Update app.py

app.py

@@ -11,41 +11,50 @@ logging.set_verbosity_error()
 model = BlipForImageTextRetrieval.from_pretrained("Salesforce/blip-itm-base-coco")
 processor = AutoProcessor.from_pretrained("Salesforce/blip-itm-base-coco")
 
-def process_image(input_type, image_url, image_upload):
+def process_image(input_type, image_url, image_upload, text):
     if input_type == "URL":
         raw_image =  Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
     else:
-        raw_image = Image.open(image_upload)
+        raw_image = image_upload
 
     inputs = processor(images=raw_image, text=text, return_tensors="pt")
     itm_scores = model(**inputs)[0]
     itm_score = torch.nn.functional.softmax(itm_scores,dim=1)
     itm_score = itm_score[0][1]
     print(itm_score)
+
     if itm_score <=.35:
-        cmnt = "Your description is not that great. Try again"
+        cmnt = "and is not that great. Try again"
     elif itm_score <= .75:
-        cmnt = "Your description is good. But you can improve it. Try again"
+        cmnt = "and is good. But you can improve it. Try again"
     else:
-        cmnt = "Your description is excellent.Can you improve on it?"
+        cmnt = "and is excellent. Can you improve on it?"
         
     formatted_text = (
-        f"""<div style='text-align: center; font-size: 15px; color: blue;'>
-        Your decription is <span style='font-size: 20px; color: orange;'>{itm_score}</span> matching, {cmnt}
+        f"""<div style='text-align: center; font-size: 40px; color: blue;'>
+        Your decription score is <span style='font-size: 60px; color: orange;'>{itm_score:.4f}</span>; {cmnt}
         </div>"""
     )
     return formatted_text
 
-demo = gr.Interface(title= "Challenge yourself best describing the image",
-                    description = 'Upload an image or type in image URL and submit',
-    fn=process_image,
-    inputs=[
-        gr.inputs.Radio(choices=["URL", "Upload"], label="Input Type"),
-        gr.inputs.Textbox(label="Image URL", visible=False),
-        gr.inputs.Image(type="file", label="Upload Image", visible=False)
-    ],
-    outputs=gr.outputs.Image(type="pil", label="Processed Image"),
-    layout="vertical"
-)
+def toggle_inputs(input_type):
+    if input_type == "URL":
+        return gr.update(visible=True), gr.update(visible=False), gr.update(visible=True)
+    else:
+        return gr.update(visible=False), gr.update(visible=True), gr.update(visible=True)
+
+with gr.Blocks() as demo:
+    input_type = gr.Radio(choices=["URL", "Upload"], label="Input Type")
+    image_url = gr.Textbox(label="Image URL", visible=False)
+    image_upload = gr.Image(type="pil", label="Upload Image", visible=False)
+    description = gr.Textbox(label="Describe the image", visible=False, lines=3)
+
+    input_type.change(fn=toggle_inputs, inputs=input_type, outputs=[image_url, image_upload, description])
+    
+    submit_btn = gr.Button("Submit")
+    processed_image = gr.HTML(label="Your challenge result")
+
+    submit_btn.click(fn=process_image, inputs=[input_type, image_url, image_upload, description], outputs=processed_image)
+
 
 demo.launch(share=True, debug=True)