Spaces:
Build error
Build error
first commit
Browse files- app.py +48 -0
- requirements.txt +4 -0
app.py
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import matplotlib.pyplot as plt
|
2 |
+
from PIL import Image
|
3 |
+
from transformers import DetrImageProcessor, DetrForObjectDetection
|
4 |
+
import torch
|
5 |
+
|
6 |
+
# colors for visualization
|
7 |
+
COLORS = [[0.000, 0.447, 0.741], [0.850, 0.325, 0.098], [0.929, 0.694, 0.125], [0.494, 0.184, 0.556], [0.466, 0.674, 0.188]]
|
8 |
+
|
9 |
+
import io
|
10 |
+
|
11 |
+
def fig2img(fig):
|
12 |
+
buf = io.BytesIO()
|
13 |
+
fig.savefig(buf)
|
14 |
+
buf.seek(0)
|
15 |
+
img = Image.open(buf)
|
16 |
+
return img
|
17 |
+
|
18 |
+
def plot_results(image, results):
|
19 |
+
plt.figure(figsize=(16, 10))
|
20 |
+
plt.imshow(image)
|
21 |
+
ax = plt.gca()
|
22 |
+
colors = COLORS * 100
|
23 |
+
for box, label, prob, color in zip(results["boxes"], results["labels"], results["scores"], colors):
|
24 |
+
xmin, xmax, ymin, ymax = box[0].item(), box[2].item(), box[1].item(), box[3].item()
|
25 |
+
ax.add_patch(plt.Rectangle((xmin, ymin), xmax - xmin, ymax - ymin,
|
26 |
+
fill=False, color=color, linewidth=3))
|
27 |
+
text = f'{model.config.id2label[label.item()]}: {prob:0.2f}'
|
28 |
+
ax.text(xmin, ymin, text, fontsize=15,
|
29 |
+
bbox=dict(facecolor='yellow', alpha=0.5))
|
30 |
+
ax.axis("off")
|
31 |
+
return fig2img(plt.gcf())
|
32 |
+
|
33 |
+
def predict(input_img):
|
34 |
+
processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
|
35 |
+
model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")
|
36 |
+
inputs = processor(images=input_img, return_tensors="pt")
|
37 |
+
outputs = model(**inputs)
|
38 |
+
|
39 |
+
target_sizes = torch.tensor([input_img.size[::-1]])
|
40 |
+
results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]
|
41 |
+
return plot_results(input_img, results)
|
42 |
+
|
43 |
+
import gradio as gr
|
44 |
+
|
45 |
+
demo = gr.Interface(fn=predict,
|
46 |
+
inputs=gr.Image(type="pil"),
|
47 |
+
outputs="image")
|
48 |
+
demo.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch==1.12.0
|
2 |
+
torchvision==0.13.0
|
3 |
+
gradio==3.1.4
|
4 |
+
transformers==4.25.1
|