Hugging Face's logo

qualcomm
/
Lightweight-Face-Detection

Object Detection
PyTorch
TF Lite
ONNX
real_time
android
Model card Files Community
bhushans commited on
Commit
f63ba2f
·
verified ·
1 Parent(s): 5d665e6

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +43 -25
README.md CHANGED
@@ -36,29 +36,31 @@ More details on model performance across various devices, can be found
36
 
37
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
- | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 1.022 ms | 0 - 3 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
40
- | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 1.103 ms | 2 - 7 MB | FP16 | NPU | [Lightweight-Face-Detection.so](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.so) |
41
- | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.237 ms | 0 - 3 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
42
- | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.682 ms | 0 - 32 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
43
- | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.735 ms | 0 - 12 MB | FP16 | NPU | [Lightweight-Face-Detection.so](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.so) |
44
- | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 1.375 ms | 0 - 40 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
45
- | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.75 ms | 0 - 19 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
46
- | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.756 ms | 0 - 10 MB | FP16 | NPU | Use Export Script |
47
- | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 1.214 ms | 0 - 23 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
48
- | Lightweight-Face-Detection | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 1.053 ms | 0 - 1 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
49
- | Lightweight-Face-Detection | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 1.074 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
50
- | Lightweight-Face-Detection | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 1.019 ms | 0 - 1 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
51
- | Lightweight-Face-Detection | SA8255 (Proxy) | SA8255P Proxy | QNN | 1.08 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
52
- | Lightweight-Face-Detection | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 1.019 ms | 0 - 1 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
53
- | Lightweight-Face-Detection | SA8775 (Proxy) | SA8775P Proxy | QNN | 1.079 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
54
- | Lightweight-Face-Detection | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 1.019 ms | 0 - 1 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
55
- | Lightweight-Face-Detection | SA8650 (Proxy) | SA8650P Proxy | QNN | 1.077 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
56
- | Lightweight-Face-Detection | SA8295P ADP | SA8295P | TFLITE | 1.769 ms | 0 - 18 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
57
- | Lightweight-Face-Detection | SA8295P ADP | SA8295P | QNN | 1.984 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
58
- | Lightweight-Face-Detection | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.277 ms | 0 - 31 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
59
- | Lightweight-Face-Detection | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.354 ms | 1 - 15 MB | FP16 | NPU | Use Export Script |
60
- | Lightweight-Face-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 1.284 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
61
- | Lightweight-Face-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.044 ms | 3 - 3 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
 
 
62
 
63
 
64
 
@@ -123,7 +125,7 @@ Lightweight-Face-Detection
123
  Device : Samsung Galaxy S23 (13)
124
  Runtime : TFLITE
125
  Estimated inference time (ms) : 1.0
126
- Estimated peak memory usage (MB): [0, 3]
127
  Total # Ops : 84
128
  Compute Unit(s) : NPU (84 ops)
129
  ```
@@ -144,13 +146,29 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
144
  import torch
145
 
146
  import qai_hub as hub
147
- from qai_hub_models.models.face_det_lite import
148
 
149
  # Load the model
 
150
 
151
  # Device
152
  device = hub.Device("Samsung Galaxy S23")
153
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
154
 
155
  ```
156
 
 
36
 
37
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 1.033 ms | 0 - 4 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
40
+ | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 1.106 ms | 1 - 27 MB | FP16 | NPU | [Lightweight-Face-Detection.so](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.so) |
41
+ | Lightweight-Face-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 2.269 ms | 1 - 3 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
42
+ | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.682 ms | 0 - 15 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
43
+ | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.733 ms | 0 - 13 MB | FP16 | NPU | [Lightweight-Face-Detection.so](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.so) |
44
+ | Lightweight-Face-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 1.375 ms | 0 - 42 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
45
+ | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.753 ms | 0 - 9 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
46
+ | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.628 ms | 1 - 12 MB | FP16 | NPU | Use Export Script |
47
+ | Lightweight-Face-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 1.216 ms | 0 - 25 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
48
+ | Lightweight-Face-Detection | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 1.025 ms | 0 - 4 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
49
+ | Lightweight-Face-Detection | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 1.092 ms | 0 - 2 MB | FP16 | NPU | Use Export Script |
50
+ | Lightweight-Face-Detection | SA7255P ADP | SA7255P | TFLITE | 22.468 ms | 0 - 11 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
51
+ | Lightweight-Face-Detection | SA7255P ADP | SA7255P | QNN | 22.643 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
52
+ | Lightweight-Face-Detection | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 1.031 ms | 0 - 5 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
53
+ | Lightweight-Face-Detection | SA8255 (Proxy) | SA8255P Proxy | QNN | 1.085 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
54
+ | Lightweight-Face-Detection | SA8295P ADP | SA8295P | TFLITE | 1.744 ms | 0 - 9 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
55
+ | Lightweight-Face-Detection | SA8295P ADP | SA8295P | QNN | 2.153 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
56
+ | Lightweight-Face-Detection | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 1.034 ms | 0 - 5 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
57
+ | Lightweight-Face-Detection | SA8650 (Proxy) | SA8650P Proxy | QNN | 1.089 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
58
+ | Lightweight-Face-Detection | SA8775P ADP | SA8775P | TFLITE | 1.925 ms | 0 - 8 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
59
+ | Lightweight-Face-Detection | SA8775P ADP | SA8775P | QNN | 2.14 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
60
+ | Lightweight-Face-Detection | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.263 ms | 0 - 13 MB | FP16 | NPU | [Lightweight-Face-Detection.tflite](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.tflite) |
61
+ | Lightweight-Face-Detection | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.345 ms | 1 - 16 MB | FP16 | NPU | Use Export Script |
62
+ | Lightweight-Face-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 1.236 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
63
+ | Lightweight-Face-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.04 ms | 1 - 1 MB | FP16 | NPU | [Lightweight-Face-Detection.onnx](https://huggingface.co/qualcomm/Lightweight-Face-Detection/blob/main/Lightweight-Face-Detection.onnx) |
64
 
65
 
66
 
 
125
  Device : Samsung Galaxy S23 (13)
126
  Runtime : TFLITE
127
  Estimated inference time (ms) : 1.0
128
+ Estimated peak memory usage (MB): [0, 4]
129
  Total # Ops : 84
130
  Compute Unit(s) : NPU (84 ops)
131
  ```
 
146
  import torch
147
 
148
  import qai_hub as hub
149
+ from qai_hub_models.models.face_det_lite import Model
150
 
151
  # Load the model
152
+ torch_model = Model.from_pretrained()
153
 
154
  # Device
155
  device = hub.Device("Samsung Galaxy S23")
156
 
157
+ # Trace model
158
+ input_shape = torch_model.get_input_spec()
159
+ sample_inputs = torch_model.sample_inputs()
160
+
161
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
162
+
163
+ # Compile model on a specific device
164
+ compile_job = hub.submit_compile_job(
165
+ model=pt_model,
166
+ device=device,
167
+ input_specs=torch_model.get_input_spec(),
168
+ )
169
+
170
+ # Get target model to run on-device
171
+ target_model = compile_job.get_target_model()
172
 
173
  ```
174