Update app.py

app.py

@@ -12,10 +12,11 @@ processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
 
 i1 = gr.inputs.Image(type="pil", label="Input image")
 i2 = gr.inputs.Textbox(label="Input text")
+i3 = gr.inputs.Number(default=0.96, label="Threshold percentage score")
 o1 = gr.outputs.Image(type="pil", label="Cropped part")
 o2 = gr.outputs.Textbox(label="Similarity score")
 
-def extract_image(image, text, num=1):
+def extract_image(image, text, num=1, prob):
     
     inputs = feature_extractor(images=image, return_tensors="pt")
     outputs = dmodel(**inputs)
@@ -25,7 +26,7 @@ def extract_image(image, text, num=1):
     bboxes = outputs.pred_boxes
     probas = outputs.logits.softmax(-1)[0, :, :-1] #removing no class as detr maps 
     
-    keep = probas.max(-1).values > 0.96
+    keep = probas.max(-1).values > prob
     outs = feature_extractor.post_process(outputs, torch.tensor(image.size[::-1]).unsqueeze(0))
     bboxes_scaled = outs[0]['boxes'][keep].detach().numpy()
     labels = outs[0]['labels'][keep].detach().numpy()
@@ -65,6 +66,6 @@ def extract_image(image, text, num=1):
 
 title = "ClipnCrop"
 description = "Extract sections of images from your image by using OpenAI's CLIP and Facebooks Detr implemented on HuggingFace Transformers" 
-examples=[['ex3.jpg', 'black bag'],['ex2.jpg', 'man in red dress']]
+examples=[['ex3.jpg', 'black bag', 0.96],['ex2.jpg', 'man in red dress', 0.85]]
 article = "<p style='text-align: center'><a href='https://github.com/Vishnunkumar/clipcrop' target='_blank'>clipcrop</a></p>"
-gr.Interface(fn=extract_image, inputs=[i1, i2], outputs=[o1, o2], title=title, description=description, article=article, examples=examples, enable_queue=True).launch()    
+gr.Interface(fn=extract_image, inputs=[i1, i2, i3], outputs=[o1, o2], title=title, description=description, article=article, examples=examples, enable_queue=True).launch()