ZhengPeng7
commited on
Commit
·
ca16a0e
1
Parent(s):
fb60ce9
Add inference endpoint feature.
Browse files- README.md +0 -4
- handler.py +12 -2
- requirements.txt +0 -2
README.md
CHANGED
@@ -8,10 +8,6 @@ tags:
|
|
8 |
- model_hub_mixin
|
9 |
repo_url: https://github.com/ZhengPeng7/BiRefNet-legacy
|
10 |
pipeline_tag: image-to-image
|
11 |
-
widget:
|
12 |
-
- src: >-
|
13 |
-
https://hips.hearstapps.com/hmg-prod/images/gettyimages-1229892983-square.jpg
|
14 |
-
example_title: Portrait
|
15 |
---
|
16 |
<h1 align="center">Bilateral Reference for High-Resolution Dichotomous Image Segmentation</h1>
|
17 |
|
|
|
8 |
- model_hub_mixin
|
9 |
repo_url: https://github.com/ZhengPeng7/BiRefNet-legacy
|
10 |
pipeline_tag: image-to-image
|
|
|
|
|
|
|
|
|
11 |
---
|
12 |
<h1 align="center">Bilateral Reference for High-Resolution Dichotomous Image Segmentation</h1>
|
13 |
|
handler.py
CHANGED
@@ -9,7 +9,6 @@ from PIL import Image
|
|
9 |
import torch
|
10 |
from torchvision import transforms
|
11 |
from transformers import AutoModelForImageSegmentation
|
12 |
-
from loadimg import load_img
|
13 |
|
14 |
torch.set_float32_matmul_precision(["high", "highest"][0])
|
15 |
|
@@ -105,7 +104,18 @@ class EndpointHandler():
|
|
105 |
A :obj:`list` | `dict`: will be serialized and returned
|
106 |
"""
|
107 |
print('data["inputs"] = ', data["inputs"])
|
108 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
109 |
# Preprocess the image
|
110 |
image_preprocessor = ImagePreprocessor(resolution=tuple(resolution))
|
111 |
image_proc = image_preprocessor.proc(image)
|
|
|
9 |
import torch
|
10 |
from torchvision import transforms
|
11 |
from transformers import AutoModelForImageSegmentation
|
|
|
12 |
|
13 |
torch.set_float32_matmul_precision(["high", "highest"][0])
|
14 |
|
|
|
104 |
A :obj:`list` | `dict`: will be serialized and returned
|
105 |
"""
|
106 |
print('data["inputs"] = ', data["inputs"])
|
107 |
+
image_src = data["inputs"]
|
108 |
+
if isinstance(image_src, str):
|
109 |
+
if os.path.isfile(image_src):
|
110 |
+
image_ori = Image.open(image_src)
|
111 |
+
else:
|
112 |
+
response = requests.get(image_src)
|
113 |
+
image_data = BytesIO(response.content)
|
114 |
+
image_ori = Image.open(image_data)
|
115 |
+
else:
|
116 |
+
image_ori = Image.fromarray(image_src)
|
117 |
+
|
118 |
+
image = image_ori.convert('RGB')
|
119 |
# Preprocess the image
|
120 |
image_preprocessor = ImagePreprocessor(resolution=tuple(resolution))
|
121 |
image_proc = image_preprocessor.proc(image)
|
requirements.txt
CHANGED
@@ -16,5 +16,3 @@ prettytable
|
|
16 |
transformers
|
17 |
huggingface-hub>0.25
|
18 |
accelerate
|
19 |
-
|
20 |
-
loadimg
|
|
|
16 |
transformers
|
17 |
huggingface-hub>0.25
|
18 |
accelerate
|
|
|
|