3f249ffad9589464d6c9ffc89d65930f9
3
+ size 176626689
grid/run/weights/openvino/model.xml ADDED
@@ -0,0 +1,2703 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="torch_jit" version="11">
3
+ <layers>
4
+ <layer id="0" name="input" type="Parameter" version="opset1">
5
+ <data shape="1,3,256,256" element_type="f32" />
6
+ <output>
7
+ <port id="0" precision="FP32" names="input">
8
+ <dim>1</dim>
9
+ <dim>3</dim>
10
+ <dim>256</dim>
11
+ <dim>256</dim>
12
+ </port>
13
+ </output>
14
+ </layer>
15
+ <layer id="1" name="onnx::Conv_260" type="Const" version="opset1">
16
+ <data element_type="f32" shape="64, 3, 7, 7" offset="0" size="37632" />
17
+ <output>
18
+ <port id="0" precision="FP32" names="onnx::Conv_260">
19
+ <dim>64</dim>
20
+ <dim>3</dim>
21
+ <dim>7</dim>
22
+ <dim>7</dim>
23
+ </port>
24
+ </output>
25
+ </layer>
26
+ <layer id="2" name="/feature_extractor/feature_extractor/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
27
+ <data strides="2, 2" dilations="1, 1" pads_begin="3, 3" pads_end="3, 3" auto_pad="explicit" />
28
+ <input>
29
+ <port id="0" precision="FP32">
30
+ <dim>1</dim>
31
+ <dim>3</dim>
32
+ <dim>256</dim>
33
+ <dim>256</dim>
34
+ </port>
35
+ <port id="1" precision="FP32">
36
+ <dim>64</dim>
37
+ <dim>3</dim>
38
+ <dim>7</dim>
39
+ <dim>7</dim>
40
+ </port>
41
+ </input>
42
+ <output>
43
+ <port id="2" precision="FP32">
44
+ <dim>1</dim>
45
+ <dim>64</dim>
46
+ <dim>128</dim>
47
+ <dim>128</dim>
48
+ </port>
49
+ </output>
50
+ </layer>
51
+ <layer id="3" name="Reshape_55" type="Const" version="opset1">
52
+ <data element_type="f32" shape="1, 64, 1, 1" offset="37632" size="256" />
53
+ <output>
54
+ <port id="0" precision="FP32">
55
+ <dim>1</dim>
56
+ <dim>64</dim>
57
+ <dim>1</dim>
58
+ <dim>1</dim>
59
+ </port>
60
+ </output>
61
+ </layer>
62
+ <layer id="4" name="/feature_extractor/feature_extractor/conv1/Conv" type="Add" version="opset1">
63
+ <data auto_broadcast="numpy" />
64
+ <input>
65
+ <port id="0" precision="FP32">
66
+ <dim>1</dim>
67
+ <dim>64</dim>
68
+ <dim>128</dim>
69
+ <dim>128</dim>
70
+ </port>
71
+ <port id="1" precision="FP32">
72
+ <dim>1</dim>
73
+ <dim>64</dim>
74
+ <dim>1</dim>
75
+ <dim>1</dim>
76
+ </port>
77
+ </input>
78
+ <output>
79
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/conv1/Conv_output_0">
80
+ <dim>1</dim>
81
+ <dim>64</dim>
82
+ <dim>128</dim>
83
+ <dim>128</dim>
84
+ </port>
85
+ </output>
86
+ </layer>
87
+ <layer id="5" name="/feature_extractor/feature_extractor/act1/Relu" type="ReLU" version="opset1">
88
+ <input>
89
+ <port id="0" precision="FP32">
90
+ <dim>1</dim>
91
+ <dim>64</dim>
92
+ <dim>128</dim>
93
+ <dim>128</dim>
94
+ </port>
95
+ </input>
96
+ <output>
97
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/act1/Relu_output_0">
98
+ <dim>1</dim>
99
+ <dim>64</dim>
100
+ <dim>128</dim>
101
+ <dim>128</dim>
102
+ </port>
103
+ </output>
104
+ </layer>
105
+ <layer id="6" name="/feature_extractor/feature_extractor/maxpool/MaxPool" type="MaxPool" version="opset8">
106
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" kernel="3, 3" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
107
+ <input>
108
+ <port id="0" precision="FP32">
109
+ <dim>1</dim>
110
+ <dim>64</dim>
111
+ <dim>128</dim>
112
+ <dim>128</dim>
113
+ </port>
114
+ </input>
115
+ <output>
116
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0">
117
+ <dim>1</dim>
118
+ <dim>64</dim>
119
+ <dim>64</dim>
120
+ <dim>64</dim>
121
+ </port>
122
+ <port id="2" precision="I64">
123
+ <dim>1</dim>
124
+ <dim>64</dim>
125
+ <dim>64</dim>
126
+ <dim>64</dim>
127
+ </port>
128
+ </output>
129
+ </layer>
130
+ <layer id="7" name="onnx::Conv_263" type="Const" version="opset1">
131
+ <data element_type="f32" shape="64, 64, 3, 3" offset="37888" size="147456" />
132
+ <output>
133
+ <port id="0" precision="FP32" names="onnx::Conv_263">
134
+ <dim>64</dim>
135
+ <dim>64</dim>
136
+ <dim>3</dim>
137
+ <dim>3</dim>
138
+ </port>
139
+ </output>
140
+ </layer>
141
+ <layer id="8" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
142
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
143
+ <input>
144
+ <port id="0" precision="FP32">
145
+ <dim>1</dim>
146
+ <dim>64</dim>
147
+ <dim>64</dim>
148
+ <dim>64</dim>
149
+ </port>
150
+ <port id="1" precision="FP32">
151
+ <dim>64</dim>
152
+ <dim>64</dim>
153
+ <dim>3</dim>
154
+ <dim>3</dim>
155
+ </port>
156
+ </input>
157
+ <output>
158
+ <port id="2" precision="FP32">
159
+ <dim>1</dim>
160
+ <dim>64</dim>
161
+ <dim>64</dim>
162
+ <dim>64</dim>
163
+ </port>
164
+ </output>
165
+ </layer>
166
+ <layer id="9" name="Reshape_105" type="Const" version="opset1">
167
+ <data element_type="f32" shape="1, 64, 1, 1" offset="185344" size="256" />
168
+ <output>
169
+ <port id="0" precision="FP32">
170
+ <dim>1</dim>
171
+ <dim>64</dim>
172
+ <dim>1</dim>
173
+ <dim>1</dim>
174
+ </port>
175
+ </output>
176
+ </layer>
177
+ <layer id="10" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" type="Add" version="opset1">
178
+ <data auto_broadcast="numpy" />
179
+ <input>
180
+ <port id="0" precision="FP32">
181
+ <dim>1</dim>
182
+ <dim>64</dim>
183
+ <dim>64</dim>
184
+ <dim>64</dim>
185
+ </port>
186
+ <port id="1" precision="FP32">
187
+ <dim>1</dim>
188
+ <dim>64</dim>
189
+ <dim>1</dim>
190
+ <dim>1</dim>
191
+ </port>
192
+ </input>
193
+ <output>
194
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0">
195
+ <dim>1</dim>
196
+ <dim>64</dim>
197
+ <dim>64</dim>
198
+ <dim>64</dim>
199
+ </port>
200
+ </output>
201
+ </layer>
202
+ <layer id="11" name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" type="ReLU" version="opset1">
203
+ <input>
204
+ <port id="0" precision="FP32">
205
+ <dim>1</dim>
206
+ <dim>64</dim>
207
+ <dim>64</dim>
208
+ <dim>64</dim>
209
+ </port>
210
+ </input>
211
+ <output>
212
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0">
213
+ <dim>1</dim>
214
+ <dim>64</dim>
215
+ <dim>64</dim>
216
+ <dim>64</dim>
217
+ </port>
218
+ </output>
219
+ </layer>
220
+ <layer id="12" name="onnx::Conv_266" type="Const" version="opset1">
221
+ <data element_type="f32" shape="64, 64, 3, 3" offset="185600" size="147456" />
222
+ <output>
223
+ <port id="0" precision="FP32" names="onnx::Conv_266">
224
+ <dim>64</dim>
225
+ <dim>64</dim>
226
+ <dim>3</dim>
227
+ <dim>3</dim>
228
+ </port>
229
+ </output>
230
+ </layer>
231
+ <layer id="13" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
232
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
233
+ <input>
234
+ <port id="0" precision="FP32">
235
+ <dim>1</dim>
236
+ <dim>64</dim>
237
+ <dim>64</dim>
238
+ <dim>64</dim>
239
+ </port>
240
+ <port id="1" precision="FP32">
241
+ <dim>64</dim>
242
+ <dim>64</dim>
243
+ <dim>3</dim>
244
+ <dim>3</dim>
245
+ </port>
246
+ </input>
247
+ <output>
248
+ <port id="2" precision="FP32">
249
+ <dim>1</dim>
250
+ <dim>64</dim>
251
+ <dim>64</dim>
252
+ <dim>64</dim>
253
+ </port>
254
+ </output>
255
+ </layer>
256
+ <layer id="14" name="Reshape_154" type="Const" version="opset1">
257
+ <data element_type="f32" shape="1, 64, 1, 1" offset="333056" size="256" />
258
+ <output>
259
+ <port id="0" precision="FP32">
260
+ <dim>1</dim>
261
+ <dim>64</dim>
262
+ <dim>1</dim>
263
+ <dim>1</dim>
264
+ </port>
265
+ </output>
266
+ </layer>
267
+ <layer id="15" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" type="Add" version="opset1">
268
+ <data auto_broadcast="numpy" />
269
+ <input>
270
+ <port id="0" precision="FP32">
271
+ <dim>1</dim>
272
+ <dim>64</dim>
273
+ <dim>64</dim>
274
+ <dim>64</dim>
275
+ </port>
276
+ <port id="1" precision="FP32">
277
+ <dim>1</dim>
278
+ <dim>64</dim>
279
+ <dim>1</dim>
280
+ <dim>1</dim>
281
+ </port>
282
+ </input>
283
+ <output>
284
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0">
285
+ <dim>1</dim>
286
+ <dim>64</dim>
287
+ <dim>64</dim>
288
+ <dim>64</dim>
289
+ </port>
290
+ </output>
291
+ </layer>
292
+ <layer id="16" name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" type="Add" version="opset1">
293
+ <data auto_broadcast="numpy" />
294
+ <input>
295
+ <port id="0" precision="FP32">
296
+ <dim>1</dim>
297
+ <dim>64</dim>
298
+ <dim>64</dim>
299
+ <dim>64</dim>
300
+ </port>
301
+ <port id="1" precision="FP32">
302
+ <dim>1</dim>
303
+ <dim>64</dim>
304
+ <dim>64</dim>
305
+ <dim>64</dim>
306
+ </port>
307
+ </input>
308
+ <output>
309
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0">
310
+ <dim>1</dim>
311
+ <dim>64</dim>
312
+ <dim>64</dim>
313
+ <dim>64</dim>
314
+ </port>
315
+ </output>
316
+ </layer>
317
+ <layer id="17" name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" type="ReLU" version="opset1">
318
+ <input>
319
+ <port id="0" precision="FP32">
320
+ <dim>1</dim>
321
+ <dim>64</dim>
322
+ <dim>64</dim>
323
+ <dim>64</dim>
324
+ </port>
325
+ </input>
326
+ <output>
327
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0">
328
+ <dim>1</dim>
329
+ <dim>64</dim>
330
+ <dim>64</dim>
331
+ <dim>64</dim>
332
+ </port>
333
+ </output>
334
+ </layer>
335
+ <layer id="18" name="onnx::Conv_269" type="Const" version="opset1">
336
+ <data element_type="f32" shape="64, 64, 3, 3" offset="333312" size="147456" />
337
+ <output>
338
+ <port id="0" precision="FP32" names="onnx::Conv_269">
339
+ <dim>64</dim>
340
+ <dim>64</dim>
341
+ <dim>3</dim>
342
+ <dim>3</dim>
343
+ </port>
344
+ </output>
345
+ </layer>
346
+ <layer id="19" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
347
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
348
+ <input>
349
+ <port id="0" precision="FP32">
350
+ <dim>1</dim>
351
+ <dim>64</dim>
352
+ <dim>64</dim>
353
+ <dim>64</dim>
354
+ </port>
355
+ <port id="1" precision="FP32">
356
+ <dim>64</dim>
357
+ <dim>64</dim>
358
+ <dim>3</dim>
359
+ <dim>3</dim>
360
+ </port>
361
+ </input>
362
+ <output>
363
+ <port id="2" precision="FP32">
364
+ <dim>1</dim>
365
+ <dim>64</dim>
366
+ <dim>64</dim>
367
+ <dim>64</dim>
368
+ </port>
369
+ </output>
370
+ </layer>
371
+ <layer id="20" name="Reshape_204" type="Const" version="opset1">
372
+ <data element_type="f32" shape="1, 64, 1, 1" offset="480768" size="256" />
373
+ <output>
374
+ <port id="0" precision="FP32">
375
+ <dim>1</dim>
376
+ <dim>64</dim>
377
+ <dim>1</dim>
378
+ <dim>1</dim>
379
+ </port>
380
+ </output>
381
+ </layer>
382
+ <layer id="21" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" type="Add" version="opset1">
383
+ <data auto_broadcast="numpy" />
384
+ <input>
385
+ <port id="0" precision="FP32">
386
+ <dim>1</dim>
387
+ <dim>64</dim>
388
+ <dim>64</dim>
389
+ <dim>64</dim>
390
+ </port>
391
+ <port id="1" precision="FP32">
392
+ <dim>1</dim>
393
+ <dim>64</dim>
394
+ <dim>1</dim>
395
+ <dim>1</dim>
396
+ </port>
397
+ </input>
398
+ <output>
399
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0">
400
+ <dim>1</dim>
401
+ <dim>64</dim>
402
+ <dim>64</dim>
403
+ <dim>64</dim>
404
+ </port>
405
+ </output>
406
+ </layer>
407
+ <layer id="22" name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" type="ReLU" version="opset1">
408
+ <input>
409
+ <port id="0" precision="FP32">
410
+ <dim>1</dim>
411
+ <dim>64</dim>
412
+ <dim>64</dim>
413
+ <dim>64</dim>
414
+ </port>
415
+ </input>
416
+ <output>
417
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0">
418
+ <dim>1</dim>
419
+ <dim>64</dim>
420
+ <dim>64</dim>
421
+ <dim>64</dim>
422
+ </port>
423
+ </output>
424
+ </layer>
425
+ <layer id="23" name="onnx::Conv_272" type="Const" version="opset1">
426
+ <data element_type="f32" shape="64, 64, 3, 3" offset="481024" size="147456" />
427
+ <output>
428
+ <port id="0" precision="FP32" names="onnx::Conv_272">
429
+ <dim>64</dim>
430
+ <dim>64</dim>
431
+ <dim>3</dim>
432
+ <dim>3</dim>
433
+ </port>
434
+ </output>
435
+ </layer>
436
+ <layer id="24" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
437
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
438
+ <input>
439
+ <port id="0" precision="FP32">
440
+ <dim>1</dim>
441
+ <dim>64</dim>
442
+ <dim>64</dim>
443
+ <dim>64</dim>
444
+ </port>
445
+ <port id="1" precision="FP32">
446
+ <dim>64</dim>
447
+ <dim>64</dim>
448
+ <dim>3</dim>
449
+ <dim>3</dim>
450
+ </port>
451
+ </input>
452
+ <output>
453
+ <port id="2" precision="FP32">
454
+ <dim>1</dim>
455
+ <dim>64</dim>
456
+ <dim>64</dim>
457
+ <dim>64</dim>
458
+ </port>
459
+ </output>
460
+ </layer>
461
+ <layer id="25" name="Reshape_253" type="Const" version="opset1">
462
+ <data element_type="f32" shape="1, 64, 1, 1" offset="628480" size="256" />
463
+ <output>
464
+ <port id="0" precision="FP32">
465
+ <dim>1</dim>
466
+ <dim>64</dim>
467
+ <dim>1</dim>
468
+ <dim>1</dim>
469
+ </port>
470
+ </output>
471
+ </layer>
472
+ <layer id="26" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" type="Add" version="opset1">
473
+ <data auto_broadcast="numpy" />
474
+ <input>
475
+ <port id="0" precision="FP32">
476
+ <dim>1</dim>
477
+ <dim>64</dim>
478
+ <dim>64</dim>
479
+ <dim>64</dim>
480
+ </port>
481
+ <port id="1" precision="FP32">
482
+ <dim>1</dim>
483
+ <dim>64</dim>
484
+ <dim>1</dim>
485
+ <dim>1</dim>
486
+ </port>
487
+ </input>
488
+ <output>
489
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0">
490
+ <dim>1</dim>
491
+ <dim>64</dim>
492
+ <dim>64</dim>
493
+ <dim>64</dim>
494
+ </port>
495
+ </output>
496
+ </layer>
497
+ <layer id="27" name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" type="Add" version="opset1">
498
+ <data auto_broadcast="numpy" />
499
+ <input>
500
+ <port id="0" precision="FP32">
501
+ <dim>1</dim>
502
+ <dim>64</dim>
503
+ <dim>64</dim>
504
+ <dim>64</dim>
505
+ </port>
506
+ <port id="1" precision="FP32">
507
+ <dim>1</dim>
508
+ <dim>64</dim>
509
+ <dim>64</dim>
510
+ <dim>64</dim>
511
+ </port>
512
+ </input>
513
+ <output>
514
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0">
515
+ <dim>1</dim>
516
+ <dim>64</dim>
517
+ <dim>64</dim>
518
+ <dim>64</dim>
519
+ </port>
520
+ </output>
521
+ </layer>
522
+ <layer id="28" name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" type="ReLU" version="opset1">
523
+ <input>
524
+ <port id="0" precision="FP32">
525
+ <dim>1</dim>
526
+ <dim>64</dim>
527
+ <dim>64</dim>
528
+ <dim>64</dim>
529
+ </port>
530
+ </input>
531
+ <output>
532
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0">
533
+ <dim>1</dim>
534
+ <dim>64</dim>
535
+ <dim>64</dim>
536
+ <dim>64</dim>
537
+ </port>
538
+ </output>
539
+ </layer>
540
+ <layer id="29" name="onnx::Conv_275" type="Const" version="opset1">
541
+ <data element_type="f32" shape="128, 64, 3, 3" offset="628736" size="294912" />
542
+ <output>
543
+ <port id="0" precision="FP32" names="onnx::Conv_275">
544
+ <dim>128</dim>
545
+ <dim>64</dim>
546
+ <dim>3</dim>
547
+ <dim>3</dim>
548
+ </port>
549
+ </output>
550
+ </layer>
551
+ <layer id="30" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
552
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
553
+ <input>
554
+ <port id="0" precision="FP32">
555
+ <dim>1</dim>
556
+ <dim>64</dim>
557
+ <dim>64</dim>
558
+ <dim>64</dim>
559
+ </port>
560
+ <port id="1" precision="FP32">
561
+ <dim>128</dim>
562
+ <dim>64</dim>
563
+ <dim>3</dim>
564
+ <dim>3</dim>
565
+ </port>
566
+ </input>
567
+ <output>
568
+ <port id="2" precision="FP32">
569
+ <dim>1</dim>
570
+ <dim>128</dim>
571
+ <dim>32</dim>
572
+ <dim>32</dim>
573
+ </port>
574
+ </output>
575
+ </layer>
576
+ <layer id="31" name="Reshape_303" type="Const" version="opset1">
577
+ <data element_type="f32" shape="1, 128, 1, 1" offset="923648" size="512" />
578
+ <output>
579
+ <port id="0" precision="FP32">
580
+ <dim>1</dim>
581
+ <dim>128</dim>
582
+ <dim>1</dim>
583
+ <dim>1</dim>
584
+ </port>
585
+ </output>
586
+ </layer>
587
+ <layer id="32" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" type="Add" version="opset1">
588
+ <data auto_broadcast="numpy" />
589
+ <input>
590
+ <port id="0" precision="FP32">
591
+ <dim>1</dim>
592
+ <dim>128</dim>
593
+ <dim>32</dim>
594
+ <dim>32</dim>
595
+ </port>
596
+ <port id="1" precision="FP32">
597
+ <dim>1</dim>
598
+ <dim>128</dim>
599
+ <dim>1</dim>
600
+ <dim>1</dim>
601
+ </port>
602
+ </input>
603
+ <output>
604
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0">
605
+ <dim>1</dim>
606
+ <dim>128</dim>
607
+ <dim>32</dim>
608
+ <dim>32</dim>
609
+ </port>
610
+ </output>
611
+ </layer>
612
+ <layer id="33" name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" type="ReLU" version="opset1">
613
+ <input>
614
+ <port id="0" precision="FP32">
615
+ <dim>1</dim>
616
+ <dim>128</dim>
617
+ <dim>32</dim>
618
+ <dim>32</dim>
619
+ </port>
620
+ </input>
621
+ <output>
622
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0">
623
+ <dim>1</dim>
624
+ <dim>128</dim>
625
+ <dim>32</dim>
626
+ <dim>32</dim>
627
+ </port>
628
+ </output>
629
+ </layer>
630
+ <layer id="34" name="onnx::Conv_278" type="Const" version="opset1">
631
+ <data element_type="f32" shape="128, 128, 3, 3" offset="924160" size="589824" />
632
+ <output>
633
+ <port id="0" precision="FP32" names="onnx::Conv_278">
634
+ <dim>128</dim>
635
+ <dim>128</dim>
636
+ <dim>3</dim>
637
+ <dim>3</dim>
638
+ </port>
639
+ </output>
640
+ </layer>
641
+ <layer id="35" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
642
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
643
+ <input>
644
+ <port id="0" precision="FP32">
645
+ <dim>1</dim>
646
+ <dim>128</dim>
647
+ <dim>32</dim>
648
+ <dim>32</dim>
649
+ </port>
650
+ <port id="1" precision="FP32">
651
+ <dim>128</dim>
652
+ <dim>128</dim>
653
+ <dim>3</dim>
654
+ <dim>3</dim>
655
+ </port>
656
+ </input>
657
+ <output>
658
+ <port id="2" precision="FP32">
659
+ <dim>1</dim>
660
+ <dim>128</dim>
661
+ <dim>32</dim>
662
+ <dim>32</dim>
663
+ </port>
664
+ </output>
665
+ </layer>
666
+ <layer id="36" name="Reshape_352" type="Const" version="opset1">
667
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1513984" size="512" />
668
+ <output>
669
+ <port id="0" precision="FP32">
670
+ <dim>1</dim>
671
+ <dim>128</dim>
672
+ <dim>1</dim>
673
+ <dim>1</dim>
674
+ </port>
675
+ </output>
676
+ </layer>
677
+ <layer id="37" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" type="Add" version="opset1">
678
+ <data auto_broadcast="numpy" />
679
+ <input>
680
+ <port id="0" precision="FP32">
681
+ <dim>1</dim>
682
+ <dim>128</dim>
683
+ <dim>32</dim>
684
+ <dim>32</dim>
685
+ </port>
686
+ <port id="1" precision="FP32">
687
+ <dim>1</dim>
688
+ <dim>128</dim>
689
+ <dim>1</dim>
690
+ <dim>1</dim>
691
+ </port>
692
+ </input>
693
+ <output>
694
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0">
695
+ <dim>1</dim>
696
+ <dim>128</dim>
697
+ <dim>32</dim>
698
+ <dim>32</dim>
699
+ </port>
700
+ </output>
701
+ </layer>
702
+ <layer id="38" name="onnx::Conv_281" type="Const" version="opset1">
703
+ <data element_type="f32" shape="128, 64, 1, 1" offset="1514496" size="32768" />
704
+ <output>
705
+ <port id="0" precision="FP32" names="onnx::Conv_281">
706
+ <dim>128</dim>
707
+ <dim>64</dim>
708
+ <dim>1</dim>
709
+ <dim>1</dim>
710
+ </port>
711
+ </output>
712
+ </layer>
713
+ <layer id="39" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
714
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
715
+ <input>
716
+ <port id="0" precision="FP32">
717
+ <dim>1</dim>
718
+ <dim>64</dim>
719
+ <dim>64</dim>
720
+ <dim>64</dim>
721
+ </port>
722
+ <port id="1" precision="FP32">
723
+ <dim>128</dim>
724
+ <dim>64</dim>
725
+ <dim>1</dim>
726
+ <dim>1</dim>
727
+ </port>
728
+ </input>
729
+ <output>
730
+ <port id="2" precision="FP32">
731
+ <dim>1</dim>
732
+ <dim>128</dim>
733
+ <dim>32</dim>
734
+ <dim>32</dim>
735
+ </port>
736
+ </output>
737
+ </layer>
738
+ <layer id="40" name="Reshape_400" type="Const" version="opset1">
739
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1547264" size="512" />
740
+ <output>
741
+ <port id="0" precision="FP32">
742
+ <dim>1</dim>
743
+ <dim>128</dim>
744
+ <dim>1</dim>
745
+ <dim>1</dim>
746
+ </port>
747
+ </output>
748
+ </layer>
749
+ <layer id="41" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" type="Add" version="opset1">
750
+ <data auto_broadcast="numpy" />
751
+ <input>
752
+ <port id="0" precision="FP32">
753
+ <dim>1</dim>
754
+ <dim>128</dim>
755
+ <dim>32</dim>
756
+ <dim>32</dim>
757
+ </port>
758
+ <port id="1" precision="FP32">
759
+ <dim>1</dim>
760
+ <dim>128</dim>
761
+ <dim>1</dim>
762
+ <dim>1</dim>
763
+ </port>
764
+ </input>
765
+ <output>
766
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0">
767
+ <dim>1</dim>
768
+ <dim>128</dim>
769
+ <dim>32</dim>
770
+ <dim>32</dim>
771
+ </port>
772
+ </output>
773
+ </layer>
774
+ <layer id="42" name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" type="Add" version="opset1">
775
+ <data auto_broadcast="numpy" />
776
+ <input>
777
+ <port id="0" precision="FP32">
778
+ <dim>1</dim>
779
+ <dim>128</dim>
780
+ <dim>32</dim>
781
+ <dim>32</dim>
782
+ </port>
783
+ <port id="1" precision="FP32">
784
+ <dim>1</dim>
785
+ <dim>128</dim>
786
+ <dim>32</dim>
787
+ <dim>32</dim>
788
+ </port>
789
+ </input>
790
+ <output>
791
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0">
792
+ <dim>1</dim>
793
+ <dim>128</dim>
794
+ <dim>32</dim>
795
+ <dim>32</dim>
796
+ </port>
797
+ </output>
798
+ </layer>
799
+ <layer id="43" name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" type="ReLU" version="opset1">
800
+ <input>
801
+ <port id="0" precision="FP32">
802
+ <dim>1</dim>
803
+ <dim>128</dim>
804
+ <dim>32</dim>
805
+ <dim>32</dim>
806
+ </port>
807
+ </input>
808
+ <output>
809
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0">
810
+ <dim>1</dim>
811
+ <dim>128</dim>
812
+ <dim>32</dim>
813
+ <dim>32</dim>
814
+ </port>
815
+ </output>
816
+ </layer>
817
+ <layer id="44" name="onnx::Conv_284" type="Const" version="opset1">
818
+ <data element_type="f32" shape="128, 128, 3, 3" offset="1547776" size="589824" />
819
+ <output>
820
+ <port id="0" precision="FP32" names="onnx::Conv_284">
821
+ <dim>128</dim>
822
+ <dim>128</dim>
823
+ <dim>3</dim>
824
+ <dim>3</dim>
825
+ </port>
826
+ </output>
827
+ </layer>
828
+ <layer id="45" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
829
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
830
+ <input>
831
+ <port id="0" precision="FP32">
832
+ <dim>1</dim>
833
+ <dim>128</dim>
834
+ <dim>32</dim>
835
+ <dim>32</dim>
836
+ </port>
837
+ <port id="1" precision="FP32">
838
+ <dim>128</dim>
839
+ <dim>128</dim>
840
+ <dim>3</dim>
841
+ <dim>3</dim>
842
+ </port>
843
+ </input>
844
+ <output>
845
+ <port id="2" precision="FP32">
846
+ <dim>1</dim>
847
+ <dim>128</dim>
848
+ <dim>32</dim>
849
+ <dim>32</dim>
850
+ </port>
851
+ </output>
852
+ </layer>
853
+ <layer id="46" name="Reshape_450" type="Const" version="opset1">
854
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2137600" size="512" />
855
+ <output>
856
+ <port id="0" precision="FP32">
857
+ <dim>1</dim>
858
+ <dim>128</dim>
859
+ <dim>1</dim>
860
+ <dim>1</dim>
861
+ </port>
862
+ </output>
863
+ </layer>
864
+ <layer id="47" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" type="Add" version="opset1">
865
+ <data auto_broadcast="numpy" />
866
+ <input>
867
+ <port id="0" precision="FP32">
868
+ <dim>1</dim>
869
+ <dim>128</dim>
870
+ <dim>32</dim>
871
+ <dim>32</dim>
872
+ </port>
873
+ <port id="1" precision="FP32">
874
+ <dim>1</dim>
875
+ <dim>128</dim>
876
+ <dim>1</dim>
877
+ <dim>1</dim>
878
+ </port>
879
+ </input>
880
+ <output>
881
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0">
882
+ <dim>1</dim>
883
+ <dim>128</dim>
884
+ <dim>32</dim>
885
+ <dim>32</dim>
886
+ </port>
887
+ </output>
888
+ </layer>
889
+ <layer id="48" name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" type="ReLU" version="opset1">
890
+ <input>
891
+ <port id="0" precision="FP32">
892
+ <dim>1</dim>
893
+ <dim>128</dim>
894
+ <dim>32</dim>
895
+ <dim>32</dim>
896
+ </port>
897
+ </input>
898
+ <output>
899
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0">
900
+ <dim>1</dim>
901
+ <dim>128</dim>
902
+ <dim>32</dim>
903
+ <dim>32</dim>
904
+ </port>
905
+ </output>
906
+ </layer>
907
+ <layer id="49" name="onnx::Conv_287" type="Const" version="opset1">
908
+ <data element_type="f32" shape="128, 128, 3, 3" offset="2138112" size="589824" />
909
+ <output>
910
+ <port id="0" precision="FP32" names="onnx::Conv_287">
911
+ <dim>128</dim>
912
+ <dim>128</dim>
913
+ <dim>3</dim>
914
+ <dim>3</dim>
915
+ </port>
916
+ </output>
917
+ </layer>
918
+ <layer id="50" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
919
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
920
+ <input>
921
+ <port id="0" precision="FP32">
922
+ <dim>1</dim>
923
+ <dim>128</dim>
924
+ <dim>32</dim>
925
+ <dim>32</dim>
926
+ </port>
927
+ <port id="1" precision="FP32">
928
+ <dim>128</dim>
929
+ <dim>128</dim>
930
+ <dim>3</dim>
931
+ <dim>3</dim>
932
+ </port>
933
+ </input>
934
+ <output>
935
+ <port id="2" precision="FP32">
936
+ <dim>1</dim>
937
+ <dim>128</dim>
938
+ <dim>32</dim>
939
+ <dim>32</dim>
940
+ </port>
941
+ </output>
942
+ </layer>
943
+ <layer id="51" name="Reshape_499" type="Const" version="opset1">
944
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2727936" size="512" />
945
+ <output>
946
+ <port id="0" precision="FP32">
947
+ <dim>1</dim>
948
+ <dim>128</dim>
949
+ <dim>1</dim>
950
+ <dim>1</dim>
951
+ </port>
952
+ </output>
953
+ </layer>
954
+ <layer id="52" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" type="Add" version="opset1">
955
+ <data auto_broadcast="numpy" />
956
+ <input>
957
+ <port id="0" precision="FP32">
958
+ <dim>1</dim>
959
+ <dim>128</dim>
960
+ <dim>32</dim>
961
+ <dim>32</dim>
962
+ </port>
963
+ <port id="1" precision="FP32">
964
+ <dim>1</dim>
965
+ <dim>128</dim>
966
+ <dim>1</dim>
967
+ <dim>1</dim>
968
+ </port>
969
+ </input>
970
+ <output>
971
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0">
972
+ <dim>1</dim>
973
+ <dim>128</dim>
974
+ <dim>32</dim>
975
+ <dim>32</dim>
976
+ </port>
977
+ </output>
978
+ </layer>
979
+ <layer id="53" name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" type="Add" version="opset1">
980
+ <data auto_broadcast="numpy" />
981
+ <input>
982
+ <port id="0" precision="FP32">
983
+ <dim>1</dim>
984
+ <dim>128</dim>
985
+ <dim>32</dim>
986
+ <dim>32</dim>
987
+ </port>
988
+ <port id="1" precision="FP32">
989
+ <dim>1</dim>
990
+ <dim>128</dim>
991
+ <dim>32</dim>
992
+ <dim>32</dim>
993
+ </port>
994
+ </input>
995
+ <output>
996
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0">
997
+ <dim>1</dim>
998
+ <dim>128</dim>
999
+ <dim>32</dim>
1000
+ <dim>32</dim>
1001
+ </port>
1002
+ </output>
1003
+ </layer>
1004
+ <layer id="54" name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" type="ReLU" version="opset1">
1005
+ <input>
1006
+ <port id="0" precision="FP32">
1007
+ <dim>1</dim>
1008
+ <dim>128</dim>
1009
+ <dim>32</dim>
1010
+ <dim>32</dim>
1011
+ </port>
1012
+ </input>
1013
+ <output>
1014
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0">
1015
+ <dim>1</dim>
1016
+ <dim>128</dim>
1017
+ <dim>32</dim>
1018
+ <dim>32</dim>
1019
+ </port>
1020
+ </output>
1021
+ </layer>
1022
+ <layer id="55" name="/Shape" type="ShapeOf" version="opset3">
1023
+ <data output_type="i64" />
1024
+ <input>
1025
+ <port id="0" precision="FP32">
1026
+ <dim>1</dim>
1027
+ <dim>128</dim>
1028
+ <dim>32</dim>
1029
+ <dim>32</dim>
1030
+ </port>
1031
+ </input>
1032
+ <output>
1033
+ <port id="1" precision="I64" names="/Shape_output_0">
1034
+ <dim>4</dim>
1035
+ </port>
1036
+ </output>
1037
+ </layer>
1038
+ <layer id="56" name="/Constant_1" type="Const" version="opset1">
1039
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1040
+ <output>
1041
+ <port id="0" precision="I64" names="/Constant_1_output_0">
1042
+ <dim>1</dim>
1043
+ </port>
1044
+ </output>
1045
+ </layer>
1046
+ <layer id="57" name="/Constant_2" type="Const" version="opset1">
1047
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1048
+ <output>
1049
+ <port id="0" precision="I64" names="/Constant_2_output_0">
1050
+ <dim>1</dim>
1051
+ </port>
1052
+ </output>
1053
+ </layer>
1054
+ <layer id="58" name="Broadcast_781" type="Const" version="opset1">
1055
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1056
+ <output>
1057
+ <port id="0" precision="I64">
1058
+ <dim>1</dim>
1059
+ </port>
1060
+ </output>
1061
+ </layer>
1062
+ <layer id="59" name="/Slice" type="StridedSlice" version="opset1">
1063
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1064
+ <input>
1065
+ <port id="0" precision="I64">
1066
+ <dim>4</dim>
1067
+ </port>
1068
+ <port id="1" precision="I64">
1069
+ <dim>1</dim>
1070
+ </port>
1071
+ <port id="2" precision="I64">
1072
+ <dim>1</dim>
1073
+ </port>
1074
+ <port id="3" precision="I64">
1075
+ <dim>1</dim>
1076
+ </port>
1077
+ </input>
1078
+ <output>
1079
+ <port id="4" precision="I64" names="/Slice_output_0">
1080
+ <dim>2</dim>
1081
+ </port>
1082
+ </output>
1083
+ </layer>
1084
+ <layer id="60" name="/Constant_3" type="Const" version="opset1">
1085
+ <data element_type="i64" shape="2" offset="2728472" size="16" />
1086
+ <output>
1087
+ <port id="0" precision="I64" names="/Constant_3_output_0">
1088
+ <dim>2</dim>
1089
+ </port>
1090
+ </output>
1091
+ </layer>
1092
+ <layer id="61" name="/Concat" type="Concat" version="opset1">
1093
+ <data axis="0" />
1094
+ <input>
1095
+ <port id="0" precision="I64">
1096
+ <dim>2</dim>
1097
+ </port>
1098
+ <port id="1" precision="I64">
1099
+ <dim>2</dim>
1100
+ </port>
1101
+ </input>
1102
+ <output>
1103
+ <port id="2" precision="I64" names="/Concat_output_0">
1104
+ <dim>4</dim>
1105
+ </port>
1106
+ </output>
1107
+ </layer>
1108
+ <layer id="62" name="Convert_804" type="Convert" version="opset1">
1109
+ <data destination_type="f32" />
1110
+ <input>
1111
+ <port id="0" precision="I64">
1112
+ <dim>4</dim>
1113
+ </port>
1114
+ </input>
1115
+ <output>
1116
+ <port id="1" precision="FP32">
1117
+ <dim>4</dim>
1118
+ </port>
1119
+ </output>
1120
+ </layer>
1121
+ <layer id="63" name="Convert_803" type="Convert" version="opset1">
1122
+ <data destination_type="f32" />
1123
+ <input>
1124
+ <port id="0" precision="I64">
1125
+ <dim>4</dim>
1126
+ </port>
1127
+ </input>
1128
+ <output>
1129
+ <port id="1" precision="FP32">
1130
+ <dim>4</dim>
1131
+ </port>
1132
+ </output>
1133
+ </layer>
1134
+ <layer id="64" name="Divide_805" type="Divide" version="opset1">
1135
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1136
+ <input>
1137
+ <port id="0" precision="FP32">
1138
+ <dim>4</dim>
1139
+ </port>
1140
+ <port id="1" precision="FP32">
1141
+ <dim>4</dim>
1142
+ </port>
1143
+ </input>
1144
+ <output>
1145
+ <port id="2" precision="FP32">
1146
+ <dim>4</dim>
1147
+ </port>
1148
+ </output>
1149
+ </layer>
1150
+ <layer id="65" name="Constant_6582" type="Const" version="opset1">
1151
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1152
+ <output>
1153
+ <port id="0" precision="FP32">
1154
+ <dim>1</dim>
1155
+ </port>
1156
+ </output>
1157
+ </layer>
1158
+ <layer id="66" name="Add_807" type="Add" version="opset1">
1159
+ <data auto_broadcast="numpy" />
1160
+ <input>
1161
+ <port id="0" precision="FP32">
1162
+ <dim>4</dim>
1163
+ </port>
1164
+ <port id="1" precision="FP32">
1165
+ <dim>1</dim>
1166
+ </port>
1167
+ </input>
1168
+ <output>
1169
+ <port id="2" precision="FP32">
1170
+ <dim>4</dim>
1171
+ </port>
1172
+ </output>
1173
+ </layer>
1174
+ <layer id="67" name="/Resize" type="Interpolate" version="opset4">
1175
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1176
+ <input>
1177
+ <port id="0" precision="FP32">
1178
+ <dim>1</dim>
1179
+ <dim>128</dim>
1180
+ <dim>32</dim>
1181
+ <dim>32</dim>
1182
+ </port>
1183
+ <port id="1" precision="I64">
1184
+ <dim>4</dim>
1185
+ </port>
1186
+ <port id="2" precision="FP32">
1187
+ <dim>4</dim>
1188
+ </port>
1189
+ </input>
1190
+ <output>
1191
+ <port id="3" precision="FP32" names="/Resize_output_0">
1192
+ <dim>1</dim>
1193
+ <dim>128</dim>
1194
+ <dim>64</dim>
1195
+ <dim>64</dim>
1196
+ </port>
1197
+ </output>
1198
+ </layer>
1199
+ <layer id="68" name="/Concat_1" type="Concat" version="opset1">
1200
+ <data axis="1" />
1201
+ <input>
1202
+ <port id="0" precision="FP32">
1203
+ <dim>1</dim>
1204
+ <dim>64</dim>
1205
+ <dim>64</dim>
1206
+ <dim>64</dim>
1207
+ </port>
1208
+ <port id="1" precision="FP32">
1209
+ <dim>1</dim>
1210
+ <dim>128</dim>
1211
+ <dim>64</dim>
1212
+ <dim>64</dim>
1213
+ </port>
1214
+ </input>
1215
+ <output>
1216
+ <port id="2" precision="FP32" names="/Concat_1_output_0">
1217
+ <dim>1</dim>
1218
+ <dim>192</dim>
1219
+ <dim>64</dim>
1220
+ <dim>64</dim>
1221
+ </port>
1222
+ </output>
1223
+ </layer>
1224
+ <layer id="69" name="onnx::Conv_290" type="Const" version="opset1">
1225
+ <data element_type="f32" shape="256, 128, 3, 3" offset="2728492" size="1179648" />
1226
+ <output>
1227
+ <port id="0" precision="FP32" names="onnx::Conv_290">
1228
+ <dim>256</dim>
1229
+ <dim>128</dim>
1230
+ <dim>3</dim>
1231
+ <dim>3</dim>
1232
+ </port>
1233
+ </output>
1234
+ </layer>
1235
+ <layer id="70" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1236
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1237
+ <input>
1238
+ <port id="0" precision="FP32">
1239
+ <dim>1</dim>
1240
+ <dim>128</dim>
1241
+ <dim>32</dim>
1242
+ <dim>32</dim>
1243
+ </port>
1244
+ <port id="1" precision="FP32">
1245
+ <dim>256</dim>
1246
+ <dim>128</dim>
1247
+ <dim>3</dim>
1248
+ <dim>3</dim>
1249
+ </port>
1250
+ </input>
1251
+ <output>
1252
+ <port id="2" precision="FP32">
1253
+ <dim>1</dim>
1254
+ <dim>256</dim>
1255
+ <dim>16</dim>
1256
+ <dim>16</dim>
1257
+ </port>
1258
+ </output>
1259
+ </layer>
1260
+ <layer id="71" name="Reshape_549" type="Const" version="opset1">
1261
+ <data element_type="f32" shape="1, 256, 1, 1" offset="3908140" size="1024" />
1262
+ <output>
1263
+ <port id="0" precision="FP32">
1264
+ <dim>1</dim>
1265
+ <dim>256</dim>
1266
+ <dim>1</dim>
1267
+ <dim>1</dim>
1268
+ </port>
1269
+ </output>
1270
+ </layer>
1271
+ <layer id="72" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" type="Add" version="opset1">
1272
+ <data auto_broadcast="numpy" />
1273
+ <input>
1274
+ <port id="0" precision="FP32">
1275
+ <dim>1</dim>
1276
+ <dim>256</dim>
1277
+ <dim>16</dim>
1278
+ <dim>16</dim>
1279
+ </port>
1280
+ <port id="1" precision="FP32">
1281
+ <dim>1</dim>
1282
+ <dim>256</dim>
1283
+ <dim>1</dim>
1284
+ <dim>1</dim>
1285
+ </port>
1286
+ </input>
1287
+ <output>
1288
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0">
1289
+ <dim>1</dim>
1290
+ <dim>256</dim>
1291
+ <dim>16</dim>
1292
+ <dim>16</dim>
1293
+ </port>
1294
+ </output>
1295
+ </layer>
1296
+ <layer id="73" name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" type="ReLU" version="opset1">
1297
+ <input>
1298
+ <port id="0" precision="FP32">
1299
+ <dim>1</dim>
1300
+ <dim>256</dim>
1301
+ <dim>16</dim>
1302
+ <dim>16</dim>
1303
+ </port>
1304
+ </input>
1305
+ <output>
1306
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0">
1307
+ <dim>1</dim>
1308
+ <dim>256</dim>
1309
+ <dim>16</dim>
1310
+ <dim>16</dim>
1311
+ </port>
1312
+ </output>
1313
+ </layer>
1314
+ <layer id="74" name="onnx::Conv_293" type="Const" version="opset1">
1315
+ <data element_type="f32" shape="256, 256, 3, 3" offset="3909164" size="2359296" />
1316
+ <output>
1317
+ <port id="0" precision="FP32" names="onnx::Conv_293">
1318
+ <dim>256</dim>
1319
+ <dim>256</dim>
1320
+ <dim>3</dim>
1321
+ <dim>3</dim>
1322
+ </port>
1323
+ </output>
1324
+ </layer>
1325
+ <layer id="75" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1326
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1327
+ <input>
1328
+ <port id="0" precision="FP32">
1329
+ <dim>1</dim>
1330
+ <dim>256</dim>
1331
+ <dim>16</dim>
1332
+ <dim>16</dim>
1333
+ </port>
1334
+ <port id="1" precision="FP32">
1335
+ <dim>256</dim>
1336
+ <dim>256</dim>
1337
+ <dim>3</dim>
1338
+ <dim>3</dim>
1339
+ </port>
1340
+ </input>
1341
+ <output>
1342
+ <port id="2" precision="FP32">
1343
+ <dim>1</dim>
1344
+ <dim>256</dim>
1345
+ <dim>16</dim>
1346
+ <dim>16</dim>
1347
+ </port>
1348
+ </output>
1349
+ </layer>
1350
+ <layer id="76" name="Reshape_598" type="Const" version="opset1">
1351
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6268460" size="1024" />
1352
+ <output>
1353
+ <port id="0" precision="FP32">
1354
+ <dim>1</dim>
1355
+ <dim>256</dim>
1356
+ <dim>1</dim>
1357
+ <dim>1</dim>
1358
+ </port>
1359
+ </output>
1360
+ </layer>
1361
+ <layer id="77" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" type="Add" version="opset1">
1362
+ <data auto_broadcast="numpy" />
1363
+ <input>
1364
+ <port id="0" precision="FP32">
1365
+ <dim>1</dim>
1366
+ <dim>256</dim>
1367
+ <dim>16</dim>
1368
+ <dim>16</dim>
1369
+ </port>
1370
+ <port id="1" precision="FP32">
1371
+ <dim>1</dim>
1372
+ <dim>256</dim>
1373
+ <dim>1</dim>
1374
+ <dim>1</dim>
1375
+ </port>
1376
+ </input>
1377
+ <output>
1378
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0">
1379
+ <dim>1</dim>
1380
+ <dim>256</dim>
1381
+ <dim>16</dim>
1382
+ <dim>16</dim>
1383
+ </port>
1384
+ </output>
1385
+ </layer>
1386
+ <layer id="78" name="onnx::Conv_296" type="Const" version="opset1">
1387
+ <data element_type="f32" shape="256, 128, 1, 1" offset="6269484" size="131072" />
1388
+ <output>
1389
+ <port id="0" precision="FP32" names="onnx::Conv_296">
1390
+ <dim>256</dim>
1391
+ <dim>128</dim>
1392
+ <dim>1</dim>
1393
+ <dim>1</dim>
1394
+ </port>
1395
+ </output>
1396
+ </layer>
1397
+ <layer id="79" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
1398
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
1399
+ <input>
1400
+ <port id="0" precision="FP32">
1401
+ <dim>1</dim>
1402
+ <dim>128</dim>
1403
+ <dim>32</dim>
1404
+ <dim>32</dim>
1405
+ </port>
1406
+ <port id="1" precision="FP32">
1407
+ <dim>256</dim>
1408
+ <dim>128</dim>
1409
+ <dim>1</dim>
1410
+ <dim>1</dim>
1411
+ </port>
1412
+ </input>
1413
+ <output>
1414
+ <port id="2" precision="FP32">
1415
+ <dim>1</dim>
1416
+ <dim>256</dim>
1417
+ <dim>16</dim>
1418
+ <dim>16</dim>
1419
+ </port>
1420
+ </output>
1421
+ </layer>
1422
+ <layer id="80" name="Reshape_646" type="Const" version="opset1">
1423
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6400556" size="1024" />
1424
+ <output>
1425
+ <port id="0" precision="FP32">
1426
+ <dim>1</dim>
1427
+ <dim>256</dim>
1428
+ <dim>1</dim>
1429
+ <dim>1</dim>
1430
+ </port>
1431
+ </output>
1432
+ </layer>
1433
+ <layer id="81" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" type="Add" version="opset1">
1434
+ <data auto_broadcast="numpy" />
1435
+ <input>
1436
+ <port id="0" precision="FP32">
1437
+ <dim>1</dim>
1438
+ <dim>256</dim>
1439
+ <dim>16</dim>
1440
+ <dim>16</dim>
1441
+ </port>
1442
+ <port id="1" precision="FP32">
1443
+ <dim>1</dim>
1444
+ <dim>256</dim>
1445
+ <dim>1</dim>
1446
+ <dim>1</dim>
1447
+ </port>
1448
+ </input>
1449
+ <output>
1450
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0">
1451
+ <dim>1</dim>
1452
+ <dim>256</dim>
1453
+ <dim>16</dim>
1454
+ <dim>16</dim>
1455
+ </port>
1456
+ </output>
1457
+ </layer>
1458
+ <layer id="82" name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" type="Add" version="opset1">
1459
+ <data auto_broadcast="numpy" />
1460
+ <input>
1461
+ <port id="0" precision="FP32">
1462
+ <dim>1</dim>
1463
+ <dim>256</dim>
1464
+ <dim>16</dim>
1465
+ <dim>16</dim>
1466
+ </port>
1467
+ <port id="1" precision="FP32">
1468
+ <dim>1</dim>
1469
+ <dim>256</dim>
1470
+ <dim>16</dim>
1471
+ <dim>16</dim>
1472
+ </port>
1473
+ </input>
1474
+ <output>
1475
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0">
1476
+ <dim>1</dim>
1477
+ <dim>256</dim>
1478
+ <dim>16</dim>
1479
+ <dim>16</dim>
1480
+ </port>
1481
+ </output>
1482
+ </layer>
1483
+ <layer id="83" name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" type="ReLU" version="opset1">
1484
+ <input>
1485
+ <port id="0" precision="FP32">
1486
+ <dim>1</dim>
1487
+ <dim>256</dim>
1488
+ <dim>16</dim>
1489
+ <dim>16</dim>
1490
+ </port>
1491
+ </input>
1492
+ <output>
1493
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0">
1494
+ <dim>1</dim>
1495
+ <dim>256</dim>
1496
+ <dim>16</dim>
1497
+ <dim>16</dim>
1498
+ </port>
1499
+ </output>
1500
+ </layer>
1501
+ <layer id="84" name="onnx::Conv_299" type="Const" version="opset1">
1502
+ <data element_type="f32" shape="256, 256, 3, 3" offset="6401580" size="2359296" />
1503
+ <output>
1504
+ <port id="0" precision="FP32" names="onnx::Conv_299">
1505
+ <dim>256</dim>
1506
+ <dim>256</dim>
1507
+ <dim>3</dim>
1508
+ <dim>3</dim>
1509
+ </port>
1510
+ </output>
1511
+ </layer>
1512
+ <layer id="85" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1513
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1514
+ <input>
1515
+ <port id="0" precision="FP32">
1516
+ <dim>1</dim>
1517
+ <dim>256</dim>
1518
+ <dim>16</dim>
1519
+ <dim>16</dim>
1520
+ </port>
1521
+ <port id="1" precision="FP32">
1522
+ <dim>256</dim>
1523
+ <dim>256</dim>
1524
+ <dim>3</dim>
1525
+ <dim>3</dim>
1526
+ </port>
1527
+ </input>
1528
+ <output>
1529
+ <port id="2" precision="FP32">
1530
+ <dim>1</dim>
1531
+ <dim>256</dim>
1532
+ <dim>16</dim>
1533
+ <dim>16</dim>
1534
+ </port>
1535
+ </output>
1536
+ </layer>
1537
+ <layer id="86" name="Reshape_696" type="Const" version="opset1">
1538
+ <data element_type="f32" shape="1, 256, 1, 1" offset="8760876" size="1024" />
1539
+ <output>
1540
+ <port id="0" precision="FP32">
1541
+ <dim>1</dim>
1542
+ <dim>256</dim>
1543
+ <dim>1</dim>
1544
+ <dim>1</dim>
1545
+ </port>
1546
+ </output>
1547
+ </layer>
1548
+ <layer id="87" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" type="Add" version="opset1">
1549
+ <data auto_broadcast="numpy" />
1550
+ <input>
1551
+ <port id="0" precision="FP32">
1552
+ <dim>1</dim>
1553
+ <dim>256</dim>
1554
+ <dim>16</dim>
1555
+ <dim>16</dim>
1556
+ </port>
1557
+ <port id="1" precision="FP32">
1558
+ <dim>1</dim>
1559
+ <dim>256</dim>
1560
+ <dim>1</dim>
1561
+ <dim>1</dim>
1562
+ </port>
1563
+ </input>
1564
+ <output>
1565
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0">
1566
+ <dim>1</dim>
1567
+ <dim>256</dim>
1568
+ <dim>16</dim>
1569
+ <dim>16</dim>
1570
+ </port>
1571
+ </output>
1572
+ </layer>
1573
+ <layer id="88" name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" type="ReLU" version="opset1">
1574
+ <input>
1575
+ <port id="0" precision="FP32">
1576
+ <dim>1</dim>
1577
+ <dim>256</dim>
1578
+ <dim>16</dim>
1579
+ <dim>16</dim>
1580
+ </port>
1581
+ </input>
1582
+ <output>
1583
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0">
1584
+ <dim>1</dim>
1585
+ <dim>256</dim>
1586
+ <dim>16</dim>
1587
+ <dim>16</dim>
1588
+ </port>
1589
+ </output>
1590
+ </layer>
1591
+ <layer id="89" name="onnx::Conv_302" type="Const" version="opset1">
1592
+ <data element_type="f32" shape="256, 256, 3, 3" offset="8761900" size="2359296" />
1593
+ <output>
1594
+ <port id="0" precision="FP32" names="onnx::Conv_302">
1595
+ <dim>256</dim>
1596
+ <dim>256</dim>
1597
+ <dim>3</dim>
1598
+ <dim>3</dim>
1599
+ </port>
1600
+ </output>
1601
+ </layer>
1602
+ <layer id="90" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1603
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1604
+ <input>
1605
+ <port id="0" precision="FP32">
1606
+ <dim>1</dim>
1607
+ <dim>256</dim>
1608
+ <dim>16</dim>
1609
+ <dim>16</dim>
1610
+ </port>
1611
+ <port id="1" precision="FP32">
1612
+ <dim>256</dim>
1613
+ <dim>256</dim>
1614
+ <dim>3</dim>
1615
+ <dim>3</dim>
1616
+ </port>
1617
+ </input>
1618
+ <output>
1619
+ <port id="2" precision="FP32">
1620
+ <dim>1</dim>
1621
+ <dim>256</dim>
1622
+ <dim>16</dim>
1623
+ <dim>16</dim>
1624
+ </port>
1625
+ </output>
1626
+ </layer>
1627
+ <layer id="91" name="Reshape_745" type="Const" version="opset1">
1628
+ <data element_type="f32" shape="1, 256, 1, 1" offset="11121196" size="1024" />
1629
+ <output>
1630
+ <port id="0" precision="FP32">
1631
+ <dim>1</dim>
1632
+ <dim>256</dim>
1633
+ <dim>1</dim>
1634
+ <dim>1</dim>
1635
+ </port>
1636
+ </output>
1637
+ </layer>
1638
+ <layer id="92" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" type="Add" version="opset1">
1639
+ <data auto_broadcast="numpy" />
1640
+ <input>
1641
+ <port id="0" precision="FP32">
1642
+ <dim>1</dim>
1643
+ <dim>256</dim>
1644
+ <dim>16</dim>
1645
+ <dim>16</dim>
1646
+ </port>
1647
+ <port id="1" precision="FP32">
1648
+ <dim>1</dim>
1649
+ <dim>256</dim>
1650
+ <dim>1</dim>
1651
+ <dim>1</dim>
1652
+ </port>
1653
+ </input>
1654
+ <output>
1655
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0">
1656
+ <dim>1</dim>
1657
+ <dim>256</dim>
1658
+ <dim>16</dim>
1659
+ <dim>16</dim>
1660
+ </port>
1661
+ </output>
1662
+ </layer>
1663
+ <layer id="93" name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" type="Add" version="opset1">
1664
+ <data auto_broadcast="numpy" />
1665
+ <input>
1666
+ <port id="0" precision="FP32">
1667
+ <dim>1</dim>
1668
+ <dim>256</dim>
1669
+ <dim>16</dim>
1670
+ <dim>16</dim>
1671
+ </port>
1672
+ <port id="1" precision="FP32">
1673
+ <dim>1</dim>
1674
+ <dim>256</dim>
1675
+ <dim>16</dim>
1676
+ <dim>16</dim>
1677
+ </port>
1678
+ </input>
1679
+ <output>
1680
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0">
1681
+ <dim>1</dim>
1682
+ <dim>256</dim>
1683
+ <dim>16</dim>
1684
+ <dim>16</dim>
1685
+ </port>
1686
+ </output>
1687
+ </layer>
1688
+ <layer id="94" name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" type="ReLU" version="opset1">
1689
+ <input>
1690
+ <port id="0" precision="FP32">
1691
+ <dim>1</dim>
1692
+ <dim>256</dim>
1693
+ <dim>16</dim>
1694
+ <dim>16</dim>
1695
+ </port>
1696
+ </input>
1697
+ <output>
1698
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0">
1699
+ <dim>1</dim>
1700
+ <dim>256</dim>
1701
+ <dim>16</dim>
1702
+ <dim>16</dim>
1703
+ </port>
1704
+ </output>
1705
+ </layer>
1706
+ <layer id="95" name="/Shape_3" type="ShapeOf" version="opset3">
1707
+ <data output_type="i64" />
1708
+ <input>
1709
+ <port id="0" precision="FP32">
1710
+ <dim>1</dim>
1711
+ <dim>256</dim>
1712
+ <dim>16</dim>
1713
+ <dim>16</dim>
1714
+ </port>
1715
+ </input>
1716
+ <output>
1717
+ <port id="1" precision="I64" names="/Shape_3_output_0">
1718
+ <dim>4</dim>
1719
+ </port>
1720
+ </output>
1721
+ </layer>
1722
+ <layer id="96" name="/Constant_9" type="Const" version="opset1">
1723
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1724
+ <output>
1725
+ <port id="0" precision="I64" names="/Constant_9_output_0">
1726
+ <dim>1</dim>
1727
+ </port>
1728
+ </output>
1729
+ </layer>
1730
+ <layer id="97" name="/Constant_10" type="Const" version="opset1">
1731
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1732
+ <output>
1733
+ <port id="0" precision="I64" names="/Constant_10_output_0">
1734
+ <dim>1</dim>
1735
+ </port>
1736
+ </output>
1737
+ </layer>
1738
+ <layer id="98" name="Broadcast_861" type="Const" version="opset1">
1739
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1740
+ <output>
1741
+ <port id="0" precision="I64">
1742
+ <dim>1</dim>
1743
+ </port>
1744
+ </output>
1745
+ </layer>
1746
+ <layer id="99" name="/Slice_1" type="StridedSlice" version="opset1">
1747
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1748
+ <input>
1749
+ <port id="0" precision="I64">
1750
+ <dim>4</dim>
1751
+ </port>
1752
+ <port id="1" precision="I64">
1753
+ <dim>1</dim>
1754
+ </port>
1755
+ <port id="2" precision="I64">
1756
+ <dim>1</dim>
1757
+ </port>
1758
+ <port id="3" precision="I64">
1759
+ <dim>1</dim>
1760
+ </port>
1761
+ </input>
1762
+ <output>
1763
+ <port id="4" precision="I64" names="/Slice_1_output_0">
1764
+ <dim>2</dim>
1765
+ </port>
1766
+ </output>
1767
+ </layer>
1768
+ <layer id="100" name="/Shape_1" type="ShapeOf" version="opset3">
1769
+ <data output_type="i64" />
1770
+ <input>
1771
+ <port id="0" precision="FP32">
1772
+ <dim>1</dim>
1773
+ <dim>192</dim>
1774
+ <dim>64</dim>
1775
+ <dim>64</dim>
1776
+ </port>
1777
+ </input>
1778
+ <output>
1779
+ <port id="1" precision="I64" names="/Shape_1_output_0,/Shape_2_output_0">
1780
+ <dim>4</dim>
1781
+ </port>
1782
+ </output>
1783
+ </layer>
1784
+ <layer id="101" name="Constant_6090" type="Const" version="opset1">
1785
+ <data element_type="i64" shape="2" offset="11122220" size="16" />
1786
+ <output>
1787
+ <port id="0" precision="I64">
1788
+ <dim>2</dim>
1789
+ </port>
1790
+ </output>
1791
+ </layer>
1792
+ <layer id="102" name="Constant_6091" type="Const" version="opset1">
1793
+ <data element_type="i64" shape="" offset="2728448" size="8" />
1794
+ <output>
1795
+ <port id="0" precision="I64" />
1796
+ </output>
1797
+ </layer>
1798
+ <layer id="103" name="Gather_6092" type="Gather" version="opset8">
1799
+ <data batch_dims="0" />
1800
+ <input>
1801
+ <port id="0" precision="I64">
1802
+ <dim>4</dim>
1803
+ </port>
1804
+ <port id="1" precision="I64">
1805
+ <dim>2</dim>
1806
+ </port>
1807
+ <port id="2" precision="I64" />
1808
+ </input>
1809
+ <output>
1810
+ <port id="3" precision="I64" names="/Cast_output_0,/Concat_2_output_0">
1811
+ <dim>2</dim>
1812
+ </port>
1813
+ </output>
1814
+ </layer>
1815
+ <layer id="104" name="/Concat_3" type="Concat" version="opset1">
1816
+ <data axis="0" />
1817
+ <input>
1818
+ <port id="0" precision="I64">
1819
+ <dim>2</dim>
1820
+ </port>
1821
+ <port id="1" precision="I64">
1822
+ <dim>2</dim>
1823
+ </port>
1824
+ </input>
1825
+ <output>
1826
+ <port id="2" precision="I64" names="/Concat_3_output_0">
1827
+ <dim>4</dim>
1828
+ </port>
1829
+ </output>
1830
+ </layer>
1831
+ <layer id="105" name="Convert_884" type="Convert" version="opset1">
1832
+ <data destination_type="f32" />
1833
+ <input>
1834
+ <port id="0" precision="I64">
1835
+ <dim>4</dim>
1836
+ </port>
1837
+ </input>
1838
+ <output>
1839
+ <port id="1" precision="FP32">
1840
+ <dim>4</dim>
1841
+ </port>
1842
+ </output>
1843
+ </layer>
1844
+ <layer id="106" name="Convert_883" type="Convert" version="opset1">
1845
+ <data destination_type="f32" />
1846
+ <input>
1847
+ <port id="0" precision="I64">
1848
+ <dim>4</dim>
1849
+ </port>
1850
+ </input>
1851
+ <output>
1852
+ <port id="1" precision="FP32">
1853
+ <dim>4</dim>
1854
+ </port>
1855
+ </output>
1856
+ </layer>
1857
+ <layer id="107" name="Divide_885" type="Divide" version="opset1">
1858
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1859
+ <input>
1860
+ <port id="0" precision="FP32">
1861
+ <dim>4</dim>
1862
+ </port>
1863
+ <port id="1" precision="FP32">
1864
+ <dim>4</dim>
1865
+ </port>
1866
+ </input>
1867
+ <output>
1868
+ <port id="2" precision="FP32">
1869
+ <dim>4</dim>
1870
+ </port>
1871
+ </output>
1872
+ </layer>
1873
+ <layer id="108" name="Constant_6583" type="Const" version="opset1">
1874
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1875
+ <output>
1876
+ <port id="0" precision="FP32">
1877
+ <dim>1</dim>
1878
+ </port>
1879
+ </output>
1880
+ </layer>
1881
+ <layer id="109" name="Add_887" type="Add" version="opset1">
1882
+ <data auto_broadcast="numpy" />
1883
+ <input>
1884
+ <port id="0" precision="FP32">
1885
+ <dim>4</dim>
1886
+ </port>
1887
+ <port id="1" precision="FP32">
1888
+ <dim>1</dim>
1889
+ </port>
1890
+ </input>
1891
+ <output>
1892
+ <port id="2" precision="FP32">
1893
+ <dim>4</dim>
1894
+ </port>
1895
+ </output>
1896
+ </layer>
1897
+ <layer id="110" name="/Resize_1" type="Interpolate" version="opset4">
1898
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1899
+ <input>
1900
+ <port id="0" precision="FP32">
1901
+ <dim>1</dim>
1902
+ <dim>256</dim>
1903
+ <dim>16</dim>
1904
+ <dim>16</dim>
1905
+ </port>
1906
+ <port id="1" precision="I64">
1907
+ <dim>4</dim>
1908
+ </port>
1909
+ <port id="2" precision="FP32">
1910
+ <dim>4</dim>
1911
+ </port>
1912
+ </input>
1913
+ <output>
1914
+ <port id="3" precision="FP32" names="/Resize_1_output_0">
1915
+ <dim>1</dim>
1916
+ <dim>256</dim>
1917
+ <dim>64</dim>
1918
+ <dim>64</dim>
1919
+ </port>
1920
+ </output>
1921
+ </layer>
1922
+ <layer id="111" name="/Concat_4" type="Concat" version="opset1">
1923
+ <data axis="1" />
1924
+ <input>
1925
+ <port id="0" precision="FP32">
1926
+ <dim>1</dim>
1927
+ <dim>192</dim>
1928
+ <dim>64</dim>
1929
+ <dim>64</dim>
1930
+ </port>
1931
+ <port id="1" precision="FP32">
1932
+ <dim>1</dim>
1933
+ <dim>256</dim>
1934
+ <dim>64</dim>
1935
+ <dim>64</dim>
1936
+ </port>
1937
+ </input>
1938
+ <output>
1939
+ <port id="2" precision="FP32" names="/Concat_4_output_0">
1940
+ <dim>1</dim>
1941
+ <dim>448</dim>
1942
+ <dim>64</dim>
1943
+ <dim>64</dim>
1944
+ </port>
1945
+ </output>
1946
+ </layer>
1947
+ <layer id="112" name="onnx::Gather_308" type="Const" version="opset1">
1948
+ <data element_type="i64" shape="100" offset="11122236" size="800" />
1949
+ <output>
1950
+ <port id="0" precision="I64" names="onnx::Gather_308">
1951
+ <dim>100</dim>
1952
+ </port>
1953
+ </output>
1954
+ </layer>
1955
+ <layer id="113" name="Constant_965" type="Const" version="opset1">
1956
+ <data element_type="i64" shape="" offset="2728464" size="8" />
1957
+ <output>
1958
+ <port id="0" precision="I64" />
1959
+ </output>
1960
+ </layer>
1961
+ <layer id="114" name="/Gather_2" type="Gather" version="opset8">
1962
+ <data batch_dims="0" />
1963
+ <input>
1964
+ <port id="0" precision="FP32">
1965
+ <dim>1</dim>
1966
+ <dim>448</dim>
1967
+ <dim>64</dim>
1968
+ <dim>64</dim>
1969
+ </port>
1970
+ <port id="1" precision="I64">
1971
+ <dim>100</dim>
1972
+ </port>
1973
+ <port id="2" precision="I64" />
1974
+ </input>
1975
+ <output>
1976
+ <port id="3" precision="FP32" names="/Gather_2_output_0">
1977
+ <dim>1</dim>
1978
+ <dim>100</dim>
1979
+ <dim>64</dim>
1980
+ <dim>64</dim>
1981
+ </port>
1982
+ </output>
1983
+ </layer>
1984
+ <layer id="115" name="/anomaly_map_generator/Constant" type="Const" version="opset1">
1985
+ <data element_type="i64" shape="3" offset="11123036" size="24" />
1986
+ <output>
1987
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_output_0">
1988
+ <dim>3</dim>
1989
+ </port>
1990
+ </output>
1991
+ </layer>
1992
+ <layer id="116" name="/anomaly_map_generator/Reshape" type="Reshape" version="opset1">
1993
+ <data special_zero="true" />
1994
+ <input>
1995
+ <port id="0" precision="FP32">
1996
+ <dim>1</dim>
1997
+ <dim>100</dim>
1998
+ <dim>64</dim>
1999
+ <dim>64</dim>
2000
+ </port>
2001
+ <port id="1" precision="I64">
2002
+ <dim>3</dim>
2003
+ </port>
2004
+ </input>
2005
+ <output>
2006
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_output_0">
2007
+ <dim>1</dim>
2008
+ <dim>100</dim>
2009
+ <dim>4096</dim>
2010
+ </port>
2011
+ </output>
2012
+ </layer>
2013
+ <layer id="117" name="Constant_6584" type="Const" version="opset1">
2014
+ <data element_type="f32" shape="1, 100, 4096" offset="11123060" size="1638400" />
2015
+ <output>
2016
+ <port id="0" precision="FP32">
2017
+ <dim>1</dim>
2018
+ <dim>100</dim>
2019
+ <dim>4096</dim>
2020
+ </port>
2021
+ </output>
2022
+ </layer>
2023
+ <layer id="118" name="/anomaly_map_generator/Sub" type="Subtract" version="opset1">
2024
+ <data auto_broadcast="numpy" />
2025
+ <input>
2026
+ <port id="0" precision="FP32">
2027
+ <dim>1</dim>
2028
+ <dim>100</dim>
2029
+ <dim>4096</dim>
2030
+ </port>
2031
+ <port id="1" precision="FP32">
2032
+ <dim>1</dim>
2033
+ <dim>100</dim>
2034
+ <dim>4096</dim>
2035
+ </port>
2036
+ </input>
2037
+ <output>
2038
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Sub_output_0">
2039
+ <dim>1</dim>
2040
+ <dim>100</dim>
2041
+ <dim>4096</dim>
2042
+ </port>
2043
+ </output>
2044
+ </layer>
2045
+ <layer id="119" name="Constant_975" type="Const" version="opset1">
2046
+ <data element_type="i64" shape="3" offset="12761460" size="24" />
2047
+ <output>
2048
+ <port id="0" precision="I64">
2049
+ <dim>3</dim>
2050
+ </port>
2051
+ </output>
2052
+ </layer>
2053
+ <layer id="120" name="/anomaly_map_generator/Transpose" type="Transpose" version="opset1">
2054
+ <input>
2055
+ <port id="0" precision="FP32">
2056
+ <dim>1</dim>
2057
+ <dim>100</dim>
2058
+ <dim>4096</dim>
2059
+ </port>
2060
+ <port id="1" precision="I64">
2061
+ <dim>3</dim>
2062
+ </port>
2063
+ </input>
2064
+ <output>
2065
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Transpose_output_0">
2066
+ <dim>4096</dim>
2067
+ <dim>1</dim>
2068
+ <dim>100</dim>
2069
+ </port>
2070
+ </output>
2071
+ </layer>
2072
+ <layer id="121" name="onnx::MatMul_310" type="Const" version="opset1">
2073
+ <data element_type="f32" shape="4096, 100, 100" offset="12761484" size="163840000" />
2074
+ <output>
2075
+ <port id="0" precision="FP32" names="onnx::MatMul_310">
2076
+ <dim>4096</dim>
2077
+ <dim>100</dim>
2078
+ <dim>100</dim>
2079
+ </port>
2080
+ </output>
2081
+ </layer>
2082
+ <layer id="122" name="/anomaly_map_generator/MatMul" type="MatMul" version="opset1">
2083
+ <data transpose_a="false" transpose_b="false" />
2084
+ <input>
2085
+ <port id="0" precision="FP32">
2086
+ <dim>4096</dim>
2087
+ <dim>1</dim>
2088
+ <dim>100</dim>
2089
+ </port>
2090
+ <port id="1" precision="FP32">
2091
+ <dim>4096</dim>
2092
+ <dim>100</dim>
2093
+ <dim>100</dim>
2094
+ </port>
2095
+ </input>
2096
+ <output>
2097
+ <port id="2" precision="FP32" names="/anomaly_map_generator/MatMul_output_0">
2098
+ <dim>4096</dim>
2099
+ <dim>1</dim>
2100
+ <dim>100</dim>
2101
+ </port>
2102
+ </output>
2103
+ </layer>
2104
+ <layer id="123" name="/anomaly_map_generator/Mul" type="Multiply" version="opset1">
2105
+ <data auto_broadcast="numpy" />
2106
+ <input>
2107
+ <port id="0" precision="FP32">
2108
+ <dim>4096</dim>
2109
+ <dim>1</dim>
2110
+ <dim>100</dim>
2111
+ </port>
2112
+ <port id="1" precision="FP32">
2113
+ <dim>4096</dim>
2114
+ <dim>1</dim>
2115
+ <dim>100</dim>
2116
+ </port>
2117
+ </input>
2118
+ <output>
2119
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Mul_output_0">
2120
+ <dim>4096</dim>
2121
+ <dim>1</dim>
2122
+ <dim>100</dim>
2123
+ </port>
2124
+ </output>
2125
+ </layer>
2126
+ <layer id="124" name="Constant_979" type="Const" version="opset1">
2127
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2128
+ <output>
2129
+ <port id="0" precision="I64">
2130
+ <dim>1</dim>
2131
+ </port>
2132
+ </output>
2133
+ </layer>
2134
+ <layer id="125" name="/anomaly_map_generator/ReduceSum" type="ReduceSum" version="opset1">
2135
+ <data keep_dims="false" />
2136
+ <input>
2137
+ <port id="0" precision="FP32">
2138
+ <dim>4096</dim>
2139
+ <dim>1</dim>
2140
+ <dim>100</dim>
2141
+ </port>
2142
+ <port id="1" precision="I64">
2143
+ <dim>1</dim>
2144
+ </port>
2145
+ </input>
2146
+ <output>
2147
+ <port id="2" precision="FP32" names="/anomaly_map_generator/ReduceSum_output_0">
2148
+ <dim>4096</dim>
2149
+ <dim>1</dim>
2150
+ </port>
2151
+ </output>
2152
+ </layer>
2153
+ <layer id="126" name="/anomaly_map_generator/Constant_1" type="Const" version="opset1">
2154
+ <data element_type="i64" shape="4" offset="176601484" size="32" />
2155
+ <output>
2156
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_1_output_0">
2157
+ <dim>4</dim>
2158
+ </port>
2159
+ </output>
2160
+ </layer>
2161
+ <layer id="127" name="/anomaly_map_generator/Reshape_1" type="Reshape" version="opset1">
2162
+ <data special_zero="true" />
2163
+ <input>
2164
+ <port id="0" precision="FP32">
2165
+ <dim>4096</dim>
2166
+ <dim>1</dim>
2167
+ </port>
2168
+ <port id="1" precision="I64">
2169
+ <dim>4</dim>
2170
+ </port>
2171
+ </input>
2172
+ <output>
2173
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_1_output_0">
2174
+ <dim>1</dim>
2175
+ <dim>1</dim>
2176
+ <dim>64</dim>
2177
+ <dim>64</dim>
2178
+ </port>
2179
+ </output>
2180
+ </layer>
2181
+ <layer id="128" name="/anomaly_map_generator/Clip" type="Clamp" version="opset1">
2182
+ <data min="0" max="3.4028234663852886e+38" />
2183
+ <input>
2184
+ <port id="0" precision="FP32">
2185
+ <dim>1</dim>
2186
+ <dim>1</dim>
2187
+ <dim>64</dim>
2188
+ <dim>64</dim>
2189
+ </port>
2190
+ </input>
2191
+ <output>
2192
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Clip_output_0">
2193
+ <dim>1</dim>
2194
+ <dim>1</dim>
2195
+ <dim>64</dim>
2196
+ <dim>64</dim>
2197
+ </port>
2198
+ </output>
2199
+ </layer>
2200
+ <layer id="129" name="/anomaly_map_generator/Sqrt" type="Sqrt" version="opset1">
2201
+ <input>
2202
+ <port id="0" precision="FP32">
2203
+ <dim>1</dim>
2204
+ <dim>1</dim>
2205
+ <dim>64</dim>
2206
+ <dim>64</dim>
2207
+ </port>
2208
+ </input>
2209
+ <output>
2210
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Sqrt_output_0">
2211
+ <dim>1</dim>
2212
+ <dim>1</dim>
2213
+ <dim>64</dim>
2214
+ <dim>64</dim>
2215
+ </port>
2216
+ </output>
2217
+ </layer>
2218
+ <layer id="130" name="/anomaly_map_generator/Shape" type="ShapeOf" version="opset3">
2219
+ <data output_type="i64" />
2220
+ <input>
2221
+ <port id="0" precision="FP32">
2222
+ <dim>1</dim>
2223
+ <dim>1</dim>
2224
+ <dim>64</dim>
2225
+ <dim>64</dim>
2226
+ </port>
2227
+ </input>
2228
+ <output>
2229
+ <port id="1" precision="I64" names="/anomaly_map_generator/Shape_output_0">
2230
+ <dim>4</dim>
2231
+ </port>
2232
+ </output>
2233
+ </layer>
2234
+ <layer id="131" name="/anomaly_map_generator/Constant_4" type="Const" version="opset1">
2235
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
2236
+ <output>
2237
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_4_output_0">
2238
+ <dim>1</dim>
2239
+ </port>
2240
+ </output>
2241
+ </layer>
2242
+ <layer id="132" name="/anomaly_map_generator/Constant_5" type="Const" version="opset1">
2243
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2244
+ <output>
2245
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_5_output_0">
2246
+ <dim>1</dim>
2247
+ </port>
2248
+ </output>
2249
+ </layer>
2250
+ <layer id="133" name="Broadcast_1002" type="Const" version="opset1">
2251
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
2252
+ <output>
2253
+ <port id="0" precision="I64">
2254
+ <dim>1</dim>
2255
+ </port>
2256
+ </output>
2257
+ </layer>
2258
+ <layer id="134" name="/anomaly_map_generator/Slice" type="StridedSlice" version="opset1">
2259
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
2260
+ <input>
2261
+ <port id="0" precision="I64">
2262
+ <dim>4</dim>
2263
+ </port>
2264
+ <port id="1" precision="I64">
2265
+ <dim>1</dim>
2266
+ </port>
2267
+ <port id="2" precision="I64">
2268
+ <dim>1</dim>
2269
+ </port>
2270
+ <port id="3" precision="I64">
2271
+ <dim>1</dim>
2272
+ </port>
2273
+ </input>
2274
+ <output>
2275
+ <port id="4" precision="I64" names="/anomaly_map_generator/Slice_output_0">
2276
+ <dim>2</dim>
2277
+ </port>
2278
+ </output>
2279
+ </layer>
2280
+ <layer id="135" name="/anomaly_map_generator/Constant_6" type="Const" version="opset1">
2281
+ <data element_type="i64" shape="2" offset="176601516" size="16" />
2282
+ <output>
2283
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_6_output_0">
2284
+ <dim>2</dim>
2285
+ </port>
2286
+ </output>
2287
+ </layer>
2288
+ <layer id="136" name="/anomaly_map_generator/Concat" type="Concat" version="opset1">
2289
+ <data axis="0" />
2290
+ <input>
2291
+ <port id="0" precision="I64">
2292
+ <dim>2</dim>
2293
+ </port>
2294
+ <port id="1" precision="I64">
2295
+ <dim>2</dim>
2296
+ </port>
2297
+ </input>
2298
+ <output>
2299
+ <port id="2" precision="I64" names="/anomaly_map_generator/Concat_output_0">
2300
+ <dim>4</dim>
2301
+ </port>
2302
+ </output>
2303
+ </layer>
2304
+ <layer id="137" name="Convert_1025" type="Convert" version="opset1">
2305
+ <data destination_type="f32" />
2306
+ <input>
2307
+ <port id="0" precision="I64">
2308
+ <dim>4</dim>
2309
+ </port>
2310
+ </input>
2311
+ <output>
2312
+ <port id="1" precision="FP32">
2313
+ <dim>4</dim>
2314
+ </port>
2315
+ </output>
2316
+ </layer>
2317
+ <layer id="138" name="Convert_1024" type="Convert" version="opset1">
2318
+ <data destination_type="f32" />
2319
+ <input>
2320
+ <port id="0" precision="I64">
2321
+ <dim>4</dim>
2322
+ </port>
2323
+ </input>
2324
+ <output>
2325
+ <port id="1" precision="FP32">
2326
+ <dim>4</dim>
2327
+ </port>
2328
+ </output>
2329
+ </layer>
2330
+ <layer id="139" name="Divide_1026" type="Divide" version="opset1">
2331
+ <data auto_broadcast="numpy" m_pythondiv="true" />
2332
+ <input>
2333
+ <port id="0" precision="FP32">
2334
+ <dim>4</dim>
2335
+ </port>
2336
+ <port id="1" precision="FP32">
2337
+ <dim>4</dim>
2338
+ </port>
2339
+ </input>
2340
+ <output>
2341
+ <port id="2" precision="FP32">
2342
+ <dim>4</dim>
2343
+ </port>
2344
+ </output>
2345
+ </layer>
2346
+ <layer id="140" name="Constant_6585" type="Const" version="opset1">
2347
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
2348
+ <output>
2349
+ <port id="0" precision="FP32">
2350
+ <dim>1</dim>
2351
+ </port>
2352
+ </output>
2353
+ </layer>
2354
+ <layer id="141" name="Add_1028" type="Add" version="opset1">
2355
+ <data auto_broadcast="numpy" />
2356
+ <input>
2357
+ <port id="0" precision="FP32">
2358
+ <dim>4</dim>
2359
+ </port>
2360
+ <port id="1" precision="FP32">
2361
+ <dim>1</dim>
2362
+ </port>
2363
+ </input>
2364
+ <output>
2365
+ <port id="2" precision="FP32">
2366
+ <dim>4</dim>
2367
+ </port>
2368
+ </output>
2369
+ </layer>
2370
+ <layer id="142" name="/anomaly_map_generator/Resize" type="Interpolate" version="opset4">
2371
+ <data mode="linear_onnx" shape_calculation_mode="sizes" coordinate_transformation_mode="half_pixel" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
2372
+ <input>
2373
+ <port id="0" precision="FP32">
2374
+ <dim>1</dim>
2375
+ <dim>1</dim>
2376
+ <dim>64</dim>
2377
+ <dim>64</dim>
2378
+ </port>
2379
+ <port id="1" precision="I64">
2380
+ <dim>4</dim>
2381
+ </port>
2382
+ <port id="2" precision="FP32">
2383
+ <dim>4</dim>
2384
+ </port>
2385
+ </input>
2386
+ <output>
2387
+ <port id="3" precision="FP32" names="/anomaly_map_generator/Resize_output_0">
2388
+ <dim>1</dim>
2389
+ <dim>1</dim>
2390
+ <dim>256</dim>
2391
+ <dim>256</dim>
2392
+ </port>
2393
+ </output>
2394
+ </layer>
2395
+ <layer id="143" name="Split_1097.0" type="Const" version="opset1">
2396
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2397
+ <output>
2398
+ <port id="0" precision="I64">
2399
+ <dim>4</dim>
2400
+ </port>
2401
+ </output>
2402
+ </layer>
2403
+ <layer id="144" name="Split_1097.1" type="Const" version="opset1">
2404
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2405
+ <output>
2406
+ <port id="0" precision="I64">
2407
+ <dim>4</dim>
2408
+ </port>
2409
+ </output>
2410
+ </layer>
2411
+ <layer id="145" name="Constant_1095" type="Const" version="opset1">
2412
+ <data element_type="f32" shape="" offset="176601564" size="4" />
2413
+ <output>
2414
+ <port id="0" precision="FP32" />
2415
+ </output>
2416
+ </layer>
2417
+ <layer id="146" name="/anomaly_map_generator/blur/Pad" type="Pad" version="opset1">
2418
+ <data pad_mode="reflect" />
2419
+ <input>
2420
+ <port id="0" precision="FP32">
2421
+ <dim>1</dim>
2422
+ <dim>1</dim>
2423
+ <dim>256</dim>
2424
+ <dim>256</dim>
2425
+ </port>
2426
+ <port id="1" precision="I64">
2427
+ <dim>4</dim>
2428
+ </port>
2429
+ <port id="2" precision="I64">
2430
+ <dim>4</dim>
2431
+ </port>
2432
+ <port id="3" precision="FP32" />
2433
+ </input>
2434
+ <output>
2435
+ <port id="4" precision="FP32" names="/anomaly_map_generator/blur/Pad_output_0">
2436
+ <dim>1</dim>
2437
+ <dim>1</dim>
2438
+ <dim>288</dim>
2439
+ <dim>288</dim>
2440
+ </port>
2441
+ </output>
2442
+ </layer>
2443
+ <layer id="147" name="anomaly_map_generator.blur.kernel" type="Const" version="opset1">
2444
+ <data element_type="f32" shape="1, 1, 33, 33" offset="176601568" size="4356" />
2445
+ <output>
2446
+ <port id="0" precision="FP32" names="anomaly_map_generator.blur.kernel">
2447
+ <dim>1</dim>
2448
+ <dim>1</dim>
2449
+ <dim>33</dim>
2450
+ <dim>33</dim>
2451
+ </port>
2452
+ </output>
2453
+ </layer>
2454
+ <layer id="148" name="/anomaly_map_generator/blur/Conv" type="Convolution" version="opset1">
2455
+ <data strides="1, 1" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
2456
+ <input>
2457
+ <port id="0" precision="FP32">
2458
+ <dim>1</dim>
2459
+ <dim>1</dim>
2460
+ <dim>288</dim>
2461
+ <dim>288</dim>
2462
+ </port>
2463
+ <port id="1" precision="FP32">
2464
+ <dim>1</dim>
2465
+ <dim>1</dim>
2466
+ <dim>33</dim>
2467
+ <dim>33</dim>
2468
+ </port>
2469
+ </input>
2470
+ <output>
2471
+ <port id="2" precision="FP32" names="/anomaly_map_generator/blur/Conv_output_0">
2472
+ <dim>1</dim>
2473
+ <dim>1</dim>
2474
+ <dim>256</dim>
2475
+ <dim>256</dim>
2476
+ </port>
2477
+ </output>
2478
+ </layer>
2479
+ <layer id="149" name="/anomaly_map_generator/blur/Constant_8" type="Const" version="opset1">
2480
+ <data element_type="i64" shape="4" offset="176605924" size="32" />
2481
+ <output>
2482
+ <port id="0" precision="I64" names="/anomaly_map_generator/blur/Constant_8_output_0">
2483
+ <dim>4</dim>
2484
+ </port>
2485
+ </output>
2486
+ </layer>
2487
+ <layer id="150" name="output" type="Reshape" version="opset1">
2488
+ <data special_zero="true" />
2489
+ <input>
2490
+ <port id="0" precision="FP32">
2491
+ <dim>1</dim>
2492
+ <dim>1</dim>
2493
+ <dim>256</dim>
2494
+ <dim>256</dim>
2495
+ </port>
2496
+ <port id="1" precision="I64">
2497
+ <dim>4</dim>
2498
+ </port>
2499
+ </input>
2500
+ <output>
2501
+ <port id="2" precision="FP32" names="output">
2502
+ <dim>1</dim>
2503
+ <dim>1</dim>
2504
+ <dim>256</dim>
2505
+ <dim>256</dim>
2506
+ </port>
2507
+ </output>
2508
+ </layer>
2509
+ <layer id="151" name="output/sink_port_0" type="Result" version="opset1">
2510
+ <input>
2511
+ <port id="0" precision="FP32">
2512
+ <dim>1</dim>
2513
+ <dim>1</dim>
2514
+ <dim>256</dim>
2515
+ <dim>256</dim>
2516
+ </port>
2517
+ </input>
2518
+ </layer>
2519
+ </layers>
2520
+ <edges>
2521
+ <edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
2522
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="1" />
2523
+ <edge from-layer="2" from-port="2" to-layer="4" to-port="0" />
2524
+ <edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
2525
+ <edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
2526
+ <edge from-layer="5" from-port="1" to-layer="6" to-port="0" />
2527
+ <edge from-layer="6" from-port="1" to-layer="8" to-port="0" />
2528
+ <edge from-layer="6" from-port="1" to-layer="16" to-port="1" />
2529
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="1" />
2530
+ <edge from-layer="8" from-port="2" to-layer="10" to-port="0" />
2531
+ <edge from-layer="9" from-port="0" to-layer="10" to-port="1" />
2532
+ <edge from-layer="10" from-port="2" to-layer="11" to-port="0" />
2533
+ <edge from-layer="11" from-port="1" to-layer="13" to-port="0" />
2534
+ <edge from-layer="12" from-port="0" to-layer="13" to-port="1" />
2535
+ <edge from-layer="13" from-port="2" to-layer="15" to-port="0" />
2536
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="1" />
2537
+ <edge from-layer="15" from-port="2" to-layer="16" to-port="0" />
2538
+ <edge from-layer="16" from-port="2" to-layer="17" to-port="0" />
2539
+ <edge from-layer="17" from-port="1" to-layer="19" to-port="0" />
2540
+ <edge from-layer="17" from-port="1" to-layer="27" to-port="1" />
2541
+ <edge from-layer="18" from-port="0" to-layer="19" to-port="1" />
2542
+ <edge from-layer="19" from-port="2" to-layer="21" to-port="0" />
2543
+ <edge from-layer="20" from-port="0" to-layer="21" to-port="1" />
2544
+ <edge from-layer="21" from-port="2" to-layer="22" to-port="0" />
2545
+ <edge from-layer="22" from-port="1" to-layer="24" to-port="0" />
2546
+ <edge from-layer="23" from-port="0" to-layer="24" to-port="1" />
2547
+ <edge from-layer="24" from-port="2" to-layer="26" to-port="0" />
2548
+ <edge from-layer="25" from-port="0" to-layer="26" to-port="1" />
2549
+ <edge from-layer="26" from-port="2" to-layer="27" to-port="0" />
2550
+ <edge from-layer="27" from-port="2" to-layer="28" to-port="0" />
2551
+ <edge from-layer="28" from-port="1" to-layer="30" to-port="0" />
2552
+ <edge from-layer="28" from-port="1" to-layer="39" to-port="0" />
2553
+ <edge from-layer="28" from-port="1" to-layer="68" to-port="0" />
2554
+ <edge from-layer="29" from-port="0" to-layer="30" to-port="1" />
2555
+ <edge from-layer="30" from-port="2" to-layer="32" to-port="0" />
2556
+ <edge from-layer="31" from-port="0" to-layer="32" to-port="1" />
2557
+ <edge from-layer="32" from-port="2" to-layer="33" to-port="0" />
2558
+ <edge from-layer="33" from-port="1" to-layer="35" to-port="0" />
2559
+ <edge from-layer="34" from-port="0" to-layer="35" to-port="1" />
2560
+ <edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
2561
+ <edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
2562
+ <edge from-layer="37" from-port="2" to-layer="42" to-port="0" />
2563
+ <edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
2564
+ <edge from-layer="39" from-port="2" to-layer="41" to-port="0" />
2565
+ <edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
2566
+ <edge from-layer="41" from-port="2" to-layer="42" to-port="1" />
2567
+ <edge from-layer="42" from-port="2" to-layer="43" to-port="0" />
2568
+ <edge from-layer="43" from-port="1" to-layer="45" to-port="0" />
2569
+ <edge from-layer="43" from-port="1" to-layer="53" to-port="1" />
2570
+ <edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
2571
+ <edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
2572
+ <edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
2573
+ <edge from-layer="47" from-port="2" to-layer="48" to-port="0" />
2574
+ <edge from-layer="48" from-port="1" to-layer="50" to-port="0" />
2575
+ <edge from-layer="49" from-port="0" to-layer="50" to-port="1" />
2576
+ <edge from-layer="50" from-port="2" to-layer="52" to-port="0" />
2577
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="1" />
2578
+ <edge from-layer="52" from-port="2" to-layer="53" to-port="0" />
2579
+ <edge from-layer="53" from-port="2" to-layer="54" to-port="0" />
2580
+ <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
2581
+ <edge from-layer="54" from-port="1" to-layer="67" to-port="0" />
2582
+ <edge from-layer="54" from-port="1" to-layer="79" to-port="0" />
2583
+ <edge from-layer="54" from-port="1" to-layer="70" to-port="0" />
2584
+ <edge from-layer="55" from-port="1" to-layer="59" to-port="0" />
2585
+ <edge from-layer="55" from-port="1" to-layer="63" to-port="0" />
2586
+ <edge from-layer="56" from-port="0" to-layer="59" to-port="1" />
2587
+ <edge from-layer="57" from-port="0" to-layer="59" to-port="2" />
2588
+ <edge from-layer="58" from-port="0" to-layer="59" to-port="3" />
2589
+ <edge from-layer="59" from-port="4" to-layer="61" to-port="0" />
2590
+ <edge from-layer="60" from-port="0" to-layer="61" to-port="1" />
2591
+ <edge from-layer="61" from-port="2" to-layer="62" to-port="0" />
2592
+ <edge from-layer="61" from-port="2" to-layer="67" to-port="1" />
2593
+ <edge from-layer="62" from-port="1" to-layer="64" to-port="0" />
2594
+ <edge from-layer="63" from-port="1" to-layer="64" to-port="1" />
2595
+ <edge from-layer="64" from-port="2" to-layer="66" to-port="0" />
2596
+ <edge from-layer="65" from-port="0" to-layer="66" to-port="1" />
2597
+ <edge from-layer="66" from-port="2" to-layer="67" to-port="2" />
2598
+ <edge from-layer="67" from-port="3" to-layer="68" to-port="1" />
2599
+ <edge from-layer="68" from-port="2" to-layer="111" to-port="0" />
2600
+ <edge from-layer="68" from-port="2" to-layer="100" to-port="0" />
2601
+ <edge from-layer="69" from-port="0" to-layer="70" to-port="1" />
2602
+ <edge from-layer="70" from-port="2" to-layer="72" to-port="0" />
2603
+ <edge from-layer="71" from-port="0" to-layer="72" to-port="1" />
2604
+ <edge from-layer="72" from-port="2" to-layer="73" to-port="0" />
2605
+ <edge from-layer="73" from-port="1" to-layer="75" to-port="0" />
2606
+ <edge from-layer="74" from-port="0" to-layer="75" to-port="1" />
2607
+ <edge from-layer="75" from-port="2" to-layer="77" to-port="0" />
2608
+ <edge from-layer="76" from-port="0" to-layer="77" to-port="1" />
2609
+ <edge from-layer="77" from-port="2" to-layer="82" to-port="0" />
2610
+ <edge from-layer="78" from-port="0" to-layer="79" to-port="1" />
2611
+ <edge from-layer="79" from-port="2" to-layer="81" to-port="0" />
2612
+ <edge from-layer="80" from-port="0" to-layer="81" to-port="1" />
2613
+ <edge from-layer="81" from-port="2" to-layer="82" to-port="1" />
2614
+ <edge from-layer="82" from-port="2" to-layer="83" to-port="0" />
2615
+ <edge from-layer="83" from-port="1" to-layer="85" to-port="0" />
2616
+ <edge from-layer="83" from-port="1" to-layer="93" to-port="1" />
2617
+ <edge from-layer="84" from-port="0" to-layer="85" to-port="1" />
2618
+ <edge from-layer="85" from-port="2" to-layer="87" to-port="0" />
2619
+ <edge from-layer="86" from-port="0" to-layer="87" to-port="1" />
2620
+ <edge from-layer="87" from-port="2" to-layer="88" to-port="0" />
2621
+ <edge from-layer="88" from-port="1" to-layer="90" to-port="0" />
2622
+ <edge from-layer="89" from-port="0" to-layer="90" to-port="1" />
2623
+ <edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
2624
+ <edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
2625
+ <edge from-layer="92" from-port="2" to-layer="93" to-port="0" />
2626
+ <edge from-layer="93" from-port="2" to-layer="94" to-port="0" />
2627
+ <edge from-layer="94" from-port="1" to-layer="95" to-port="0" />
2628
+ <edge from-layer="94" from-port="1" to-layer="110" to-port="0" />
2629
+ <edge from-layer="95" from-port="1" to-layer="99" to-port="0" />
2630
+ <edge from-layer="95" from-port="1" to-layer="106" to-port="0" />
2631
+ <edge from-layer="96" from-port="0" to-layer="99" to-port="1" />
2632
+ <edge from-layer="97" from-port="0" to-layer="99" to-port="2" />
2633
+ <edge from-layer="98" from-port="0" to-layer="99" to-port="3" />
2634
+ <edge from-layer="99" from-port="4" to-layer="104" to-port="0" />
2635
+ <edge from-layer="100" from-port="1" to-layer="103" to-port="0" />
2636
+ <edge from-layer="101" from-port="0" to-layer="103" to-port="1" />
2637
+ <edge from-layer="102" from-port="0" to-layer="103" to-port="2" />
2638
+ <edge from-layer="103" from-port="3" to-layer="104" to-port="1" />
2639
+ <edge from-layer="104" from-port="2" to-layer="105" to-port="0" />
2640
+ <edge from-layer="104" from-port="2" to-layer="110" to-port="1" />
2641
+ <edge from-layer="105" from-port="1" to-layer="107" to-port="0" />
2642
+ <edge from-layer="106" from-port="1" to-layer="107" to-port="1" />
2643
+ <edge from-layer="107" from-port="2" to-layer="109" to-port="0" />
2644
+ <edge from-layer="108" from-port="0" to-layer="109" to-port="1" />
2645
+ <edge from-layer="109" from-port="2" to-layer="110" to-port="2" />
2646
+ <edge from-layer="110" from-port="3" to-layer="111" to-port="1" />
2647
+ <edge from-layer="111" from-port="2" to-layer="114" to-port="0" />
2648
+ <edge from-layer="112" from-port="0" to-layer="114" to-port="1" />
2649
+ <edge from-layer="113" from-port="0" to-layer="114" to-port="2" />
2650
+ <edge from-layer="114" from-port="3" to-layer="116" to-port="0" />
2651
+ <edge from-layer="115" from-port="0" to-layer="116" to-port="1" />
2652
+ <edge from-layer="116" from-port="2" to-layer="118" to-port="0" />
2653
+ <edge from-layer="117" from-port="0" to-layer="118" to-port="1" />
2654
+ <edge from-layer="118" from-port="2" to-layer="120" to-port="0" />
2655
+ <edge from-layer="119" from-port="0" to-layer="120" to-port="1" />
2656
+ <edge from-layer="120" from-port="2" to-layer="122" to-port="0" />
2657
+ <edge from-layer="120" from-port="2" to-layer="123" to-port="1" />
2658
+ <edge from-layer="121" from-port="0" to-layer="122" to-port="1" />
2659
+ <edge from-layer="122" from-port="2" to-layer="123" to-port="0" />
2660
+ <edge from-layer="123" from-port="2" to-layer="125" to-port="0" />
2661
+ <edge from-layer="124" from-port="0" to-layer="125" to-port="1" />
2662
+ <edge from-layer="125" from-port="2" to-layer="127" to-port="0" />
2663
+ <edge from-layer="126" from-port="0" to-layer="127" to-port="1" />
2664
+ <edge from-layer="127" from-port="2" to-layer="128" to-port="0" />
2665
+ <edge from-layer="128" from-port="1" to-layer="129" to-port="0" />
2666
+ <edge from-layer="129" from-port="1" to-layer="130" to-port="0" />
2667
+ <edge from-layer="129" from-port="1" to-layer="142" to-port="0" />
2668
+ <edge from-layer="130" from-port="1" to-layer="134" to-port="0" />
2669
+ <edge from-layer="130" from-port="1" to-layer="138" to-port="0" />
2670
+ <edge from-layer="131" from-port="0" to-layer="134" to-port="1" />
2671
+ <edge from-layer="132" from-port="0" to-layer="134" to-port="2" />
2672
+ <edge from-layer="133" from-port="0" to-layer="134" to-port="3" />
2673
+ <edge from-layer="134" from-port="4" to-layer="136" to-port="0" />
2674
+ <edge from-layer="135" from-port="0" to-layer="136" to-port="1" />
2675
+ <edge from-layer="136" from-port="2" to-layer="137" to-port="0" />
2676
+ <edge from-layer="136" from-port="2" to-layer="142" to-port="1" />
2677
+ <edge from-layer="137" from-port="1" to-layer="139" to-port="0" />
2678
+ <edge from-layer="138" from-port="1" to-layer="139" to-port="1" />
2679
+ <edge from-layer="139" from-port="2" to-layer="141" to-port="0" />
2680
+ <edge from-layer="140" from-port="0" to-layer="141" to-port="1" />
2681
+ <edge from-layer="141" from-port="2" to-layer="142" to-port="2" />
2682
+ <edge from-layer="142" from-port="3" to-layer="146" to-port="0" />
2683
+ <edge from-layer="143" from-port="0" to-layer="146" to-port="1" />
2684
+ <edge from-layer="144" from-port="0" to-layer="146" to-port="2" />
2685
+ <edge from-layer="145" from-port="0" to-layer="146" to-port="3" />
2686
+ <edge from-layer="146" from-port="4" to-layer="148" to-port="0" />
2687
+ <edge from-layer="147" from-port="0" to-layer="148" to-port="1" />
2688
+ <edge from-layer="148" from-port="2" to-layer="150" to-port="0" />
2689
+ <edge from-layer="149" from-port="0" to-layer="150" to-port="1" />
2690
+ <edge from-layer="150" from-port="2" to-layer="151" to-port="0" />
2691
+ </edges>
2692
+ <rt_info>
2693
+ <MO_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2694
+ <Runtime_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2695
+ <conversion_parameters>
2696
+ <framework value="onnx" />
2697
+ <input_model value="DIR\model.onnx" />
2698
+ <model_name value="model" />
2699
+ <output_dir value="DIR" />
2700
+ </conversion_parameters>
2701
+ <legacy_frontend value="False" />
2702
+ </rt_info>
2703
+ </net>
pill/examples/001.png ADDED
pill/examples/007.png ADDED
pill/examples/013.png ADDED
pill/examples/014.png ADDED
pill/examples/019.png ADDED
pill/examples/021.png ADDED
pill/examples/022.png ADDED
pill/run/weights/lightning/model.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e38efa86026af35d3ccf5e76916f36a53146f42f60d50f58d24bfd759b54952
3
+ size 202329404
pill/run/weights/openvino/metadata.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "task": "segmentation",
3
+ "transform": {
4
+ "__version__": "1.3.0",
5
+ "transform": {
6
+ "__class_fullname__": "Compose",
7
+ "p": 1.0,
8
+ "transforms": [
9
+ {
10
+ "__class_fullname__": "Resize",
11
+ "always_apply": true,
12
+ "p": 1,
13
+ "height": 256,
14
+ "width": 256,
15
+ "interpolation": 1
16
+ },
17
+ {
18
+ "__class_fullname__": "Normalize",
19
+ "always_apply": false,
20
+ "p": 1.0,
21
+ "mean": [
22
+ 0.485,
23
+ 0.456,
24
+ 0.406
25
+ ],
26
+ "std": [
27
+ 0.229,
28
+ 0.224,
29
+ 0.225
30
+ ],
31
+ "max_pixel_value": 255.0
32
+ },
33
+ {
34
+ "__class_fullname__": "ToTensorV2",
35
+ "always_apply": true,
36
+ "p": 1.0,
37
+ "transpose_mask": false
38
+ }
39
+ ],
40
+ "bbox_params": null,
41
+ "keypoint_params": null,
42
+ "additional_targets": {
43
+ "image": "image",
44
+ "depth_image": "image"
45
+ }
46
+ }
47
+ },
48
+ "image_threshold": 11.986470222473145,
49
+ "pixel_threshold": 9.849000930786133,
50
+ "min": 1.8367986679077148,
51
+ "max": 28.677541732788086
52
+ }
pill/run/weights/openvino/model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e12fc7290827c23dd6ed40fb3bd48f6787339c6d8fe8f78563af14da79409dea
3
+ size 176605956
pill/run/weights/openvino/model.mapping ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <mapping>
3
+ <map>
4
+ <framework name="input" output_port_id="input" />
5
+ <IR name="input" output_port_id="0" />
6
+ </map>
7
+ <map>
8
+ <framework name="onnx::Conv_260" output_port_id="onnx::Conv_260" />
9
+ <IR name="onnx::Conv_260" output_port_id="0" />
10
+ </map>
11
+ <map>
12
+ <framework name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/conv1/Conv_output_0" />
13
+ <IR name="/feature_extractor/feature_extractor/conv1/Conv" output_port_id="2" />
14
+ </map>
15
+ <map>
16
+ <framework name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="/feature_extractor/feature_extractor/act1/Relu_output_0" />
17
+ <IR name="/feature_extractor/feature_extractor/act1/Relu" output_port_id="1" />
18
+ </map>
19
+ <map>
20
+ <framework name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0" />
21
+ <IR name="/feature_extractor/feature_extractor/maxpool/MaxPool" output_port_id="1" />
22
+ </map>
23
+ <map>
24
+ <framework name="onnx::Conv_263" output_port_id="onnx::Conv_263" />
25
+ <IR name="onnx::Conv_263" output_port_id="0" />
26
+ </map>
27
+ <map>
28
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0" />
29
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" output_port_id="2" />
30
+ </map>
31
+ <map>
32
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0" />
33
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" output_port_id="1" />
34
+ </map>
35
+ <map>
36
+ <framework name="onnx::Conv_266" output_port_id="onnx::Conv_266" />
37
+ <IR name="onnx::Conv_266" output_port_id="0" />
38
+ </map>
39
+ <map>
40
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0" />
41
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" output_port_id="2" />
42
+ </map>
43
+ <map>
44
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0" />
45
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" output_port_id="2" />
46
+ </map>
47
+ <map>
48
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0" />
49
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" output_port_id="1" />
50
+ </map>
51
+ <map>
52
+ <framework name="onnx::Conv_269" output_port_id="onnx::Conv_269" />
53
+ <IR name="onnx::Conv_269" output_port_id="0" />
54
+ </map>
55
+ <map>
56
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0" />
57
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" output_port_id="2" />
58
+ </map>
59
+ <map>
60
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0" />
61
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" output_port_id="1" />
62
+ </map>
63
+ <map>
64
+ <framework name="onnx::Conv_272" output_port_id="onnx::Conv_272" />
65
+ <IR name="onnx::Conv_272" output_port_id="0" />
66
+ </map>
67
+ <map>
68
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0" />
69
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" output_port_id="2" />
70
+ </map>
71
+ <map>
72
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0" />
73
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" output_port_id="2" />
74
+ </map>
75
+ <map>
76
+ <framework name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0" />
77
+ <IR name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" output_port_id="1" />
78
+ </map>
79
+ <map>
80
+ <framework name="onnx::Conv_275" output_port_id="onnx::Conv_275" />
81
+ <IR name="onnx::Conv_275" output_port_id="0" />
82
+ </map>
83
+ <map>
84
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0" />
85
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" output_port_id="2" />
86
+ </map>
87
+ <map>
88
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0" />
89
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" output_port_id="1" />
90
+ </map>
91
+ <map>
92
+ <framework name="onnx::Conv_278" output_port_id="onnx::Conv_278" />
93
+ <IR name="onnx::Conv_278" output_port_id="0" />
94
+ </map>
95
+ <map>
96
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0" />
97
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" output_port_id="2" />
98
+ </map>
99
+ <map>
100
+ <framework name="onnx::Conv_281" output_port_id="onnx::Conv_281" />
101
+ <IR name="onnx::Conv_281" output_port_id="0" />
102
+ </map>
103
+ <map>
104
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0" />
105
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" output_port_id="2" />
106
+ </map>
107
+ <map>
108
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0" />
109
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" output_port_id="2" />
110
+ </map>
111
+ <map>
112
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0" />
113
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" output_port_id="1" />
114
+ </map>
115
+ <map>
116
+ <framework name="onnx::Conv_284" output_port_id="onnx::Conv_284" />
117
+ <IR name="onnx::Conv_284" output_port_id="0" />
118
+ </map>
119
+ <map>
120
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0" />
121
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" output_port_id="2" />
122
+ </map>
123
+ <map>
124
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0" />
125
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" output_port_id="1" />
126
+ </map>
127
+ <map>
128
+ <framework name="onnx::Conv_287" output_port_id="onnx::Conv_287" />
129
+ <IR name="onnx::Conv_287" output_port_id="0" />
130
+ </map>
131
+ <map>
132
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0" />
133
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" output_port_id="2" />
134
+ </map>
135
+ <map>
136
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0" />
137
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" output_port_id="2" />
138
+ </map>
139
+ <map>
140
+ <framework name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0" />
141
+ <IR name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" output_port_id="1" />
142
+ </map>
143
+ <map>
144
+ <framework name="/Shape" output_port_id="/Shape_output_0" />
145
+ <IR name="/Shape" output_port_id="1" />
146
+ </map>
147
+ <map>
148
+ <framework name="/Constant_1" output_port_id="/Constant_1_output_0" />
149
+ <IR name="/Constant_1" output_port_id="0" />
150
+ </map>
151
+ <map>
152
+ <framework name="/Constant_2" output_port_id="/Constant_2_output_0" />
153
+ <IR name="/Constant_2" output_port_id="0" />
154
+ </map>
155
+ <map>
156
+ <framework name="/Slice" output_port_id="/Slice_output_0" />
157
+ <IR name="/Slice" output_port_id="4" />
158
+ </map>
159
+ <map>
160
+ <framework name="/Constant_3" output_port_id="/Constant_3_output_0" />
161
+ <IR name="/Constant_3" output_port_id="0" />
162
+ </map>
163
+ <map>
164
+ <framework name="/Concat" output_port_id="/Concat_output_0" />
165
+ <IR name="/Concat" output_port_id="2" />
166
+ </map>
167
+ <map>
168
+ <framework name="/Resize" output_port_id="/Resize_output_0" />
169
+ <IR name="/Resize" output_port_id="3" />
170
+ </map>
171
+ <map>
172
+ <framework name="/Concat_1" output_port_id="/Concat_1_output_0" />
173
+ <IR name="/Concat_1" output_port_id="2" />
174
+ </map>
175
+ <map>
176
+ <framework name="onnx::Conv_290" output_port_id="onnx::Conv_290" />
177
+ <IR name="onnx::Conv_290" output_port_id="0" />
178
+ </map>
179
+ <map>
180
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0" />
181
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" output_port_id="2" />
182
+ </map>
183
+ <map>
184
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0" />
185
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" output_port_id="1" />
186
+ </map>
187
+ <map>
188
+ <framework name="onnx::Conv_293" output_port_id="onnx::Conv_293" />
189
+ <IR name="onnx::Conv_293" output_port_id="0" />
190
+ </map>
191
+ <map>
192
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0" />
193
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" output_port_id="2" />
194
+ </map>
195
+ <map>
196
+ <framework name="onnx::Conv_296" output_port_id="onnx::Conv_296" />
197
+ <IR name="onnx::Conv_296" output_port_id="0" />
198
+ </map>
199
+ <map>
200
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0" />
201
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" output_port_id="2" />
202
+ </map>
203
+ <map>
204
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0" />
205
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" output_port_id="2" />
206
+ </map>
207
+ <map>
208
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0" />
209
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" output_port_id="1" />
210
+ </map>
211
+ <map>
212
+ <framework name="onnx::Conv_299" output_port_id="onnx::Conv_299" />
213
+ <IR name="onnx::Conv_299" output_port_id="0" />
214
+ </map>
215
+ <map>
216
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0" />
217
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" output_port_id="2" />
218
+ </map>
219
+ <map>
220
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0" />
221
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" output_port_id="1" />
222
+ </map>
223
+ <map>
224
+ <framework name="onnx::Conv_302" output_port_id="onnx::Conv_302" />
225
+ <IR name="onnx::Conv_302" output_port_id="0" />
226
+ </map>
227
+ <map>
228
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0" />
229
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" output_port_id="2" />
230
+ </map>
231
+ <map>
232
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0" />
233
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" output_port_id="2" />
234
+ </map>
235
+ <map>
236
+ <framework name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0" />
237
+ <IR name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" output_port_id="1" />
238
+ </map>
239
+ <map>
240
+ <framework name="/Shape_3" output_port_id="/Shape_3_output_0" />
241
+ <IR name="/Shape_3" output_port_id="1" />
242
+ </map>
243
+ <map>
244
+ <framework name="/Constant_9" output_port_id="/Constant_9_output_0" />
245
+ <IR name="/Constant_9" output_port_id="0" />
246
+ </map>
247
+ <map>
248
+ <framework name="/Constant_10" output_port_id="/Constant_10_output_0" />
249
+ <IR name="/Constant_10" output_port_id="0" />
250
+ </map>
251
+ <map>
252
+ <framework name="/Slice_1" output_port_id="/Slice_1_output_0" />
253
+ <IR name="/Slice_1" output_port_id="4" />
254
+ </map>
255
+ <map>
256
+ <framework name="/Shape_1" output_port_id="/Shape_1_output_0" />
257
+ <IR name="/Shape_1" output_port_id="1" />
258
+ </map>
259
+ <map>
260
+ <framework name="/Shape_1" output_port_id="/Shape_2_output_0" />
261
+ <IR name="/Shape_1" output_port_id="1" />
262
+ </map>
263
+ <map>
264
+ <framework name="Gather_6092" output_port_id="/Concat_2_output_0" />
265
+ <IR name="Gather_6092" output_port_id="3" />
266
+ </map>
267
+ <map>
268
+ <framework name="Gather_6092" output_port_id="/Cast_output_0" />
269
+ <IR name="Gather_6092" output_port_id="3" />
270
+ </map>
271
+ <map>
272
+ <framework name="/Concat_3" output_port_id="/Concat_3_output_0" />
273
+ <IR name="/Concat_3" output_port_id="2" />
274
+ </map>
275
+ <map>
276
+ <framework name="/Resize_1" output_port_id="/Resize_1_output_0" />
277
+ <IR name="/Resize_1" output_port_id="3" />
278
+ </map>
279
+ <map>
280
+ <framework name="/Concat_4" output_port_id="/Concat_4_output_0" />
281
+ <IR name="/Concat_4" output_port_id="2" />
282
+ </map>
283
+ <map>
284
+ <framework name="onnx::Gather_308" output_port_id="onnx::Gather_308" />
285
+ <IR name="onnx::Gather_308" output_port_id="0" />
286
+ </map>
287
+ <map>
288
+ <framework name="/Gather_2" output_port_id="/Gather_2_output_0" />
289
+ <IR name="/Gather_2" output_port_id="3" />
290
+ </map>
291
+ <map>
292
+ <framework name="/anomaly_map_generator/Constant" output_port_id="/anomaly_map_generator/Constant_output_0" />
293
+ <IR name="/anomaly_map_generator/Constant" output_port_id="0" />
294
+ </map>
295
+ <map>
296
+ <framework name="/anomaly_map_generator/Reshape" output_port_id="/anomaly_map_generator/Reshape_output_0" />
297
+ <IR name="/anomaly_map_generator/Reshape" output_port_id="2" />
298
+ </map>
299
+ <map>
300
+ <framework name="/anomaly_map_generator/Sub" output_port_id="/anomaly_map_generator/Sub_output_0" />
301
+ <IR name="/anomaly_map_generator/Sub" output_port_id="2" />
302
+ </map>
303
+ <map>
304
+ <framework name="/anomaly_map_generator/Transpose" output_port_id="/anomaly_map_generator/Transpose_output_0" />
305
+ <IR name="/anomaly_map_generator/Transpose" output_port_id="2" />
306
+ </map>
307
+ <map>
308
+ <framework name="onnx::MatMul_310" output_port_id="onnx::MatMul_310" />
309
+ <IR name="onnx::MatMul_310" output_port_id="0" />
310
+ </map>
311
+ <map>
312
+ <framework name="/anomaly_map_generator/MatMul" output_port_id="/anomaly_map_generator/MatMul_output_0" />
313
+ <IR name="/anomaly_map_generator/MatMul" output_port_id="2" />
314
+ </map>
315
+ <map>
316
+ <framework name="/anomaly_map_generator/Mul" output_port_id="/anomaly_map_generator/Mul_output_0" />
317
+ <IR name="/anomaly_map_generator/Mul" output_port_id="2" />
318
+ </map>
319
+ <map>
320
+ <framework name="/anomaly_map_generator/ReduceSum" output_port_id="/anomaly_map_generator/ReduceSum_output_0" />
321
+ <IR name="/anomaly_map_generator/ReduceSum" output_port_id="2" />
322
+ </map>
323
+ <map>
324
+ <framework name="/anomaly_map_generator/Constant_1" output_port_id="/anomaly_map_generator/Constant_1_output_0" />
325
+ <IR name="/anomaly_map_generator/Constant_1" output_port_id="0" />
326
+ </map>
327
+ <map>
328
+ <framework name="/anomaly_map_generator/Reshape_1" output_port_id="/anomaly_map_generator/Reshape_1_output_0" />
329
+ <IR name="/anomaly_map_generator/Reshape_1" output_port_id="2" />
330
+ </map>
331
+ <map>
332
+ <framework name="/anomaly_map_generator/Clip" output_port_id="/anomaly_map_generator/Clip_output_0" />
333
+ <IR name="/anomaly_map_generator/Clip" output_port_id="1" />
334
+ </map>
335
+ <map>
336
+ <framework name="/anomaly_map_generator/Sqrt" output_port_id="/anomaly_map_generator/Sqrt_output_0" />
337
+ <IR name="/anomaly_map_generator/Sqrt" output_port_id="1" />
338
+ </map>
339
+ <map>
340
+ <framework name="/anomaly_map_generator/Shape" output_port_id="/anomaly_map_generator/Shape_output_0" />
341
+ <IR name="/anomaly_map_generator/Shape" output_port_id="1" />
342
+ </map>
343
+ <map>
344
+ <framework name="/anomaly_map_generator/Constant_4" output_port_id="/anomaly_map_generator/Constant_4_output_0" />
345
+ <IR name="/anomaly_map_generator/Constant_4" output_port_id="0" />
346
+ </map>
347
+ <map>
348
+ <framework name="/anomaly_map_generator/Constant_5" output_port_id="/anomaly_map_generator/Constant_5_output_0" />
349
+ <IR name="/anomaly_map_generator/Constant_5" output_port_id="0" />
350
+ </map>
351
+ <map>
352
+ <framework name="/anomaly_map_generator/Slice" output_port_id="/anomaly_map_generator/Slice_output_0" />
353
+ <IR name="/anomaly_map_generator/Slice" output_port_id="4" />
354
+ </map>
355
+ <map>
356
+ <framework name="/anomaly_map_generator/Constant_6" output_port_id="/anomaly_map_generator/Constant_6_output_0" />
357
+ <IR name="/anomaly_map_generator/Constant_6" output_port_id="0" />
358
+ </map>
359
+ <map>
360
+ <framework name="/anomaly_map_generator/Concat" output_port_id="/anomaly_map_generator/Concat_output_0" />
361
+ <IR name="/anomaly_map_generator/Concat" output_port_id="2" />
362
+ </map>
363
+ <map>
364
+ <framework name="/anomaly_map_generator/Resize" output_port_id="/anomaly_map_generator/Resize_output_0" />
365
+ <IR name="/anomaly_map_generator/Resize" output_port_id="3" />
366
+ </map>
367
+ <map>
368
+ <framework name="/anomaly_map_generator/blur/Pad" output_port_id="/anomaly_map_generator/blur/Pad_output_0" />
369
+ <IR name="/anomaly_map_generator/blur/Pad" output_port_id="4" />
370
+ </map>
371
+ <map>
372
+ <framework name="anomaly_map_generator.blur.kernel" output_port_id="anomaly_map_generator.blur.kernel" />
373
+ <IR name="anomaly_map_generator.blur.kernel" output_port_id="0" />
374
+ </map>
375
+ <map>
376
+ <framework name="/anomaly_map_generator/blur/Conv" output_port_id="/anomaly_map_generator/blur/Conv_output_0" />
377
+ <IR name="/anomaly_map_generator/blur/Conv" output_port_id="2" />
378
+ </map>
379
+ <map>
380
+ <framework name="/anomaly_map_generator/blur/Constant_8" output_port_id="/anomaly_map_generator/blur/Constant_8_output_0" />
381
+ <IR name="/anomaly_map_generator/blur/Constant_8" output_port_id="0" />
382
+ </map>
383
+ <map>
384
+ <framework name="output" output_port_id="output" />
385
+ <IR name="output" output_port_id="2" />
386
+ </map>
387
+ </mapping>
pill/run/weights/openvino/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1d3587f01391ac27e8289838cbb9e396eb6c383bc059c5ac24745d5d2c3e10b
3
+ size 176626689
pill/run/weights/openvino/model.xml ADDED
@@ -0,0 +1,2703 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="torch_jit" version="11">
3
+ <layers>
4
+ <layer id="0" name="input" type="Parameter" version="opset1">
5
+ <data shape="1,3,256,256" element_type="f32" />
6
+ <output>
7
+ <port id="0" precision="FP32" names="input">
8
+ <dim>1</dim>
9
+ <dim>3</dim>
10
+ <dim>256</dim>
11
+ <dim>256</dim>
12
+ </port>
13
+ </output>
14
+ </layer>
15
+ <layer id="1" name="onnx::Conv_260" type="Const" version="opset1">
16
+ <data element_type="f32" shape="64, 3, 7, 7" offset="0" size="37632" />
17
+ <output>
18
+ <port id="0" precision="FP32" names="onnx::Conv_260">
19
+ <dim>64</dim>
20
+ <dim>3</dim>
21
+ <dim>7</dim>
22
+ <dim>7</dim>
23
+ </port>
24
+ </output>
25
+ </layer>
26
+ <layer id="2" name="/feature_extractor/feature_extractor/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
27
+ <data strides="2, 2" dilations="1, 1" pads_begin="3, 3" pads_end="3, 3" auto_pad="explicit" />
28
+ <input>
29
+ <port id="0" precision="FP32">
30
+ <dim>1</dim>
31
+ <dim>3</dim>
32
+ <dim>256</dim>
33
+ <dim>256</dim>
34
+ </port>
35
+ <port id="1" precision="FP32">
36
+ <dim>64</dim>
37
+ <dim>3</dim>
38
+ <dim>7</dim>
39
+ <dim>7</dim>
40
+ </port>
41
+ </input>
42
+ <output>
43
+ <port id="2" precision="FP32">
44
+ <dim>1</dim>
45
+ <dim>64</dim>
46
+ <dim>128</dim>
47
+ <dim>128</dim>
48
+ </port>
49
+ </output>
50
+ </layer>
51
+ <layer id="3" name="Reshape_55" type="Const" version="opset1">
52
+ <data element_type="f32" shape="1, 64, 1, 1" offset="37632" size="256" />
53
+ <output>
54
+ <port id="0" precision="FP32">
55
+ <dim>1</dim>
56
+ <dim>64</dim>
57
+ <dim>1</dim>
58
+ <dim>1</dim>
59
+ </port>
60
+ </output>
61
+ </layer>
62
+ <layer id="4" name="/feature_extractor/feature_extractor/conv1/Conv" type="Add" version="opset1">
63
+ <data auto_broadcast="numpy" />
64
+ <input>
65
+ <port id="0" precision="FP32">
66
+ <dim>1</dim>
67
+ <dim>64</dim>
68
+ <dim>128</dim>
69
+ <dim>128</dim>
70
+ </port>
71
+ <port id="1" precision="FP32">
72
+ <dim>1</dim>
73
+ <dim>64</dim>
74
+ <dim>1</dim>
75
+ <dim>1</dim>
76
+ </port>
77
+ </input>
78
+ <output>
79
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/conv1/Conv_output_0">
80
+ <dim>1</dim>
81
+ <dim>64</dim>
82
+ <dim>128</dim>
83
+ <dim>128</dim>
84
+ </port>
85
+ </output>
86
+ </layer>
87
+ <layer id="5" name="/feature_extractor/feature_extractor/act1/Relu" type="ReLU" version="opset1">
88
+ <input>
89
+ <port id="0" precision="FP32">
90
+ <dim>1</dim>
91
+ <dim>64</dim>
92
+ <dim>128</dim>
93
+ <dim>128</dim>
94
+ </port>
95
+ </input>
96
+ <output>
97
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/act1/Relu_output_0">
98
+ <dim>1</dim>
99
+ <dim>64</dim>
100
+ <dim>128</dim>
101
+ <dim>128</dim>
102
+ </port>
103
+ </output>
104
+ </layer>
105
+ <layer id="6" name="/feature_extractor/feature_extractor/maxpool/MaxPool" type="MaxPool" version="opset8">
106
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" kernel="3, 3" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
107
+ <input>
108
+ <port id="0" precision="FP32">
109
+ <dim>1</dim>
110
+ <dim>64</dim>
111
+ <dim>128</dim>
112
+ <dim>128</dim>
113
+ </port>
114
+ </input>
115
+ <output>
116
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/maxpool/MaxPool_output_0">
117
+ <dim>1</dim>
118
+ <dim>64</dim>
119
+ <dim>64</dim>
120
+ <dim>64</dim>
121
+ </port>
122
+ <port id="2" precision="I64">
123
+ <dim>1</dim>
124
+ <dim>64</dim>
125
+ <dim>64</dim>
126
+ <dim>64</dim>
127
+ </port>
128
+ </output>
129
+ </layer>
130
+ <layer id="7" name="onnx::Conv_263" type="Const" version="opset1">
131
+ <data element_type="f32" shape="64, 64, 3, 3" offset="37888" size="147456" />
132
+ <output>
133
+ <port id="0" precision="FP32" names="onnx::Conv_263">
134
+ <dim>64</dim>
135
+ <dim>64</dim>
136
+ <dim>3</dim>
137
+ <dim>3</dim>
138
+ </port>
139
+ </output>
140
+ </layer>
141
+ <layer id="8" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
142
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
143
+ <input>
144
+ <port id="0" precision="FP32">
145
+ <dim>1</dim>
146
+ <dim>64</dim>
147
+ <dim>64</dim>
148
+ <dim>64</dim>
149
+ </port>
150
+ <port id="1" precision="FP32">
151
+ <dim>64</dim>
152
+ <dim>64</dim>
153
+ <dim>3</dim>
154
+ <dim>3</dim>
155
+ </port>
156
+ </input>
157
+ <output>
158
+ <port id="2" precision="FP32">
159
+ <dim>1</dim>
160
+ <dim>64</dim>
161
+ <dim>64</dim>
162
+ <dim>64</dim>
163
+ </port>
164
+ </output>
165
+ </layer>
166
+ <layer id="9" name="Reshape_105" type="Const" version="opset1">
167
+ <data element_type="f32" shape="1, 64, 1, 1" offset="185344" size="256" />
168
+ <output>
169
+ <port id="0" precision="FP32">
170
+ <dim>1</dim>
171
+ <dim>64</dim>
172
+ <dim>1</dim>
173
+ <dim>1</dim>
174
+ </port>
175
+ </output>
176
+ </layer>
177
+ <layer id="10" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv" type="Add" version="opset1">
178
+ <data auto_broadcast="numpy" />
179
+ <input>
180
+ <port id="0" precision="FP32">
181
+ <dim>1</dim>
182
+ <dim>64</dim>
183
+ <dim>64</dim>
184
+ <dim>64</dim>
185
+ </port>
186
+ <port id="1" precision="FP32">
187
+ <dim>1</dim>
188
+ <dim>64</dim>
189
+ <dim>1</dim>
190
+ <dim>1</dim>
191
+ </port>
192
+ </input>
193
+ <output>
194
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv1/Conv_output_0">
195
+ <dim>1</dim>
196
+ <dim>64</dim>
197
+ <dim>64</dim>
198
+ <dim>64</dim>
199
+ </port>
200
+ </output>
201
+ </layer>
202
+ <layer id="11" name="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu" type="ReLU" version="opset1">
203
+ <input>
204
+ <port id="0" precision="FP32">
205
+ <dim>1</dim>
206
+ <dim>64</dim>
207
+ <dim>64</dim>
208
+ <dim>64</dim>
209
+ </port>
210
+ </input>
211
+ <output>
212
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act1/Relu_output_0">
213
+ <dim>1</dim>
214
+ <dim>64</dim>
215
+ <dim>64</dim>
216
+ <dim>64</dim>
217
+ </port>
218
+ </output>
219
+ </layer>
220
+ <layer id="12" name="onnx::Conv_266" type="Const" version="opset1">
221
+ <data element_type="f32" shape="64, 64, 3, 3" offset="185600" size="147456" />
222
+ <output>
223
+ <port id="0" precision="FP32" names="onnx::Conv_266">
224
+ <dim>64</dim>
225
+ <dim>64</dim>
226
+ <dim>3</dim>
227
+ <dim>3</dim>
228
+ </port>
229
+ </output>
230
+ </layer>
231
+ <layer id="13" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
232
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
233
+ <input>
234
+ <port id="0" precision="FP32">
235
+ <dim>1</dim>
236
+ <dim>64</dim>
237
+ <dim>64</dim>
238
+ <dim>64</dim>
239
+ </port>
240
+ <port id="1" precision="FP32">
241
+ <dim>64</dim>
242
+ <dim>64</dim>
243
+ <dim>3</dim>
244
+ <dim>3</dim>
245
+ </port>
246
+ </input>
247
+ <output>
248
+ <port id="2" precision="FP32">
249
+ <dim>1</dim>
250
+ <dim>64</dim>
251
+ <dim>64</dim>
252
+ <dim>64</dim>
253
+ </port>
254
+ </output>
255
+ </layer>
256
+ <layer id="14" name="Reshape_154" type="Const" version="opset1">
257
+ <data element_type="f32" shape="1, 64, 1, 1" offset="333056" size="256" />
258
+ <output>
259
+ <port id="0" precision="FP32">
260
+ <dim>1</dim>
261
+ <dim>64</dim>
262
+ <dim>1</dim>
263
+ <dim>1</dim>
264
+ </port>
265
+ </output>
266
+ </layer>
267
+ <layer id="15" name="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv" type="Add" version="opset1">
268
+ <data auto_broadcast="numpy" />
269
+ <input>
270
+ <port id="0" precision="FP32">
271
+ <dim>1</dim>
272
+ <dim>64</dim>
273
+ <dim>64</dim>
274
+ <dim>64</dim>
275
+ </port>
276
+ <port id="1" precision="FP32">
277
+ <dim>1</dim>
278
+ <dim>64</dim>
279
+ <dim>1</dim>
280
+ <dim>1</dim>
281
+ </port>
282
+ </input>
283
+ <output>
284
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/conv2/Conv_output_0">
285
+ <dim>1</dim>
286
+ <dim>64</dim>
287
+ <dim>64</dim>
288
+ <dim>64</dim>
289
+ </port>
290
+ </output>
291
+ </layer>
292
+ <layer id="16" name="/feature_extractor/feature_extractor/layer1/layer1.0/Add" type="Add" version="opset1">
293
+ <data auto_broadcast="numpy" />
294
+ <input>
295
+ <port id="0" precision="FP32">
296
+ <dim>1</dim>
297
+ <dim>64</dim>
298
+ <dim>64</dim>
299
+ <dim>64</dim>
300
+ </port>
301
+ <port id="1" precision="FP32">
302
+ <dim>1</dim>
303
+ <dim>64</dim>
304
+ <dim>64</dim>
305
+ <dim>64</dim>
306
+ </port>
307
+ </input>
308
+ <output>
309
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/Add_output_0">
310
+ <dim>1</dim>
311
+ <dim>64</dim>
312
+ <dim>64</dim>
313
+ <dim>64</dim>
314
+ </port>
315
+ </output>
316
+ </layer>
317
+ <layer id="17" name="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu" type="ReLU" version="opset1">
318
+ <input>
319
+ <port id="0" precision="FP32">
320
+ <dim>1</dim>
321
+ <dim>64</dim>
322
+ <dim>64</dim>
323
+ <dim>64</dim>
324
+ </port>
325
+ </input>
326
+ <output>
327
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.0/act2/Relu_output_0">
328
+ <dim>1</dim>
329
+ <dim>64</dim>
330
+ <dim>64</dim>
331
+ <dim>64</dim>
332
+ </port>
333
+ </output>
334
+ </layer>
335
+ <layer id="18" name="onnx::Conv_269" type="Const" version="opset1">
336
+ <data element_type="f32" shape="64, 64, 3, 3" offset="333312" size="147456" />
337
+ <output>
338
+ <port id="0" precision="FP32" names="onnx::Conv_269">
339
+ <dim>64</dim>
340
+ <dim>64</dim>
341
+ <dim>3</dim>
342
+ <dim>3</dim>
343
+ </port>
344
+ </output>
345
+ </layer>
346
+ <layer id="19" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
347
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
348
+ <input>
349
+ <port id="0" precision="FP32">
350
+ <dim>1</dim>
351
+ <dim>64</dim>
352
+ <dim>64</dim>
353
+ <dim>64</dim>
354
+ </port>
355
+ <port id="1" precision="FP32">
356
+ <dim>64</dim>
357
+ <dim>64</dim>
358
+ <dim>3</dim>
359
+ <dim>3</dim>
360
+ </port>
361
+ </input>
362
+ <output>
363
+ <port id="2" precision="FP32">
364
+ <dim>1</dim>
365
+ <dim>64</dim>
366
+ <dim>64</dim>
367
+ <dim>64</dim>
368
+ </port>
369
+ </output>
370
+ </layer>
371
+ <layer id="20" name="Reshape_204" type="Const" version="opset1">
372
+ <data element_type="f32" shape="1, 64, 1, 1" offset="480768" size="256" />
373
+ <output>
374
+ <port id="0" precision="FP32">
375
+ <dim>1</dim>
376
+ <dim>64</dim>
377
+ <dim>1</dim>
378
+ <dim>1</dim>
379
+ </port>
380
+ </output>
381
+ </layer>
382
+ <layer id="21" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv" type="Add" version="opset1">
383
+ <data auto_broadcast="numpy" />
384
+ <input>
385
+ <port id="0" precision="FP32">
386
+ <dim>1</dim>
387
+ <dim>64</dim>
388
+ <dim>64</dim>
389
+ <dim>64</dim>
390
+ </port>
391
+ <port id="1" precision="FP32">
392
+ <dim>1</dim>
393
+ <dim>64</dim>
394
+ <dim>1</dim>
395
+ <dim>1</dim>
396
+ </port>
397
+ </input>
398
+ <output>
399
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv1/Conv_output_0">
400
+ <dim>1</dim>
401
+ <dim>64</dim>
402
+ <dim>64</dim>
403
+ <dim>64</dim>
404
+ </port>
405
+ </output>
406
+ </layer>
407
+ <layer id="22" name="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu" type="ReLU" version="opset1">
408
+ <input>
409
+ <port id="0" precision="FP32">
410
+ <dim>1</dim>
411
+ <dim>64</dim>
412
+ <dim>64</dim>
413
+ <dim>64</dim>
414
+ </port>
415
+ </input>
416
+ <output>
417
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act1/Relu_output_0">
418
+ <dim>1</dim>
419
+ <dim>64</dim>
420
+ <dim>64</dim>
421
+ <dim>64</dim>
422
+ </port>
423
+ </output>
424
+ </layer>
425
+ <layer id="23" name="onnx::Conv_272" type="Const" version="opset1">
426
+ <data element_type="f32" shape="64, 64, 3, 3" offset="481024" size="147456" />
427
+ <output>
428
+ <port id="0" precision="FP32" names="onnx::Conv_272">
429
+ <dim>64</dim>
430
+ <dim>64</dim>
431
+ <dim>3</dim>
432
+ <dim>3</dim>
433
+ </port>
434
+ </output>
435
+ </layer>
436
+ <layer id="24" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
437
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
438
+ <input>
439
+ <port id="0" precision="FP32">
440
+ <dim>1</dim>
441
+ <dim>64</dim>
442
+ <dim>64</dim>
443
+ <dim>64</dim>
444
+ </port>
445
+ <port id="1" precision="FP32">
446
+ <dim>64</dim>
447
+ <dim>64</dim>
448
+ <dim>3</dim>
449
+ <dim>3</dim>
450
+ </port>
451
+ </input>
452
+ <output>
453
+ <port id="2" precision="FP32">
454
+ <dim>1</dim>
455
+ <dim>64</dim>
456
+ <dim>64</dim>
457
+ <dim>64</dim>
458
+ </port>
459
+ </output>
460
+ </layer>
461
+ <layer id="25" name="Reshape_253" type="Const" version="opset1">
462
+ <data element_type="f32" shape="1, 64, 1, 1" offset="628480" size="256" />
463
+ <output>
464
+ <port id="0" precision="FP32">
465
+ <dim>1</dim>
466
+ <dim>64</dim>
467
+ <dim>1</dim>
468
+ <dim>1</dim>
469
+ </port>
470
+ </output>
471
+ </layer>
472
+ <layer id="26" name="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv" type="Add" version="opset1">
473
+ <data auto_broadcast="numpy" />
474
+ <input>
475
+ <port id="0" precision="FP32">
476
+ <dim>1</dim>
477
+ <dim>64</dim>
478
+ <dim>64</dim>
479
+ <dim>64</dim>
480
+ </port>
481
+ <port id="1" precision="FP32">
482
+ <dim>1</dim>
483
+ <dim>64</dim>
484
+ <dim>1</dim>
485
+ <dim>1</dim>
486
+ </port>
487
+ </input>
488
+ <output>
489
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/conv2/Conv_output_0">
490
+ <dim>1</dim>
491
+ <dim>64</dim>
492
+ <dim>64</dim>
493
+ <dim>64</dim>
494
+ </port>
495
+ </output>
496
+ </layer>
497
+ <layer id="27" name="/feature_extractor/feature_extractor/layer1/layer1.1/Add" type="Add" version="opset1">
498
+ <data auto_broadcast="numpy" />
499
+ <input>
500
+ <port id="0" precision="FP32">
501
+ <dim>1</dim>
502
+ <dim>64</dim>
503
+ <dim>64</dim>
504
+ <dim>64</dim>
505
+ </port>
506
+ <port id="1" precision="FP32">
507
+ <dim>1</dim>
508
+ <dim>64</dim>
509
+ <dim>64</dim>
510
+ <dim>64</dim>
511
+ </port>
512
+ </input>
513
+ <output>
514
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/Add_output_0">
515
+ <dim>1</dim>
516
+ <dim>64</dim>
517
+ <dim>64</dim>
518
+ <dim>64</dim>
519
+ </port>
520
+ </output>
521
+ </layer>
522
+ <layer id="28" name="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu" type="ReLU" version="opset1">
523
+ <input>
524
+ <port id="0" precision="FP32">
525
+ <dim>1</dim>
526
+ <dim>64</dim>
527
+ <dim>64</dim>
528
+ <dim>64</dim>
529
+ </port>
530
+ </input>
531
+ <output>
532
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer1/layer1.1/act2/Relu_output_0">
533
+ <dim>1</dim>
534
+ <dim>64</dim>
535
+ <dim>64</dim>
536
+ <dim>64</dim>
537
+ </port>
538
+ </output>
539
+ </layer>
540
+ <layer id="29" name="onnx::Conv_275" type="Const" version="opset1">
541
+ <data element_type="f32" shape="128, 64, 3, 3" offset="628736" size="294912" />
542
+ <output>
543
+ <port id="0" precision="FP32" names="onnx::Conv_275">
544
+ <dim>128</dim>
545
+ <dim>64</dim>
546
+ <dim>3</dim>
547
+ <dim>3</dim>
548
+ </port>
549
+ </output>
550
+ </layer>
551
+ <layer id="30" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
552
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
553
+ <input>
554
+ <port id="0" precision="FP32">
555
+ <dim>1</dim>
556
+ <dim>64</dim>
557
+ <dim>64</dim>
558
+ <dim>64</dim>
559
+ </port>
560
+ <port id="1" precision="FP32">
561
+ <dim>128</dim>
562
+ <dim>64</dim>
563
+ <dim>3</dim>
564
+ <dim>3</dim>
565
+ </port>
566
+ </input>
567
+ <output>
568
+ <port id="2" precision="FP32">
569
+ <dim>1</dim>
570
+ <dim>128</dim>
571
+ <dim>32</dim>
572
+ <dim>32</dim>
573
+ </port>
574
+ </output>
575
+ </layer>
576
+ <layer id="31" name="Reshape_303" type="Const" version="opset1">
577
+ <data element_type="f32" shape="1, 128, 1, 1" offset="923648" size="512" />
578
+ <output>
579
+ <port id="0" precision="FP32">
580
+ <dim>1</dim>
581
+ <dim>128</dim>
582
+ <dim>1</dim>
583
+ <dim>1</dim>
584
+ </port>
585
+ </output>
586
+ </layer>
587
+ <layer id="32" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv" type="Add" version="opset1">
588
+ <data auto_broadcast="numpy" />
589
+ <input>
590
+ <port id="0" precision="FP32">
591
+ <dim>1</dim>
592
+ <dim>128</dim>
593
+ <dim>32</dim>
594
+ <dim>32</dim>
595
+ </port>
596
+ <port id="1" precision="FP32">
597
+ <dim>1</dim>
598
+ <dim>128</dim>
599
+ <dim>1</dim>
600
+ <dim>1</dim>
601
+ </port>
602
+ </input>
603
+ <output>
604
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv1/Conv_output_0">
605
+ <dim>1</dim>
606
+ <dim>128</dim>
607
+ <dim>32</dim>
608
+ <dim>32</dim>
609
+ </port>
610
+ </output>
611
+ </layer>
612
+ <layer id="33" name="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu" type="ReLU" version="opset1">
613
+ <input>
614
+ <port id="0" precision="FP32">
615
+ <dim>1</dim>
616
+ <dim>128</dim>
617
+ <dim>32</dim>
618
+ <dim>32</dim>
619
+ </port>
620
+ </input>
621
+ <output>
622
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act1/Relu_output_0">
623
+ <dim>1</dim>
624
+ <dim>128</dim>
625
+ <dim>32</dim>
626
+ <dim>32</dim>
627
+ </port>
628
+ </output>
629
+ </layer>
630
+ <layer id="34" name="onnx::Conv_278" type="Const" version="opset1">
631
+ <data element_type="f32" shape="128, 128, 3, 3" offset="924160" size="589824" />
632
+ <output>
633
+ <port id="0" precision="FP32" names="onnx::Conv_278">
634
+ <dim>128</dim>
635
+ <dim>128</dim>
636
+ <dim>3</dim>
637
+ <dim>3</dim>
638
+ </port>
639
+ </output>
640
+ </layer>
641
+ <layer id="35" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
642
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
643
+ <input>
644
+ <port id="0" precision="FP32">
645
+ <dim>1</dim>
646
+ <dim>128</dim>
647
+ <dim>32</dim>
648
+ <dim>32</dim>
649
+ </port>
650
+ <port id="1" precision="FP32">
651
+ <dim>128</dim>
652
+ <dim>128</dim>
653
+ <dim>3</dim>
654
+ <dim>3</dim>
655
+ </port>
656
+ </input>
657
+ <output>
658
+ <port id="2" precision="FP32">
659
+ <dim>1</dim>
660
+ <dim>128</dim>
661
+ <dim>32</dim>
662
+ <dim>32</dim>
663
+ </port>
664
+ </output>
665
+ </layer>
666
+ <layer id="36" name="Reshape_352" type="Const" version="opset1">
667
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1513984" size="512" />
668
+ <output>
669
+ <port id="0" precision="FP32">
670
+ <dim>1</dim>
671
+ <dim>128</dim>
672
+ <dim>1</dim>
673
+ <dim>1</dim>
674
+ </port>
675
+ </output>
676
+ </layer>
677
+ <layer id="37" name="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv" type="Add" version="opset1">
678
+ <data auto_broadcast="numpy" />
679
+ <input>
680
+ <port id="0" precision="FP32">
681
+ <dim>1</dim>
682
+ <dim>128</dim>
683
+ <dim>32</dim>
684
+ <dim>32</dim>
685
+ </port>
686
+ <port id="1" precision="FP32">
687
+ <dim>1</dim>
688
+ <dim>128</dim>
689
+ <dim>1</dim>
690
+ <dim>1</dim>
691
+ </port>
692
+ </input>
693
+ <output>
694
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/conv2/Conv_output_0">
695
+ <dim>1</dim>
696
+ <dim>128</dim>
697
+ <dim>32</dim>
698
+ <dim>32</dim>
699
+ </port>
700
+ </output>
701
+ </layer>
702
+ <layer id="38" name="onnx::Conv_281" type="Const" version="opset1">
703
+ <data element_type="f32" shape="128, 64, 1, 1" offset="1514496" size="32768" />
704
+ <output>
705
+ <port id="0" precision="FP32" names="onnx::Conv_281">
706
+ <dim>128</dim>
707
+ <dim>64</dim>
708
+ <dim>1</dim>
709
+ <dim>1</dim>
710
+ </port>
711
+ </output>
712
+ </layer>
713
+ <layer id="39" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
714
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
715
+ <input>
716
+ <port id="0" precision="FP32">
717
+ <dim>1</dim>
718
+ <dim>64</dim>
719
+ <dim>64</dim>
720
+ <dim>64</dim>
721
+ </port>
722
+ <port id="1" precision="FP32">
723
+ <dim>128</dim>
724
+ <dim>64</dim>
725
+ <dim>1</dim>
726
+ <dim>1</dim>
727
+ </port>
728
+ </input>
729
+ <output>
730
+ <port id="2" precision="FP32">
731
+ <dim>1</dim>
732
+ <dim>128</dim>
733
+ <dim>32</dim>
734
+ <dim>32</dim>
735
+ </port>
736
+ </output>
737
+ </layer>
738
+ <layer id="40" name="Reshape_400" type="Const" version="opset1">
739
+ <data element_type="f32" shape="1, 128, 1, 1" offset="1547264" size="512" />
740
+ <output>
741
+ <port id="0" precision="FP32">
742
+ <dim>1</dim>
743
+ <dim>128</dim>
744
+ <dim>1</dim>
745
+ <dim>1</dim>
746
+ </port>
747
+ </output>
748
+ </layer>
749
+ <layer id="41" name="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv" type="Add" version="opset1">
750
+ <data auto_broadcast="numpy" />
751
+ <input>
752
+ <port id="0" precision="FP32">
753
+ <dim>1</dim>
754
+ <dim>128</dim>
755
+ <dim>32</dim>
756
+ <dim>32</dim>
757
+ </port>
758
+ <port id="1" precision="FP32">
759
+ <dim>1</dim>
760
+ <dim>128</dim>
761
+ <dim>1</dim>
762
+ <dim>1</dim>
763
+ </port>
764
+ </input>
765
+ <output>
766
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/downsample/downsample.0/Conv_output_0">
767
+ <dim>1</dim>
768
+ <dim>128</dim>
769
+ <dim>32</dim>
770
+ <dim>32</dim>
771
+ </port>
772
+ </output>
773
+ </layer>
774
+ <layer id="42" name="/feature_extractor/feature_extractor/layer2/layer2.0/Add" type="Add" version="opset1">
775
+ <data auto_broadcast="numpy" />
776
+ <input>
777
+ <port id="0" precision="FP32">
778
+ <dim>1</dim>
779
+ <dim>128</dim>
780
+ <dim>32</dim>
781
+ <dim>32</dim>
782
+ </port>
783
+ <port id="1" precision="FP32">
784
+ <dim>1</dim>
785
+ <dim>128</dim>
786
+ <dim>32</dim>
787
+ <dim>32</dim>
788
+ </port>
789
+ </input>
790
+ <output>
791
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/Add_output_0">
792
+ <dim>1</dim>
793
+ <dim>128</dim>
794
+ <dim>32</dim>
795
+ <dim>32</dim>
796
+ </port>
797
+ </output>
798
+ </layer>
799
+ <layer id="43" name="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu" type="ReLU" version="opset1">
800
+ <input>
801
+ <port id="0" precision="FP32">
802
+ <dim>1</dim>
803
+ <dim>128</dim>
804
+ <dim>32</dim>
805
+ <dim>32</dim>
806
+ </port>
807
+ </input>
808
+ <output>
809
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.0/act2/Relu_output_0">
810
+ <dim>1</dim>
811
+ <dim>128</dim>
812
+ <dim>32</dim>
813
+ <dim>32</dim>
814
+ </port>
815
+ </output>
816
+ </layer>
817
+ <layer id="44" name="onnx::Conv_284" type="Const" version="opset1">
818
+ <data element_type="f32" shape="128, 128, 3, 3" offset="1547776" size="589824" />
819
+ <output>
820
+ <port id="0" precision="FP32" names="onnx::Conv_284">
821
+ <dim>128</dim>
822
+ <dim>128</dim>
823
+ <dim>3</dim>
824
+ <dim>3</dim>
825
+ </port>
826
+ </output>
827
+ </layer>
828
+ <layer id="45" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
829
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
830
+ <input>
831
+ <port id="0" precision="FP32">
832
+ <dim>1</dim>
833
+ <dim>128</dim>
834
+ <dim>32</dim>
835
+ <dim>32</dim>
836
+ </port>
837
+ <port id="1" precision="FP32">
838
+ <dim>128</dim>
839
+ <dim>128</dim>
840
+ <dim>3</dim>
841
+ <dim>3</dim>
842
+ </port>
843
+ </input>
844
+ <output>
845
+ <port id="2" precision="FP32">
846
+ <dim>1</dim>
847
+ <dim>128</dim>
848
+ <dim>32</dim>
849
+ <dim>32</dim>
850
+ </port>
851
+ </output>
852
+ </layer>
853
+ <layer id="46" name="Reshape_450" type="Const" version="opset1">
854
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2137600" size="512" />
855
+ <output>
856
+ <port id="0" precision="FP32">
857
+ <dim>1</dim>
858
+ <dim>128</dim>
859
+ <dim>1</dim>
860
+ <dim>1</dim>
861
+ </port>
862
+ </output>
863
+ </layer>
864
+ <layer id="47" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv" type="Add" version="opset1">
865
+ <data auto_broadcast="numpy" />
866
+ <input>
867
+ <port id="0" precision="FP32">
868
+ <dim>1</dim>
869
+ <dim>128</dim>
870
+ <dim>32</dim>
871
+ <dim>32</dim>
872
+ </port>
873
+ <port id="1" precision="FP32">
874
+ <dim>1</dim>
875
+ <dim>128</dim>
876
+ <dim>1</dim>
877
+ <dim>1</dim>
878
+ </port>
879
+ </input>
880
+ <output>
881
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv1/Conv_output_0">
882
+ <dim>1</dim>
883
+ <dim>128</dim>
884
+ <dim>32</dim>
885
+ <dim>32</dim>
886
+ </port>
887
+ </output>
888
+ </layer>
889
+ <layer id="48" name="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu" type="ReLU" version="opset1">
890
+ <input>
891
+ <port id="0" precision="FP32">
892
+ <dim>1</dim>
893
+ <dim>128</dim>
894
+ <dim>32</dim>
895
+ <dim>32</dim>
896
+ </port>
897
+ </input>
898
+ <output>
899
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act1/Relu_output_0">
900
+ <dim>1</dim>
901
+ <dim>128</dim>
902
+ <dim>32</dim>
903
+ <dim>32</dim>
904
+ </port>
905
+ </output>
906
+ </layer>
907
+ <layer id="49" name="onnx::Conv_287" type="Const" version="opset1">
908
+ <data element_type="f32" shape="128, 128, 3, 3" offset="2138112" size="589824" />
909
+ <output>
910
+ <port id="0" precision="FP32" names="onnx::Conv_287">
911
+ <dim>128</dim>
912
+ <dim>128</dim>
913
+ <dim>3</dim>
914
+ <dim>3</dim>
915
+ </port>
916
+ </output>
917
+ </layer>
918
+ <layer id="50" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
919
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
920
+ <input>
921
+ <port id="0" precision="FP32">
922
+ <dim>1</dim>
923
+ <dim>128</dim>
924
+ <dim>32</dim>
925
+ <dim>32</dim>
926
+ </port>
927
+ <port id="1" precision="FP32">
928
+ <dim>128</dim>
929
+ <dim>128</dim>
930
+ <dim>3</dim>
931
+ <dim>3</dim>
932
+ </port>
933
+ </input>
934
+ <output>
935
+ <port id="2" precision="FP32">
936
+ <dim>1</dim>
937
+ <dim>128</dim>
938
+ <dim>32</dim>
939
+ <dim>32</dim>
940
+ </port>
941
+ </output>
942
+ </layer>
943
+ <layer id="51" name="Reshape_499" type="Const" version="opset1">
944
+ <data element_type="f32" shape="1, 128, 1, 1" offset="2727936" size="512" />
945
+ <output>
946
+ <port id="0" precision="FP32">
947
+ <dim>1</dim>
948
+ <dim>128</dim>
949
+ <dim>1</dim>
950
+ <dim>1</dim>
951
+ </port>
952
+ </output>
953
+ </layer>
954
+ <layer id="52" name="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv" type="Add" version="opset1">
955
+ <data auto_broadcast="numpy" />
956
+ <input>
957
+ <port id="0" precision="FP32">
958
+ <dim>1</dim>
959
+ <dim>128</dim>
960
+ <dim>32</dim>
961
+ <dim>32</dim>
962
+ </port>
963
+ <port id="1" precision="FP32">
964
+ <dim>1</dim>
965
+ <dim>128</dim>
966
+ <dim>1</dim>
967
+ <dim>1</dim>
968
+ </port>
969
+ </input>
970
+ <output>
971
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/conv2/Conv_output_0">
972
+ <dim>1</dim>
973
+ <dim>128</dim>
974
+ <dim>32</dim>
975
+ <dim>32</dim>
976
+ </port>
977
+ </output>
978
+ </layer>
979
+ <layer id="53" name="/feature_extractor/feature_extractor/layer2/layer2.1/Add" type="Add" version="opset1">
980
+ <data auto_broadcast="numpy" />
981
+ <input>
982
+ <port id="0" precision="FP32">
983
+ <dim>1</dim>
984
+ <dim>128</dim>
985
+ <dim>32</dim>
986
+ <dim>32</dim>
987
+ </port>
988
+ <port id="1" precision="FP32">
989
+ <dim>1</dim>
990
+ <dim>128</dim>
991
+ <dim>32</dim>
992
+ <dim>32</dim>
993
+ </port>
994
+ </input>
995
+ <output>
996
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/Add_output_0">
997
+ <dim>1</dim>
998
+ <dim>128</dim>
999
+ <dim>32</dim>
1000
+ <dim>32</dim>
1001
+ </port>
1002
+ </output>
1003
+ </layer>
1004
+ <layer id="54" name="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu" type="ReLU" version="opset1">
1005
+ <input>
1006
+ <port id="0" precision="FP32">
1007
+ <dim>1</dim>
1008
+ <dim>128</dim>
1009
+ <dim>32</dim>
1010
+ <dim>32</dim>
1011
+ </port>
1012
+ </input>
1013
+ <output>
1014
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer2/layer2.1/act2/Relu_output_0">
1015
+ <dim>1</dim>
1016
+ <dim>128</dim>
1017
+ <dim>32</dim>
1018
+ <dim>32</dim>
1019
+ </port>
1020
+ </output>
1021
+ </layer>
1022
+ <layer id="55" name="/Shape" type="ShapeOf" version="opset3">
1023
+ <data output_type="i64" />
1024
+ <input>
1025
+ <port id="0" precision="FP32">
1026
+ <dim>1</dim>
1027
+ <dim>128</dim>
1028
+ <dim>32</dim>
1029
+ <dim>32</dim>
1030
+ </port>
1031
+ </input>
1032
+ <output>
1033
+ <port id="1" precision="I64" names="/Shape_output_0">
1034
+ <dim>4</dim>
1035
+ </port>
1036
+ </output>
1037
+ </layer>
1038
+ <layer id="56" name="/Constant_1" type="Const" version="opset1">
1039
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1040
+ <output>
1041
+ <port id="0" precision="I64" names="/Constant_1_output_0">
1042
+ <dim>1</dim>
1043
+ </port>
1044
+ </output>
1045
+ </layer>
1046
+ <layer id="57" name="/Constant_2" type="Const" version="opset1">
1047
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1048
+ <output>
1049
+ <port id="0" precision="I64" names="/Constant_2_output_0">
1050
+ <dim>1</dim>
1051
+ </port>
1052
+ </output>
1053
+ </layer>
1054
+ <layer id="58" name="Broadcast_781" type="Const" version="opset1">
1055
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1056
+ <output>
1057
+ <port id="0" precision="I64">
1058
+ <dim>1</dim>
1059
+ </port>
1060
+ </output>
1061
+ </layer>
1062
+ <layer id="59" name="/Slice" type="StridedSlice" version="opset1">
1063
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1064
+ <input>
1065
+ <port id="0" precision="I64">
1066
+ <dim>4</dim>
1067
+ </port>
1068
+ <port id="1" precision="I64">
1069
+ <dim>1</dim>
1070
+ </port>
1071
+ <port id="2" precision="I64">
1072
+ <dim>1</dim>
1073
+ </port>
1074
+ <port id="3" precision="I64">
1075
+ <dim>1</dim>
1076
+ </port>
1077
+ </input>
1078
+ <output>
1079
+ <port id="4" precision="I64" names="/Slice_output_0">
1080
+ <dim>2</dim>
1081
+ </port>
1082
+ </output>
1083
+ </layer>
1084
+ <layer id="60" name="/Constant_3" type="Const" version="opset1">
1085
+ <data element_type="i64" shape="2" offset="2728472" size="16" />
1086
+ <output>
1087
+ <port id="0" precision="I64" names="/Constant_3_output_0">
1088
+ <dim>2</dim>
1089
+ </port>
1090
+ </output>
1091
+ </layer>
1092
+ <layer id="61" name="/Concat" type="Concat" version="opset1">
1093
+ <data axis="0" />
1094
+ <input>
1095
+ <port id="0" precision="I64">
1096
+ <dim>2</dim>
1097
+ </port>
1098
+ <port id="1" precision="I64">
1099
+ <dim>2</dim>
1100
+ </port>
1101
+ </input>
1102
+ <output>
1103
+ <port id="2" precision="I64" names="/Concat_output_0">
1104
+ <dim>4</dim>
1105
+ </port>
1106
+ </output>
1107
+ </layer>
1108
+ <layer id="62" name="Convert_804" type="Convert" version="opset1">
1109
+ <data destination_type="f32" />
1110
+ <input>
1111
+ <port id="0" precision="I64">
1112
+ <dim>4</dim>
1113
+ </port>
1114
+ </input>
1115
+ <output>
1116
+ <port id="1" precision="FP32">
1117
+ <dim>4</dim>
1118
+ </port>
1119
+ </output>
1120
+ </layer>
1121
+ <layer id="63" name="Convert_803" type="Convert" version="opset1">
1122
+ <data destination_type="f32" />
1123
+ <input>
1124
+ <port id="0" precision="I64">
1125
+ <dim>4</dim>
1126
+ </port>
1127
+ </input>
1128
+ <output>
1129
+ <port id="1" precision="FP32">
1130
+ <dim>4</dim>
1131
+ </port>
1132
+ </output>
1133
+ </layer>
1134
+ <layer id="64" name="Divide_805" type="Divide" version="opset1">
1135
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1136
+ <input>
1137
+ <port id="0" precision="FP32">
1138
+ <dim>4</dim>
1139
+ </port>
1140
+ <port id="1" precision="FP32">
1141
+ <dim>4</dim>
1142
+ </port>
1143
+ </input>
1144
+ <output>
1145
+ <port id="2" precision="FP32">
1146
+ <dim>4</dim>
1147
+ </port>
1148
+ </output>
1149
+ </layer>
1150
+ <layer id="65" name="Constant_6582" type="Const" version="opset1">
1151
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1152
+ <output>
1153
+ <port id="0" precision="FP32">
1154
+ <dim>1</dim>
1155
+ </port>
1156
+ </output>
1157
+ </layer>
1158
+ <layer id="66" name="Add_807" type="Add" version="opset1">
1159
+ <data auto_broadcast="numpy" />
1160
+ <input>
1161
+ <port id="0" precision="FP32">
1162
+ <dim>4</dim>
1163
+ </port>
1164
+ <port id="1" precision="FP32">
1165
+ <dim>1</dim>
1166
+ </port>
1167
+ </input>
1168
+ <output>
1169
+ <port id="2" precision="FP32">
1170
+ <dim>4</dim>
1171
+ </port>
1172
+ </output>
1173
+ </layer>
1174
+ <layer id="67" name="/Resize" type="Interpolate" version="opset4">
1175
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1176
+ <input>
1177
+ <port id="0" precision="FP32">
1178
+ <dim>1</dim>
1179
+ <dim>128</dim>
1180
+ <dim>32</dim>
1181
+ <dim>32</dim>
1182
+ </port>
1183
+ <port id="1" precision="I64">
1184
+ <dim>4</dim>
1185
+ </port>
1186
+ <port id="2" precision="FP32">
1187
+ <dim>4</dim>
1188
+ </port>
1189
+ </input>
1190
+ <output>
1191
+ <port id="3" precision="FP32" names="/Resize_output_0">
1192
+ <dim>1</dim>
1193
+ <dim>128</dim>
1194
+ <dim>64</dim>
1195
+ <dim>64</dim>
1196
+ </port>
1197
+ </output>
1198
+ </layer>
1199
+ <layer id="68" name="/Concat_1" type="Concat" version="opset1">
1200
+ <data axis="1" />
1201
+ <input>
1202
+ <port id="0" precision="FP32">
1203
+ <dim>1</dim>
1204
+ <dim>64</dim>
1205
+ <dim>64</dim>
1206
+ <dim>64</dim>
1207
+ </port>
1208
+ <port id="1" precision="FP32">
1209
+ <dim>1</dim>
1210
+ <dim>128</dim>
1211
+ <dim>64</dim>
1212
+ <dim>64</dim>
1213
+ </port>
1214
+ </input>
1215
+ <output>
1216
+ <port id="2" precision="FP32" names="/Concat_1_output_0">
1217
+ <dim>1</dim>
1218
+ <dim>192</dim>
1219
+ <dim>64</dim>
1220
+ <dim>64</dim>
1221
+ </port>
1222
+ </output>
1223
+ </layer>
1224
+ <layer id="69" name="onnx::Conv_290" type="Const" version="opset1">
1225
+ <data element_type="f32" shape="256, 128, 3, 3" offset="2728492" size="1179648" />
1226
+ <output>
1227
+ <port id="0" precision="FP32" names="onnx::Conv_290">
1228
+ <dim>256</dim>
1229
+ <dim>128</dim>
1230
+ <dim>3</dim>
1231
+ <dim>3</dim>
1232
+ </port>
1233
+ </output>
1234
+ </layer>
1235
+ <layer id="70" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1236
+ <data strides="2, 2" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1237
+ <input>
1238
+ <port id="0" precision="FP32">
1239
+ <dim>1</dim>
1240
+ <dim>128</dim>
1241
+ <dim>32</dim>
1242
+ <dim>32</dim>
1243
+ </port>
1244
+ <port id="1" precision="FP32">
1245
+ <dim>256</dim>
1246
+ <dim>128</dim>
1247
+ <dim>3</dim>
1248
+ <dim>3</dim>
1249
+ </port>
1250
+ </input>
1251
+ <output>
1252
+ <port id="2" precision="FP32">
1253
+ <dim>1</dim>
1254
+ <dim>256</dim>
1255
+ <dim>16</dim>
1256
+ <dim>16</dim>
1257
+ </port>
1258
+ </output>
1259
+ </layer>
1260
+ <layer id="71" name="Reshape_549" type="Const" version="opset1">
1261
+ <data element_type="f32" shape="1, 256, 1, 1" offset="3908140" size="1024" />
1262
+ <output>
1263
+ <port id="0" precision="FP32">
1264
+ <dim>1</dim>
1265
+ <dim>256</dim>
1266
+ <dim>1</dim>
1267
+ <dim>1</dim>
1268
+ </port>
1269
+ </output>
1270
+ </layer>
1271
+ <layer id="72" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv" type="Add" version="opset1">
1272
+ <data auto_broadcast="numpy" />
1273
+ <input>
1274
+ <port id="0" precision="FP32">
1275
+ <dim>1</dim>
1276
+ <dim>256</dim>
1277
+ <dim>16</dim>
1278
+ <dim>16</dim>
1279
+ </port>
1280
+ <port id="1" precision="FP32">
1281
+ <dim>1</dim>
1282
+ <dim>256</dim>
1283
+ <dim>1</dim>
1284
+ <dim>1</dim>
1285
+ </port>
1286
+ </input>
1287
+ <output>
1288
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv1/Conv_output_0">
1289
+ <dim>1</dim>
1290
+ <dim>256</dim>
1291
+ <dim>16</dim>
1292
+ <dim>16</dim>
1293
+ </port>
1294
+ </output>
1295
+ </layer>
1296
+ <layer id="73" name="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu" type="ReLU" version="opset1">
1297
+ <input>
1298
+ <port id="0" precision="FP32">
1299
+ <dim>1</dim>
1300
+ <dim>256</dim>
1301
+ <dim>16</dim>
1302
+ <dim>16</dim>
1303
+ </port>
1304
+ </input>
1305
+ <output>
1306
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act1/Relu_output_0">
1307
+ <dim>1</dim>
1308
+ <dim>256</dim>
1309
+ <dim>16</dim>
1310
+ <dim>16</dim>
1311
+ </port>
1312
+ </output>
1313
+ </layer>
1314
+ <layer id="74" name="onnx::Conv_293" type="Const" version="opset1">
1315
+ <data element_type="f32" shape="256, 256, 3, 3" offset="3909164" size="2359296" />
1316
+ <output>
1317
+ <port id="0" precision="FP32" names="onnx::Conv_293">
1318
+ <dim>256</dim>
1319
+ <dim>256</dim>
1320
+ <dim>3</dim>
1321
+ <dim>3</dim>
1322
+ </port>
1323
+ </output>
1324
+ </layer>
1325
+ <layer id="75" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1326
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1327
+ <input>
1328
+ <port id="0" precision="FP32">
1329
+ <dim>1</dim>
1330
+ <dim>256</dim>
1331
+ <dim>16</dim>
1332
+ <dim>16</dim>
1333
+ </port>
1334
+ <port id="1" precision="FP32">
1335
+ <dim>256</dim>
1336
+ <dim>256</dim>
1337
+ <dim>3</dim>
1338
+ <dim>3</dim>
1339
+ </port>
1340
+ </input>
1341
+ <output>
1342
+ <port id="2" precision="FP32">
1343
+ <dim>1</dim>
1344
+ <dim>256</dim>
1345
+ <dim>16</dim>
1346
+ <dim>16</dim>
1347
+ </port>
1348
+ </output>
1349
+ </layer>
1350
+ <layer id="76" name="Reshape_598" type="Const" version="opset1">
1351
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6268460" size="1024" />
1352
+ <output>
1353
+ <port id="0" precision="FP32">
1354
+ <dim>1</dim>
1355
+ <dim>256</dim>
1356
+ <dim>1</dim>
1357
+ <dim>1</dim>
1358
+ </port>
1359
+ </output>
1360
+ </layer>
1361
+ <layer id="77" name="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv" type="Add" version="opset1">
1362
+ <data auto_broadcast="numpy" />
1363
+ <input>
1364
+ <port id="0" precision="FP32">
1365
+ <dim>1</dim>
1366
+ <dim>256</dim>
1367
+ <dim>16</dim>
1368
+ <dim>16</dim>
1369
+ </port>
1370
+ <port id="1" precision="FP32">
1371
+ <dim>1</dim>
1372
+ <dim>256</dim>
1373
+ <dim>1</dim>
1374
+ <dim>1</dim>
1375
+ </port>
1376
+ </input>
1377
+ <output>
1378
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/conv2/Conv_output_0">
1379
+ <dim>1</dim>
1380
+ <dim>256</dim>
1381
+ <dim>16</dim>
1382
+ <dim>16</dim>
1383
+ </port>
1384
+ </output>
1385
+ </layer>
1386
+ <layer id="78" name="onnx::Conv_296" type="Const" version="opset1">
1387
+ <data element_type="f32" shape="256, 128, 1, 1" offset="6269484" size="131072" />
1388
+ <output>
1389
+ <port id="0" precision="FP32" names="onnx::Conv_296">
1390
+ <dim>256</dim>
1391
+ <dim>128</dim>
1392
+ <dim>1</dim>
1393
+ <dim>1</dim>
1394
+ </port>
1395
+ </output>
1396
+ </layer>
1397
+ <layer id="79" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv/WithoutBiases" type="Convolution" version="opset1">
1398
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
1399
+ <input>
1400
+ <port id="0" precision="FP32">
1401
+ <dim>1</dim>
1402
+ <dim>128</dim>
1403
+ <dim>32</dim>
1404
+ <dim>32</dim>
1405
+ </port>
1406
+ <port id="1" precision="FP32">
1407
+ <dim>256</dim>
1408
+ <dim>128</dim>
1409
+ <dim>1</dim>
1410
+ <dim>1</dim>
1411
+ </port>
1412
+ </input>
1413
+ <output>
1414
+ <port id="2" precision="FP32">
1415
+ <dim>1</dim>
1416
+ <dim>256</dim>
1417
+ <dim>16</dim>
1418
+ <dim>16</dim>
1419
+ </port>
1420
+ </output>
1421
+ </layer>
1422
+ <layer id="80" name="Reshape_646" type="Const" version="opset1">
1423
+ <data element_type="f32" shape="1, 256, 1, 1" offset="6400556" size="1024" />
1424
+ <output>
1425
+ <port id="0" precision="FP32">
1426
+ <dim>1</dim>
1427
+ <dim>256</dim>
1428
+ <dim>1</dim>
1429
+ <dim>1</dim>
1430
+ </port>
1431
+ </output>
1432
+ </layer>
1433
+ <layer id="81" name="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv" type="Add" version="opset1">
1434
+ <data auto_broadcast="numpy" />
1435
+ <input>
1436
+ <port id="0" precision="FP32">
1437
+ <dim>1</dim>
1438
+ <dim>256</dim>
1439
+ <dim>16</dim>
1440
+ <dim>16</dim>
1441
+ </port>
1442
+ <port id="1" precision="FP32">
1443
+ <dim>1</dim>
1444
+ <dim>256</dim>
1445
+ <dim>1</dim>
1446
+ <dim>1</dim>
1447
+ </port>
1448
+ </input>
1449
+ <output>
1450
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/downsample/downsample.0/Conv_output_0">
1451
+ <dim>1</dim>
1452
+ <dim>256</dim>
1453
+ <dim>16</dim>
1454
+ <dim>16</dim>
1455
+ </port>
1456
+ </output>
1457
+ </layer>
1458
+ <layer id="82" name="/feature_extractor/feature_extractor/layer3/layer3.0/Add" type="Add" version="opset1">
1459
+ <data auto_broadcast="numpy" />
1460
+ <input>
1461
+ <port id="0" precision="FP32">
1462
+ <dim>1</dim>
1463
+ <dim>256</dim>
1464
+ <dim>16</dim>
1465
+ <dim>16</dim>
1466
+ </port>
1467
+ <port id="1" precision="FP32">
1468
+ <dim>1</dim>
1469
+ <dim>256</dim>
1470
+ <dim>16</dim>
1471
+ <dim>16</dim>
1472
+ </port>
1473
+ </input>
1474
+ <output>
1475
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/Add_output_0">
1476
+ <dim>1</dim>
1477
+ <dim>256</dim>
1478
+ <dim>16</dim>
1479
+ <dim>16</dim>
1480
+ </port>
1481
+ </output>
1482
+ </layer>
1483
+ <layer id="83" name="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu" type="ReLU" version="opset1">
1484
+ <input>
1485
+ <port id="0" precision="FP32">
1486
+ <dim>1</dim>
1487
+ <dim>256</dim>
1488
+ <dim>16</dim>
1489
+ <dim>16</dim>
1490
+ </port>
1491
+ </input>
1492
+ <output>
1493
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.0/act2/Relu_output_0">
1494
+ <dim>1</dim>
1495
+ <dim>256</dim>
1496
+ <dim>16</dim>
1497
+ <dim>16</dim>
1498
+ </port>
1499
+ </output>
1500
+ </layer>
1501
+ <layer id="84" name="onnx::Conv_299" type="Const" version="opset1">
1502
+ <data element_type="f32" shape="256, 256, 3, 3" offset="6401580" size="2359296" />
1503
+ <output>
1504
+ <port id="0" precision="FP32" names="onnx::Conv_299">
1505
+ <dim>256</dim>
1506
+ <dim>256</dim>
1507
+ <dim>3</dim>
1508
+ <dim>3</dim>
1509
+ </port>
1510
+ </output>
1511
+ </layer>
1512
+ <layer id="85" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv/WithoutBiases" type="Convolution" version="opset1">
1513
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1514
+ <input>
1515
+ <port id="0" precision="FP32">
1516
+ <dim>1</dim>
1517
+ <dim>256</dim>
1518
+ <dim>16</dim>
1519
+ <dim>16</dim>
1520
+ </port>
1521
+ <port id="1" precision="FP32">
1522
+ <dim>256</dim>
1523
+ <dim>256</dim>
1524
+ <dim>3</dim>
1525
+ <dim>3</dim>
1526
+ </port>
1527
+ </input>
1528
+ <output>
1529
+ <port id="2" precision="FP32">
1530
+ <dim>1</dim>
1531
+ <dim>256</dim>
1532
+ <dim>16</dim>
1533
+ <dim>16</dim>
1534
+ </port>
1535
+ </output>
1536
+ </layer>
1537
+ <layer id="86" name="Reshape_696" type="Const" version="opset1">
1538
+ <data element_type="f32" shape="1, 256, 1, 1" offset="8760876" size="1024" />
1539
+ <output>
1540
+ <port id="0" precision="FP32">
1541
+ <dim>1</dim>
1542
+ <dim>256</dim>
1543
+ <dim>1</dim>
1544
+ <dim>1</dim>
1545
+ </port>
1546
+ </output>
1547
+ </layer>
1548
+ <layer id="87" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv" type="Add" version="opset1">
1549
+ <data auto_broadcast="numpy" />
1550
+ <input>
1551
+ <port id="0" precision="FP32">
1552
+ <dim>1</dim>
1553
+ <dim>256</dim>
1554
+ <dim>16</dim>
1555
+ <dim>16</dim>
1556
+ </port>
1557
+ <port id="1" precision="FP32">
1558
+ <dim>1</dim>
1559
+ <dim>256</dim>
1560
+ <dim>1</dim>
1561
+ <dim>1</dim>
1562
+ </port>
1563
+ </input>
1564
+ <output>
1565
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv1/Conv_output_0">
1566
+ <dim>1</dim>
1567
+ <dim>256</dim>
1568
+ <dim>16</dim>
1569
+ <dim>16</dim>
1570
+ </port>
1571
+ </output>
1572
+ </layer>
1573
+ <layer id="88" name="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu" type="ReLU" version="opset1">
1574
+ <input>
1575
+ <port id="0" precision="FP32">
1576
+ <dim>1</dim>
1577
+ <dim>256</dim>
1578
+ <dim>16</dim>
1579
+ <dim>16</dim>
1580
+ </port>
1581
+ </input>
1582
+ <output>
1583
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act1/Relu_output_0">
1584
+ <dim>1</dim>
1585
+ <dim>256</dim>
1586
+ <dim>16</dim>
1587
+ <dim>16</dim>
1588
+ </port>
1589
+ </output>
1590
+ </layer>
1591
+ <layer id="89" name="onnx::Conv_302" type="Const" version="opset1">
1592
+ <data element_type="f32" shape="256, 256, 3, 3" offset="8761900" size="2359296" />
1593
+ <output>
1594
+ <port id="0" precision="FP32" names="onnx::Conv_302">
1595
+ <dim>256</dim>
1596
+ <dim>256</dim>
1597
+ <dim>3</dim>
1598
+ <dim>3</dim>
1599
+ </port>
1600
+ </output>
1601
+ </layer>
1602
+ <layer id="90" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv/WithoutBiases" type="Convolution" version="opset1">
1603
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1604
+ <input>
1605
+ <port id="0" precision="FP32">
1606
+ <dim>1</dim>
1607
+ <dim>256</dim>
1608
+ <dim>16</dim>
1609
+ <dim>16</dim>
1610
+ </port>
1611
+ <port id="1" precision="FP32">
1612
+ <dim>256</dim>
1613
+ <dim>256</dim>
1614
+ <dim>3</dim>
1615
+ <dim>3</dim>
1616
+ </port>
1617
+ </input>
1618
+ <output>
1619
+ <port id="2" precision="FP32">
1620
+ <dim>1</dim>
1621
+ <dim>256</dim>
1622
+ <dim>16</dim>
1623
+ <dim>16</dim>
1624
+ </port>
1625
+ </output>
1626
+ </layer>
1627
+ <layer id="91" name="Reshape_745" type="Const" version="opset1">
1628
+ <data element_type="f32" shape="1, 256, 1, 1" offset="11121196" size="1024" />
1629
+ <output>
1630
+ <port id="0" precision="FP32">
1631
+ <dim>1</dim>
1632
+ <dim>256</dim>
1633
+ <dim>1</dim>
1634
+ <dim>1</dim>
1635
+ </port>
1636
+ </output>
1637
+ </layer>
1638
+ <layer id="92" name="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv" type="Add" version="opset1">
1639
+ <data auto_broadcast="numpy" />
1640
+ <input>
1641
+ <port id="0" precision="FP32">
1642
+ <dim>1</dim>
1643
+ <dim>256</dim>
1644
+ <dim>16</dim>
1645
+ <dim>16</dim>
1646
+ </port>
1647
+ <port id="1" precision="FP32">
1648
+ <dim>1</dim>
1649
+ <dim>256</dim>
1650
+ <dim>1</dim>
1651
+ <dim>1</dim>
1652
+ </port>
1653
+ </input>
1654
+ <output>
1655
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/conv2/Conv_output_0">
1656
+ <dim>1</dim>
1657
+ <dim>256</dim>
1658
+ <dim>16</dim>
1659
+ <dim>16</dim>
1660
+ </port>
1661
+ </output>
1662
+ </layer>
1663
+ <layer id="93" name="/feature_extractor/feature_extractor/layer3/layer3.1/Add" type="Add" version="opset1">
1664
+ <data auto_broadcast="numpy" />
1665
+ <input>
1666
+ <port id="0" precision="FP32">
1667
+ <dim>1</dim>
1668
+ <dim>256</dim>
1669
+ <dim>16</dim>
1670
+ <dim>16</dim>
1671
+ </port>
1672
+ <port id="1" precision="FP32">
1673
+ <dim>1</dim>
1674
+ <dim>256</dim>
1675
+ <dim>16</dim>
1676
+ <dim>16</dim>
1677
+ </port>
1678
+ </input>
1679
+ <output>
1680
+ <port id="2" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/Add_output_0">
1681
+ <dim>1</dim>
1682
+ <dim>256</dim>
1683
+ <dim>16</dim>
1684
+ <dim>16</dim>
1685
+ </port>
1686
+ </output>
1687
+ </layer>
1688
+ <layer id="94" name="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu" type="ReLU" version="opset1">
1689
+ <input>
1690
+ <port id="0" precision="FP32">
1691
+ <dim>1</dim>
1692
+ <dim>256</dim>
1693
+ <dim>16</dim>
1694
+ <dim>16</dim>
1695
+ </port>
1696
+ </input>
1697
+ <output>
1698
+ <port id="1" precision="FP32" names="/feature_extractor/feature_extractor/layer3/layer3.1/act2/Relu_output_0">
1699
+ <dim>1</dim>
1700
+ <dim>256</dim>
1701
+ <dim>16</dim>
1702
+ <dim>16</dim>
1703
+ </port>
1704
+ </output>
1705
+ </layer>
1706
+ <layer id="95" name="/Shape_3" type="ShapeOf" version="opset3">
1707
+ <data output_type="i64" />
1708
+ <input>
1709
+ <port id="0" precision="FP32">
1710
+ <dim>1</dim>
1711
+ <dim>256</dim>
1712
+ <dim>16</dim>
1713
+ <dim>16</dim>
1714
+ </port>
1715
+ </input>
1716
+ <output>
1717
+ <port id="1" precision="I64" names="/Shape_3_output_0">
1718
+ <dim>4</dim>
1719
+ </port>
1720
+ </output>
1721
+ </layer>
1722
+ <layer id="96" name="/Constant_9" type="Const" version="opset1">
1723
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
1724
+ <output>
1725
+ <port id="0" precision="I64" names="/Constant_9_output_0">
1726
+ <dim>1</dim>
1727
+ </port>
1728
+ </output>
1729
+ </layer>
1730
+ <layer id="97" name="/Constant_10" type="Const" version="opset1">
1731
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
1732
+ <output>
1733
+ <port id="0" precision="I64" names="/Constant_10_output_0">
1734
+ <dim>1</dim>
1735
+ </port>
1736
+ </output>
1737
+ </layer>
1738
+ <layer id="98" name="Broadcast_861" type="Const" version="opset1">
1739
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
1740
+ <output>
1741
+ <port id="0" precision="I64">
1742
+ <dim>1</dim>
1743
+ </port>
1744
+ </output>
1745
+ </layer>
1746
+ <layer id="99" name="/Slice_1" type="StridedSlice" version="opset1">
1747
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
1748
+ <input>
1749
+ <port id="0" precision="I64">
1750
+ <dim>4</dim>
1751
+ </port>
1752
+ <port id="1" precision="I64">
1753
+ <dim>1</dim>
1754
+ </port>
1755
+ <port id="2" precision="I64">
1756
+ <dim>1</dim>
1757
+ </port>
1758
+ <port id="3" precision="I64">
1759
+ <dim>1</dim>
1760
+ </port>
1761
+ </input>
1762
+ <output>
1763
+ <port id="4" precision="I64" names="/Slice_1_output_0">
1764
+ <dim>2</dim>
1765
+ </port>
1766
+ </output>
1767
+ </layer>
1768
+ <layer id="100" name="/Shape_1" type="ShapeOf" version="opset3">
1769
+ <data output_type="i64" />
1770
+ <input>
1771
+ <port id="0" precision="FP32">
1772
+ <dim>1</dim>
1773
+ <dim>192</dim>
1774
+ <dim>64</dim>
1775
+ <dim>64</dim>
1776
+ </port>
1777
+ </input>
1778
+ <output>
1779
+ <port id="1" precision="I64" names="/Shape_1_output_0,/Shape_2_output_0">
1780
+ <dim>4</dim>
1781
+ </port>
1782
+ </output>
1783
+ </layer>
1784
+ <layer id="101" name="Constant_6090" type="Const" version="opset1">
1785
+ <data element_type="i64" shape="2" offset="11122220" size="16" />
1786
+ <output>
1787
+ <port id="0" precision="I64">
1788
+ <dim>2</dim>
1789
+ </port>
1790
+ </output>
1791
+ </layer>
1792
+ <layer id="102" name="Constant_6091" type="Const" version="opset1">
1793
+ <data element_type="i64" shape="" offset="2728448" size="8" />
1794
+ <output>
1795
+ <port id="0" precision="I64" />
1796
+ </output>
1797
+ </layer>
1798
+ <layer id="103" name="Gather_6092" type="Gather" version="opset8">
1799
+ <data batch_dims="0" />
1800
+ <input>
1801
+ <port id="0" precision="I64">
1802
+ <dim>4</dim>
1803
+ </port>
1804
+ <port id="1" precision="I64">
1805
+ <dim>2</dim>
1806
+ </port>
1807
+ <port id="2" precision="I64" />
1808
+ </input>
1809
+ <output>
1810
+ <port id="3" precision="I64" names="/Cast_output_0,/Concat_2_output_0">
1811
+ <dim>2</dim>
1812
+ </port>
1813
+ </output>
1814
+ </layer>
1815
+ <layer id="104" name="/Concat_3" type="Concat" version="opset1">
1816
+ <data axis="0" />
1817
+ <input>
1818
+ <port id="0" precision="I64">
1819
+ <dim>2</dim>
1820
+ </port>
1821
+ <port id="1" precision="I64">
1822
+ <dim>2</dim>
1823
+ </port>
1824
+ </input>
1825
+ <output>
1826
+ <port id="2" precision="I64" names="/Concat_3_output_0">
1827
+ <dim>4</dim>
1828
+ </port>
1829
+ </output>
1830
+ </layer>
1831
+ <layer id="105" name="Convert_884" type="Convert" version="opset1">
1832
+ <data destination_type="f32" />
1833
+ <input>
1834
+ <port id="0" precision="I64">
1835
+ <dim>4</dim>
1836
+ </port>
1837
+ </input>
1838
+ <output>
1839
+ <port id="1" precision="FP32">
1840
+ <dim>4</dim>
1841
+ </port>
1842
+ </output>
1843
+ </layer>
1844
+ <layer id="106" name="Convert_883" type="Convert" version="opset1">
1845
+ <data destination_type="f32" />
1846
+ <input>
1847
+ <port id="0" precision="I64">
1848
+ <dim>4</dim>
1849
+ </port>
1850
+ </input>
1851
+ <output>
1852
+ <port id="1" precision="FP32">
1853
+ <dim>4</dim>
1854
+ </port>
1855
+ </output>
1856
+ </layer>
1857
+ <layer id="107" name="Divide_885" type="Divide" version="opset1">
1858
+ <data auto_broadcast="numpy" m_pythondiv="true" />
1859
+ <input>
1860
+ <port id="0" precision="FP32">
1861
+ <dim>4</dim>
1862
+ </port>
1863
+ <port id="1" precision="FP32">
1864
+ <dim>4</dim>
1865
+ </port>
1866
+ </input>
1867
+ <output>
1868
+ <port id="2" precision="FP32">
1869
+ <dim>4</dim>
1870
+ </port>
1871
+ </output>
1872
+ </layer>
1873
+ <layer id="108" name="Constant_6583" type="Const" version="opset1">
1874
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
1875
+ <output>
1876
+ <port id="0" precision="FP32">
1877
+ <dim>1</dim>
1878
+ </port>
1879
+ </output>
1880
+ </layer>
1881
+ <layer id="109" name="Add_887" type="Add" version="opset1">
1882
+ <data auto_broadcast="numpy" />
1883
+ <input>
1884
+ <port id="0" precision="FP32">
1885
+ <dim>4</dim>
1886
+ </port>
1887
+ <port id="1" precision="FP32">
1888
+ <dim>1</dim>
1889
+ </port>
1890
+ </input>
1891
+ <output>
1892
+ <port id="2" precision="FP32">
1893
+ <dim>4</dim>
1894
+ </port>
1895
+ </output>
1896
+ </layer>
1897
+ <layer id="110" name="/Resize_1" type="Interpolate" version="opset4">
1898
+ <data mode="nearest" shape_calculation_mode="sizes" coordinate_transformation_mode="asymmetric" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
1899
+ <input>
1900
+ <port id="0" precision="FP32">
1901
+ <dim>1</dim>
1902
+ <dim>256</dim>
1903
+ <dim>16</dim>
1904
+ <dim>16</dim>
1905
+ </port>
1906
+ <port id="1" precision="I64">
1907
+ <dim>4</dim>
1908
+ </port>
1909
+ <port id="2" precision="FP32">
1910
+ <dim>4</dim>
1911
+ </port>
1912
+ </input>
1913
+ <output>
1914
+ <port id="3" precision="FP32" names="/Resize_1_output_0">
1915
+ <dim>1</dim>
1916
+ <dim>256</dim>
1917
+ <dim>64</dim>
1918
+ <dim>64</dim>
1919
+ </port>
1920
+ </output>
1921
+ </layer>
1922
+ <layer id="111" name="/Concat_4" type="Concat" version="opset1">
1923
+ <data axis="1" />
1924
+ <input>
1925
+ <port id="0" precision="FP32">
1926
+ <dim>1</dim>
1927
+ <dim>192</dim>
1928
+ <dim>64</dim>
1929
+ <dim>64</dim>
1930
+ </port>
1931
+ <port id="1" precision="FP32">
1932
+ <dim>1</dim>
1933
+ <dim>256</dim>
1934
+ <dim>64</dim>
1935
+ <dim>64</dim>
1936
+ </port>
1937
+ </input>
1938
+ <output>
1939
+ <port id="2" precision="FP32" names="/Concat_4_output_0">
1940
+ <dim>1</dim>
1941
+ <dim>448</dim>
1942
+ <dim>64</dim>
1943
+ <dim>64</dim>
1944
+ </port>
1945
+ </output>
1946
+ </layer>
1947
+ <layer id="112" name="onnx::Gather_308" type="Const" version="opset1">
1948
+ <data element_type="i64" shape="100" offset="11122236" size="800" />
1949
+ <output>
1950
+ <port id="0" precision="I64" names="onnx::Gather_308">
1951
+ <dim>100</dim>
1952
+ </port>
1953
+ </output>
1954
+ </layer>
1955
+ <layer id="113" name="Constant_965" type="Const" version="opset1">
1956
+ <data element_type="i64" shape="" offset="2728464" size="8" />
1957
+ <output>
1958
+ <port id="0" precision="I64" />
1959
+ </output>
1960
+ </layer>
1961
+ <layer id="114" name="/Gather_2" type="Gather" version="opset8">
1962
+ <data batch_dims="0" />
1963
+ <input>
1964
+ <port id="0" precision="FP32">
1965
+ <dim>1</dim>
1966
+ <dim>448</dim>
1967
+ <dim>64</dim>
1968
+ <dim>64</dim>
1969
+ </port>
1970
+ <port id="1" precision="I64">
1971
+ <dim>100</dim>
1972
+ </port>
1973
+ <port id="2" precision="I64" />
1974
+ </input>
1975
+ <output>
1976
+ <port id="3" precision="FP32" names="/Gather_2_output_0">
1977
+ <dim>1</dim>
1978
+ <dim>100</dim>
1979
+ <dim>64</dim>
1980
+ <dim>64</dim>
1981
+ </port>
1982
+ </output>
1983
+ </layer>
1984
+ <layer id="115" name="/anomaly_map_generator/Constant" type="Const" version="opset1">
1985
+ <data element_type="i64" shape="3" offset="11123036" size="24" />
1986
+ <output>
1987
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_output_0">
1988
+ <dim>3</dim>
1989
+ </port>
1990
+ </output>
1991
+ </layer>
1992
+ <layer id="116" name="/anomaly_map_generator/Reshape" type="Reshape" version="opset1">
1993
+ <data special_zero="true" />
1994
+ <input>
1995
+ <port id="0" precision="FP32">
1996
+ <dim>1</dim>
1997
+ <dim>100</dim>
1998
+ <dim>64</dim>
1999
+ <dim>64</dim>
2000
+ </port>
2001
+ <port id="1" precision="I64">
2002
+ <dim>3</dim>
2003
+ </port>
2004
+ </input>
2005
+ <output>
2006
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_output_0">
2007
+ <dim>1</dim>
2008
+ <dim>100</dim>
2009
+ <dim>4096</dim>
2010
+ </port>
2011
+ </output>
2012
+ </layer>
2013
+ <layer id="117" name="Constant_6584" type="Const" version="opset1">
2014
+ <data element_type="f32" shape="1, 100, 4096" offset="11123060" size="1638400" />
2015
+ <output>
2016
+ <port id="0" precision="FP32">
2017
+ <dim>1</dim>
2018
+ <dim>100</dim>
2019
+ <dim>4096</dim>
2020
+ </port>
2021
+ </output>
2022
+ </layer>
2023
+ <layer id="118" name="/anomaly_map_generator/Sub" type="Subtract" version="opset1">
2024
+ <data auto_broadcast="numpy" />
2025
+ <input>
2026
+ <port id="0" precision="FP32">
2027
+ <dim>1</dim>
2028
+ <dim>100</dim>
2029
+ <dim>4096</dim>
2030
+ </port>
2031
+ <port id="1" precision="FP32">
2032
+ <dim>1</dim>
2033
+ <dim>100</dim>
2034
+ <dim>4096</dim>
2035
+ </port>
2036
+ </input>
2037
+ <output>
2038
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Sub_output_0">
2039
+ <dim>1</dim>
2040
+ <dim>100</dim>
2041
+ <dim>4096</dim>
2042
+ </port>
2043
+ </output>
2044
+ </layer>
2045
+ <layer id="119" name="Constant_975" type="Const" version="opset1">
2046
+ <data element_type="i64" shape="3" offset="12761460" size="24" />
2047
+ <output>
2048
+ <port id="0" precision="I64">
2049
+ <dim>3</dim>
2050
+ </port>
2051
+ </output>
2052
+ </layer>
2053
+ <layer id="120" name="/anomaly_map_generator/Transpose" type="Transpose" version="opset1">
2054
+ <input>
2055
+ <port id="0" precision="FP32">
2056
+ <dim>1</dim>
2057
+ <dim>100</dim>
2058
+ <dim>4096</dim>
2059
+ </port>
2060
+ <port id="1" precision="I64">
2061
+ <dim>3</dim>
2062
+ </port>
2063
+ </input>
2064
+ <output>
2065
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Transpose_output_0">
2066
+ <dim>4096</dim>
2067
+ <dim>1</dim>
2068
+ <dim>100</dim>
2069
+ </port>
2070
+ </output>
2071
+ </layer>
2072
+ <layer id="121" name="onnx::MatMul_310" type="Const" version="opset1">
2073
+ <data element_type="f32" shape="4096, 100, 100" offset="12761484" size="163840000" />
2074
+ <output>
2075
+ <port id="0" precision="FP32" names="onnx::MatMul_310">
2076
+ <dim>4096</dim>
2077
+ <dim>100</dim>
2078
+ <dim>100</dim>
2079
+ </port>
2080
+ </output>
2081
+ </layer>
2082
+ <layer id="122" name="/anomaly_map_generator/MatMul" type="MatMul" version="opset1">
2083
+ <data transpose_a="false" transpose_b="false" />
2084
+ <input>
2085
+ <port id="0" precision="FP32">
2086
+ <dim>4096</dim>
2087
+ <dim>1</dim>
2088
+ <dim>100</dim>
2089
+ </port>
2090
+ <port id="1" precision="FP32">
2091
+ <dim>4096</dim>
2092
+ <dim>100</dim>
2093
+ <dim>100</dim>
2094
+ </port>
2095
+ </input>
2096
+ <output>
2097
+ <port id="2" precision="FP32" names="/anomaly_map_generator/MatMul_output_0">
2098
+ <dim>4096</dim>
2099
+ <dim>1</dim>
2100
+ <dim>100</dim>
2101
+ </port>
2102
+ </output>
2103
+ </layer>
2104
+ <layer id="123" name="/anomaly_map_generator/Mul" type="Multiply" version="opset1">
2105
+ <data auto_broadcast="numpy" />
2106
+ <input>
2107
+ <port id="0" precision="FP32">
2108
+ <dim>4096</dim>
2109
+ <dim>1</dim>
2110
+ <dim>100</dim>
2111
+ </port>
2112
+ <port id="1" precision="FP32">
2113
+ <dim>4096</dim>
2114
+ <dim>1</dim>
2115
+ <dim>100</dim>
2116
+ </port>
2117
+ </input>
2118
+ <output>
2119
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Mul_output_0">
2120
+ <dim>4096</dim>
2121
+ <dim>1</dim>
2122
+ <dim>100</dim>
2123
+ </port>
2124
+ </output>
2125
+ </layer>
2126
+ <layer id="124" name="Constant_979" type="Const" version="opset1">
2127
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2128
+ <output>
2129
+ <port id="0" precision="I64">
2130
+ <dim>1</dim>
2131
+ </port>
2132
+ </output>
2133
+ </layer>
2134
+ <layer id="125" name="/anomaly_map_generator/ReduceSum" type="ReduceSum" version="opset1">
2135
+ <data keep_dims="false" />
2136
+ <input>
2137
+ <port id="0" precision="FP32">
2138
+ <dim>4096</dim>
2139
+ <dim>1</dim>
2140
+ <dim>100</dim>
2141
+ </port>
2142
+ <port id="1" precision="I64">
2143
+ <dim>1</dim>
2144
+ </port>
2145
+ </input>
2146
+ <output>
2147
+ <port id="2" precision="FP32" names="/anomaly_map_generator/ReduceSum_output_0">
2148
+ <dim>4096</dim>
2149
+ <dim>1</dim>
2150
+ </port>
2151
+ </output>
2152
+ </layer>
2153
+ <layer id="126" name="/anomaly_map_generator/Constant_1" type="Const" version="opset1">
2154
+ <data element_type="i64" shape="4" offset="176601484" size="32" />
2155
+ <output>
2156
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_1_output_0">
2157
+ <dim>4</dim>
2158
+ </port>
2159
+ </output>
2160
+ </layer>
2161
+ <layer id="127" name="/anomaly_map_generator/Reshape_1" type="Reshape" version="opset1">
2162
+ <data special_zero="true" />
2163
+ <input>
2164
+ <port id="0" precision="FP32">
2165
+ <dim>4096</dim>
2166
+ <dim>1</dim>
2167
+ </port>
2168
+ <port id="1" precision="I64">
2169
+ <dim>4</dim>
2170
+ </port>
2171
+ </input>
2172
+ <output>
2173
+ <port id="2" precision="FP32" names="/anomaly_map_generator/Reshape_1_output_0">
2174
+ <dim>1</dim>
2175
+ <dim>1</dim>
2176
+ <dim>64</dim>
2177
+ <dim>64</dim>
2178
+ </port>
2179
+ </output>
2180
+ </layer>
2181
+ <layer id="128" name="/anomaly_map_generator/Clip" type="Clamp" version="opset1">
2182
+ <data min="0" max="3.4028234663852886e+38" />
2183
+ <input>
2184
+ <port id="0" precision="FP32">
2185
+ <dim>1</dim>
2186
+ <dim>1</dim>
2187
+ <dim>64</dim>
2188
+ <dim>64</dim>
2189
+ </port>
2190
+ </input>
2191
+ <output>
2192
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Clip_output_0">
2193
+ <dim>1</dim>
2194
+ <dim>1</dim>
2195
+ <dim>64</dim>
2196
+ <dim>64</dim>
2197
+ </port>
2198
+ </output>
2199
+ </layer>
2200
+ <layer id="129" name="/anomaly_map_generator/Sqrt" type="Sqrt" version="opset1">
2201
+ <input>
2202
+ <port id="0" precision="FP32">
2203
+ <dim>1</dim>
2204
+ <dim>1</dim>
2205
+ <dim>64</dim>
2206
+ <dim>64</dim>
2207
+ </port>
2208
+ </input>
2209
+ <output>
2210
+ <port id="1" precision="FP32" names="/anomaly_map_generator/Sqrt_output_0">
2211
+ <dim>1</dim>
2212
+ <dim>1</dim>
2213
+ <dim>64</dim>
2214
+ <dim>64</dim>
2215
+ </port>
2216
+ </output>
2217
+ </layer>
2218
+ <layer id="130" name="/anomaly_map_generator/Shape" type="ShapeOf" version="opset3">
2219
+ <data output_type="i64" />
2220
+ <input>
2221
+ <port id="0" precision="FP32">
2222
+ <dim>1</dim>
2223
+ <dim>1</dim>
2224
+ <dim>64</dim>
2225
+ <dim>64</dim>
2226
+ </port>
2227
+ </input>
2228
+ <output>
2229
+ <port id="1" precision="I64" names="/anomaly_map_generator/Shape_output_0">
2230
+ <dim>4</dim>
2231
+ </port>
2232
+ </output>
2233
+ </layer>
2234
+ <layer id="131" name="/anomaly_map_generator/Constant_4" type="Const" version="opset1">
2235
+ <data element_type="i64" shape="1" offset="2728448" size="8" />
2236
+ <output>
2237
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_4_output_0">
2238
+ <dim>1</dim>
2239
+ </port>
2240
+ </output>
2241
+ </layer>
2242
+ <layer id="132" name="/anomaly_map_generator/Constant_5" type="Const" version="opset1">
2243
+ <data element_type="i64" shape="1" offset="2728456" size="8" />
2244
+ <output>
2245
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_5_output_0">
2246
+ <dim>1</dim>
2247
+ </port>
2248
+ </output>
2249
+ </layer>
2250
+ <layer id="133" name="Broadcast_1002" type="Const" version="opset1">
2251
+ <data element_type="i64" shape="1" offset="2728464" size="8" />
2252
+ <output>
2253
+ <port id="0" precision="I64">
2254
+ <dim>1</dim>
2255
+ </port>
2256
+ </output>
2257
+ </layer>
2258
+ <layer id="134" name="/anomaly_map_generator/Slice" type="StridedSlice" version="opset1">
2259
+ <data begin_mask="0" end_mask="0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask="" />
2260
+ <input>
2261
+ <port id="0" precision="I64">
2262
+ <dim>4</dim>
2263
+ </port>
2264
+ <port id="1" precision="I64">
2265
+ <dim>1</dim>
2266
+ </port>
2267
+ <port id="2" precision="I64">
2268
+ <dim>1</dim>
2269
+ </port>
2270
+ <port id="3" precision="I64">
2271
+ <dim>1</dim>
2272
+ </port>
2273
+ </input>
2274
+ <output>
2275
+ <port id="4" precision="I64" names="/anomaly_map_generator/Slice_output_0">
2276
+ <dim>2</dim>
2277
+ </port>
2278
+ </output>
2279
+ </layer>
2280
+ <layer id="135" name="/anomaly_map_generator/Constant_6" type="Const" version="opset1">
2281
+ <data element_type="i64" shape="2" offset="176601516" size="16" />
2282
+ <output>
2283
+ <port id="0" precision="I64" names="/anomaly_map_generator/Constant_6_output_0">
2284
+ <dim>2</dim>
2285
+ </port>
2286
+ </output>
2287
+ </layer>
2288
+ <layer id="136" name="/anomaly_map_generator/Concat" type="Concat" version="opset1">
2289
+ <data axis="0" />
2290
+ <input>
2291
+ <port id="0" precision="I64">
2292
+ <dim>2</dim>
2293
+ </port>
2294
+ <port id="1" precision="I64">
2295
+ <dim>2</dim>
2296
+ </port>
2297
+ </input>
2298
+ <output>
2299
+ <port id="2" precision="I64" names="/anomaly_map_generator/Concat_output_0">
2300
+ <dim>4</dim>
2301
+ </port>
2302
+ </output>
2303
+ </layer>
2304
+ <layer id="137" name="Convert_1025" type="Convert" version="opset1">
2305
+ <data destination_type="f32" />
2306
+ <input>
2307
+ <port id="0" precision="I64">
2308
+ <dim>4</dim>
2309
+ </port>
2310
+ </input>
2311
+ <output>
2312
+ <port id="1" precision="FP32">
2313
+ <dim>4</dim>
2314
+ </port>
2315
+ </output>
2316
+ </layer>
2317
+ <layer id="138" name="Convert_1024" type="Convert" version="opset1">
2318
+ <data destination_type="f32" />
2319
+ <input>
2320
+ <port id="0" precision="I64">
2321
+ <dim>4</dim>
2322
+ </port>
2323
+ </input>
2324
+ <output>
2325
+ <port id="1" precision="FP32">
2326
+ <dim>4</dim>
2327
+ </port>
2328
+ </output>
2329
+ </layer>
2330
+ <layer id="139" name="Divide_1026" type="Divide" version="opset1">
2331
+ <data auto_broadcast="numpy" m_pythondiv="true" />
2332
+ <input>
2333
+ <port id="0" precision="FP32">
2334
+ <dim>4</dim>
2335
+ </port>
2336
+ <port id="1" precision="FP32">
2337
+ <dim>4</dim>
2338
+ </port>
2339
+ </input>
2340
+ <output>
2341
+ <port id="2" precision="FP32">
2342
+ <dim>4</dim>
2343
+ </port>
2344
+ </output>
2345
+ </layer>
2346
+ <layer id="140" name="Constant_6585" type="Const" version="opset1">
2347
+ <data element_type="f32" shape="1" offset="2728488" size="4" />
2348
+ <output>
2349
+ <port id="0" precision="FP32">
2350
+ <dim>1</dim>
2351
+ </port>
2352
+ </output>
2353
+ </layer>
2354
+ <layer id="141" name="Add_1028" type="Add" version="opset1">
2355
+ <data auto_broadcast="numpy" />
2356
+ <input>
2357
+ <port id="0" precision="FP32">
2358
+ <dim>4</dim>
2359
+ </port>
2360
+ <port id="1" precision="FP32">
2361
+ <dim>1</dim>
2362
+ </port>
2363
+ </input>
2364
+ <output>
2365
+ <port id="2" precision="FP32">
2366
+ <dim>4</dim>
2367
+ </port>
2368
+ </output>
2369
+ </layer>
2370
+ <layer id="142" name="/anomaly_map_generator/Resize" type="Interpolate" version="opset4">
2371
+ <data mode="linear_onnx" shape_calculation_mode="sizes" coordinate_transformation_mode="half_pixel" nearest_mode="floor" antialias="false" pads_begin="0, 0, 0, 0" pads_end="0, 0, 0, 0" cube_coeff="-0.75" />
2372
+ <input>
2373
+ <port id="0" precision="FP32">
2374
+ <dim>1</dim>
2375
+ <dim>1</dim>
2376
+ <dim>64</dim>
2377
+ <dim>64</dim>
2378
+ </port>
2379
+ <port id="1" precision="I64">
2380
+ <dim>4</dim>
2381
+ </port>
2382
+ <port id="2" precision="FP32">
2383
+ <dim>4</dim>
2384
+ </port>
2385
+ </input>
2386
+ <output>
2387
+ <port id="3" precision="FP32" names="/anomaly_map_generator/Resize_output_0">
2388
+ <dim>1</dim>
2389
+ <dim>1</dim>
2390
+ <dim>256</dim>
2391
+ <dim>256</dim>
2392
+ </port>
2393
+ </output>
2394
+ </layer>
2395
+ <layer id="143" name="Split_1097.0" type="Const" version="opset1">
2396
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2397
+ <output>
2398
+ <port id="0" precision="I64">
2399
+ <dim>4</dim>
2400
+ </port>
2401
+ </output>
2402
+ </layer>
2403
+ <layer id="144" name="Split_1097.1" type="Const" version="opset1">
2404
+ <data element_type="i64" shape="4" offset="176601532" size="32" />
2405
+ <output>
2406
+ <port id="0" precision="I64">
2407
+ <dim>4</dim>
2408
+ </port>
2409
+ </output>
2410
+ </layer>
2411
+ <layer id="145" name="Constant_1095" type="Const" version="opset1">
2412
+ <data element_type="f32" shape="" offset="176601564" size="4" />
2413
+ <output>
2414
+ <port id="0" precision="FP32" />
2415
+ </output>
2416
+ </layer>
2417
+ <layer id="146" name="/anomaly_map_generator/blur/Pad" type="Pad" version="opset1">
2418
+ <data pad_mode="reflect" />
2419
+ <input>
2420
+ <port id="0" precision="FP32">
2421
+ <dim>1</dim>
2422
+ <dim>1</dim>
2423
+ <dim>256</dim>
2424
+ <dim>256</dim>
2425
+ </port>
2426
+ <port id="1" precision="I64">
2427
+ <dim>4</dim>
2428
+ </port>
2429
+ <port id="2" precision="I64">
2430
+ <dim>4</dim>
2431
+ </port>
2432
+ <port id="3" precision="FP32" />
2433
+ </input>
2434
+ <output>
2435
+ <port id="4" precision="FP32" names="/anomaly_map_generator/blur/Pad_output_0">
2436
+ <dim>1</dim>
2437
+ <dim>1</dim>
2438
+ <dim>288</dim>
2439
+ <dim>288</dim>
2440
+ </port>
2441
+ </output>
2442
+ </layer>
2443
+ <layer id="147" name="anomaly_map_generator.blur.kernel" type="Const" version="opset1">
2444
+ <data element_type="f32" shape="1, 1, 33, 33" offset="176601568" size="4356" />
2445
+ <output>
2446
+ <port id="0" precision="FP32" names="anomaly_map_generator.blur.kernel">
2447
+ <dim>1</dim>
2448
+ <dim>1</dim>
2449
+ <dim>33</dim>
2450
+ <dim>33</dim>
2451
+ </port>
2452
+ </output>
2453
+ </layer>
2454
+ <layer id="148" name="/anomaly_map_generator/blur/Conv" type="Convolution" version="opset1">
2455
+ <data strides="1, 1" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" auto_pad="explicit" />
2456
+ <input>
2457
+ <port id="0" precision="FP32">
2458
+ <dim>1</dim>
2459
+ <dim>1</dim>
2460
+ <dim>288</dim>
2461
+ <dim>288</dim>
2462
+ </port>
2463
+ <port id="1" precision="FP32">
2464
+ <dim>1</dim>
2465
+ <dim>1</dim>
2466
+ <dim>33</dim>
2467
+ <dim>33</dim>
2468
+ </port>
2469
+ </input>
2470
+ <output>
2471
+ <port id="2" precision="FP32" names="/anomaly_map_generator/blur/Conv_output_0">
2472
+ <dim>1</dim>
2473
+ <dim>1</dim>
2474
+ <dim>256</dim>
2475
+ <dim>256</dim>
2476
+ </port>
2477
+ </output>
2478
+ </layer>
2479
+ <layer id="149" name="/anomaly_map_generator/blur/Constant_8" type="Const" version="opset1">
2480
+ <data element_type="i64" shape="4" offset="176605924" size="32" />
2481
+ <output>
2482
+ <port id="0" precision="I64" names="/anomaly_map_generator/blur/Constant_8_output_0">
2483
+ <dim>4</dim>
2484
+ </port>
2485
+ </output>
2486
+ </layer>
2487
+ <layer id="150" name="output" type="Reshape" version="opset1">
2488
+ <data special_zero="true" />
2489
+ <input>
2490
+ <port id="0" precision="FP32">
2491
+ <dim>1</dim>
2492
+ <dim>1</dim>
2493
+ <dim>256</dim>
2494
+ <dim>256</dim>
2495
+ </port>
2496
+ <port id="1" precision="I64">
2497
+ <dim>4</dim>
2498
+ </port>
2499
+ </input>
2500
+ <output>
2501
+ <port id="2" precision="FP32" names="output">
2502
+ <dim>1</dim>
2503
+ <dim>1</dim>
2504
+ <dim>256</dim>
2505
+ <dim>256</dim>
2506
+ </port>
2507
+ </output>
2508
+ </layer>
2509
+ <layer id="151" name="output/sink_port_0" type="Result" version="opset1">
2510
+ <input>
2511
+ <port id="0" precision="FP32">
2512
+ <dim>1</dim>
2513
+ <dim>1</dim>
2514
+ <dim>256</dim>
2515
+ <dim>256</dim>
2516
+ </port>
2517
+ </input>
2518
+ </layer>
2519
+ </layers>
2520
+ <edges>
2521
+ <edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
2522
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="1" />
2523
+ <edge from-layer="2" from-port="2" to-layer="4" to-port="0" />
2524
+ <edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
2525
+ <edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
2526
+ <edge from-layer="5" from-port="1" to-layer="6" to-port="0" />
2527
+ <edge from-layer="6" from-port="1" to-layer="8" to-port="0" />
2528
+ <edge from-layer="6" from-port="1" to-layer="16" to-port="1" />
2529
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="1" />
2530
+ <edge from-layer="8" from-port="2" to-layer="10" to-port="0" />
2531
+ <edge from-layer="9" from-port="0" to-layer="10" to-port="1" />
2532
+ <edge from-layer="10" from-port="2" to-layer="11" to-port="0" />
2533
+ <edge from-layer="11" from-port="1" to-layer="13" to-port="0" />
2534
+ <edge from-layer="12" from-port="0" to-layer="13" to-port="1" />
2535
+ <edge from-layer="13" from-port="2" to-layer="15" to-port="0" />
2536
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="1" />
2537
+ <edge from-layer="15" from-port="2" to-layer="16" to-port="0" />
2538
+ <edge from-layer="16" from-port="2" to-layer="17" to-port="0" />
2539
+ <edge from-layer="17" from-port="1" to-layer="19" to-port="0" />
2540
+ <edge from-layer="17" from-port="1" to-layer="27" to-port="1" />
2541
+ <edge from-layer="18" from-port="0" to-layer="19" to-port="1" />
2542
+ <edge from-layer="19" from-port="2" to-layer="21" to-port="0" />
2543
+ <edge from-layer="20" from-port="0" to-layer="21" to-port="1" />
2544
+ <edge from-layer="21" from-port="2" to-layer="22" to-port="0" />
2545
+ <edge from-layer="22" from-port="1" to-layer="24" to-port="0" />
2546
+ <edge from-layer="23" from-port="0" to-layer="24" to-port="1" />
2547
+ <edge from-layer="24" from-port="2" to-layer="26" to-port="0" />
2548
+ <edge from-layer="25" from-port="0" to-layer="26" to-port="1" />
2549
+ <edge from-layer="26" from-port="2" to-layer="27" to-port="0" />
2550
+ <edge from-layer="27" from-port="2" to-layer="28" to-port="0" />
2551
+ <edge from-layer="28" from-port="1" to-layer="30" to-port="0" />
2552
+ <edge from-layer="28" from-port="1" to-layer="39" to-port="0" />
2553
+ <edge from-layer="28" from-port="1" to-layer="68" to-port="0" />
2554
+ <edge from-layer="29" from-port="0" to-layer="30" to-port="1" />
2555
+ <edge from-layer="30" from-port="2" to-layer="32" to-port="0" />
2556
+ <edge from-layer="31" from-port="0" to-layer="32" to-port="1" />
2557
+ <edge from-layer="32" from-port="2" to-layer="33" to-port="0" />
2558
+ <edge from-layer="33" from-port="1" to-layer="35" to-port="0" />
2559
+ <edge from-layer="34" from-port="0" to-layer="35" to-port="1" />
2560
+ <edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
2561
+ <edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
2562
+ <edge from-layer="37" from-port="2" to-layer="42" to-port="0" />
2563
+ <edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
2564
+ <edge from-layer="39" from-port="2" to-layer="41" to-port="0" />
2565
+ <edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
2566
+ <edge from-layer="41" from-port="2" to-layer="42" to-port="1" />
2567
+ <edge from-layer="42" from-port="2" to-layer="43" to-port="0" />
2568
+ <edge from-layer="43" from-port="1" to-layer="45" to-port="0" />
2569
+ <edge from-layer="43" from-port="1" to-layer="53" to-port="1" />
2570
+ <edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
2571
+ <edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
2572
+ <edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
2573
+ <edge from-layer="47" from-port="2" to-layer="48" to-port="0" />
2574
+ <edge from-layer="48" from-port="1" to-layer="50" to-port="0" />
2575
+ <edge from-layer="49" from-port="0" to-layer="50" to-port="1" />
2576
+ <edge from-layer="50" from-port="2" to-layer="52" to-port="0" />
2577
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="1" />
2578
+ <edge from-layer="52" from-port="2" to-layer="53" to-port="0" />
2579
+ <edge from-layer="53" from-port="2" to-layer="54" to-port="0" />
2580
+ <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
2581
+ <edge from-layer="54" from-port="1" to-layer="67" to-port="0" />
2582
+ <edge from-layer="54" from-port="1" to-layer="79" to-port="0" />
2583
+ <edge from-layer="54" from-port="1" to-layer="70" to-port="0" />
2584
+ <edge from-layer="55" from-port="1" to-layer="59" to-port="0" />
2585
+ <edge from-layer="55" from-port="1" to-layer="63" to-port="0" />
2586
+ <edge from-layer="56" from-port="0" to-layer="59" to-port="1" />
2587
+ <edge from-layer="57" from-port="0" to-layer="59" to-port="2" />
2588
+ <edge from-layer="58" from-port="0" to-layer="59" to-port="3" />
2589
+ <edge from-layer="59" from-port="4" to-layer="61" to-port="0" />
2590
+ <edge from-layer="60" from-port="0" to-layer="61" to-port="1" />
2591
+ <edge from-layer="61" from-port="2" to-layer="62" to-port="0" />
2592
+ <edge from-layer="61" from-port="2" to-layer="67" to-port="1" />
2593
+ <edge from-layer="62" from-port="1" to-layer="64" to-port="0" />
2594
+ <edge from-layer="63" from-port="1" to-layer="64" to-port="1" />
2595
+ <edge from-layer="64" from-port="2" to-layer="66" to-port="0" />
2596
+ <edge from-layer="65" from-port="0" to-layer="66" to-port="1" />
2597
+ <edge from-layer="66" from-port="2" to-layer="67" to-port="2" />
2598
+ <edge from-layer="67" from-port="3" to-layer="68" to-port="1" />
2599
+ <edge from-layer="68" from-port="2" to-layer="111" to-port="0" />
2600
+ <edge from-layer="68" from-port="2" to-layer="100" to-port="0" />
2601
+ <edge from-layer="69" from-port="0" to-layer="70" to-port="1" />
2602
+ <edge from-layer="70" from-port="2" to-layer="72" to-port="0" />
2603
+ <edge from-layer="71" from-port="0" to-layer="72" to-port="1" />
2604
+ <edge from-layer="72" from-port="2" to-layer="73" to-port="0" />
2605
+ <edge from-layer="73" from-port="1" to-layer="75" to-port="0" />
2606
+ <edge from-layer="74" from-port="0" to-layer="75" to-port="1" />
2607
+ <edge from-layer="75" from-port="2" to-layer="77" to-port="0" />
2608
+ <edge from-layer="76" from-port="0" to-layer="77" to-port="1" />
2609
+ <edge from-layer="77" from-port="2" to-layer="82" to-port="0" />
2610
+ <edge from-layer="78" from-port="0" to-layer="79" to-port="1" />
2611
+ <edge from-layer="79" from-port="2" to-layer="81" to-port="0" />
2612
+ <edge from-layer="80" from-port="0" to-layer="81" to-port="1" />
2613
+ <edge from-layer="81" from-port="2" to-layer="82" to-port="1" />
2614
+ <edge from-layer="82" from-port="2" to-layer="83" to-port="0" />
2615
+ <edge from-layer="83" from-port="1" to-layer="85" to-port="0" />
2616
+ <edge from-layer="83" from-port="1" to-layer="93" to-port="1" />
2617
+ <edge from-layer="84" from-port="0" to-layer="85" to-port="1" />
2618
+ <edge from-layer="85" from-port="2" to-layer="87" to-port="0" />
2619
+ <edge from-layer="86" from-port="0" to-layer="87" to-port="1" />
2620
+ <edge from-layer="87" from-port="2" to-layer="88" to-port="0" />
2621
+ <edge from-layer="88" from-port="1" to-layer="90" to-port="0" />
2622
+ <edge from-layer="89" from-port="0" to-layer="90" to-port="1" />
2623
+ <edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
2624
+ <edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
2625
+ <edge from-layer="92" from-port="2" to-layer="93" to-port="0" />
2626
+ <edge from-layer="93" from-port="2" to-layer="94" to-port="0" />
2627
+ <edge from-layer="94" from-port="1" to-layer="95" to-port="0" />
2628
+ <edge from-layer="94" from-port="1" to-layer="110" to-port="0" />
2629
+ <edge from-layer="95" from-port="1" to-layer="99" to-port="0" />
2630
+ <edge from-layer="95" from-port="1" to-layer="106" to-port="0" />
2631
+ <edge from-layer="96" from-port="0" to-layer="99" to-port="1" />
2632
+ <edge from-layer="97" from-port="0" to-layer="99" to-port="2" />
2633
+ <edge from-layer="98" from-port="0" to-layer="99" to-port="3" />
2634
+ <edge from-layer="99" from-port="4" to-layer="104" to-port="0" />
2635
+ <edge from-layer="100" from-port="1" to-layer="103" to-port="0" />
2636
+ <edge from-layer="101" from-port="0" to-layer="103" to-port="1" />
2637
+ <edge from-layer="102" from-port="0" to-layer="103" to-port="2" />
2638
+ <edge from-layer="103" from-port="3" to-layer="104" to-port="1" />
2639
+ <edge from-layer="104" from-port="2" to-layer="105" to-port="0" />
2640
+ <edge from-layer="104" from-port="2" to-layer="110" to-port="1" />
2641
+ <edge from-layer="105" from-port="1" to-layer="107" to-port="0" />
2642
+ <edge from-layer="106" from-port="1" to-layer="107" to-port="1" />
2643
+ <edge from-layer="107" from-port="2" to-layer="109" to-port="0" />
2644
+ <edge from-layer="108" from-port="0" to-layer="109" to-port="1" />
2645
+ <edge from-layer="109" from-port="2" to-layer="110" to-port="2" />
2646
+ <edge from-layer="110" from-port="3" to-layer="111" to-port="1" />
2647
+ <edge from-layer="111" from-port="2" to-layer="114" to-port="0" />
2648
+ <edge from-layer="112" from-port="0" to-layer="114" to-port="1" />
2649
+ <edge from-layer="113" from-port="0" to-layer="114" to-port="2" />
2650
+ <edge from-layer="114" from-port="3" to-layer="116" to-port="0" />
2651
+ <edge from-layer="115" from-port="0" to-layer="116" to-port="1" />
2652
+ <edge from-layer="116" from-port="2" to-layer="118" to-port="0" />
2653
+ <edge from-layer="117" from-port="0" to-layer="118" to-port="1" />
2654
+ <edge from-layer="118" from-port="2" to-layer="120" to-port="0" />
2655
+ <edge from-layer="119" from-port="0" to-layer="120" to-port="1" />
2656
+ <edge from-layer="120" from-port="2" to-layer="122" to-port="0" />
2657
+ <edge from-layer="120" from-port="2" to-layer="123" to-port="1" />
2658
+ <edge from-layer="121" from-port="0" to-layer="122" to-port="1" />
2659
+ <edge from-layer="122" from-port="2" to-layer="123" to-port="0" />
2660
+ <edge from-layer="123" from-port="2" to-layer="125" to-port="0" />
2661
+ <edge from-layer="124" from-port="0" to-layer="125" to-port="1" />
2662
+ <edge from-layer="125" from-port="2" to-layer="127" to-port="0" />
2663
+ <edge from-layer="126" from-port="0" to-layer="127" to-port="1" />
2664
+ <edge from-layer="127" from-port="2" to-layer="128" to-port="0" />
2665
+ <edge from-layer="128" from-port="1" to-layer="129" to-port="0" />
2666
+ <edge from-layer="129" from-port="1" to-layer="130" to-port="0" />
2667
+ <edge from-layer="129" from-port="1" to-layer="142" to-port="0" />
2668
+ <edge from-layer="130" from-port="1" to-layer="134" to-port="0" />
2669
+ <edge from-layer="130" from-port="1" to-layer="138" to-port="0" />
2670
+ <edge from-layer="131" from-port="0" to-layer="134" to-port="1" />
2671
+ <edge from-layer="132" from-port="0" to-layer="134" to-port="2" />
2672
+ <edge from-layer="133" from-port="0" to-layer="134" to-port="3" />
2673
+ <edge from-layer="134" from-port="4" to-layer="136" to-port="0" />
2674
+ <edge from-layer="135" from-port="0" to-layer="136" to-port="1" />
2675
+ <edge from-layer="136" from-port="2" to-layer="137" to-port="0" />
2676
+ <edge from-layer="136" from-port="2" to-layer="142" to-port="1" />
2677
+ <edge from-layer="137" from-port="1" to-layer="139" to-port="0" />
2678
+ <edge from-layer="138" from-port="1" to-layer="139" to-port="1" />
2679
+ <edge from-layer="139" from-port="2" to-layer="141" to-port="0" />
2680
+ <edge from-layer="140" from-port="0" to-layer="141" to-port="1" />
2681
+ <edge from-layer="141" from-port="2" to-layer="142" to-port="2" />
2682
+ <edge from-layer="142" from-port="3" to-layer="146" to-port="0" />
2683
+ <edge from-layer="143" from-port="0" to-layer="146" to-port="1" />
2684
+ <edge from-layer="144" from-port="0" to-layer="146" to-port="2" />
2685
+ <edge from-layer="145" from-port="0" to-layer="146" to-port="3" />
2686
+ <edge from-layer="146" from-port="4" to-layer="148" to-port="0" />
2687
+ <edge from-layer="147" from-port="0" to-layer="148" to-port="1" />
2688
+ <edge from-layer="148" from-port="2" to-layer="150" to-port="0" />
2689
+ <edge from-layer="149" from-port="0" to-layer="150" to-port="1" />
2690
+ <edge from-layer="150" from-port="2" to-layer="151" to-port="0" />
2691
+ </edges>
2692
+ <rt_info>
2693
+ <MO_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2694
+ <Runtime_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3" />
2695
+ <conversion_parameters>
2696
+ <framework value="onnx" />
2697
+ <input_model value="DIR\model.onnx" />
2698
+ <model_name value="model" />
2699
+ <output_dir value="DIR" />
2700
+ </conversion_parameters>
2701
+ <legacy_frontend value="False" />
2702
+ </rt_info>
2703
+ </net>
requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ anomalib[full]