Update app.py

app.py

@@ -3,73 +3,79 @@ from transformers import BlipForQuestionAnswering, AutoProcessor
 from PIL import Image
 import gradio as gr
 
-# Load the BLIP model and processor
 model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
 processor = AutoProcessor.from_pretrained("Salesforce/blip-vqa-base")
 
-# Define your Groq API key and endpoint
-groq_api_key = "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"  # Replace with your Groq API key
-groq_api_url = "https://api.groq.com/openai/v1/chat/completions"  # Replace with the appropriate Groq endpoint
+groq_api_key = "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"
+groq_api_url = "https://api.groq.com/openai/v1/chat/completions"
 
-def qna(image, question, context):
+def qna(image, question, history):
     try:
-        # Step 1: Get initial short answer from BLIP
         inputs = processor(image, question, return_tensors="pt")
         out = model.generate(**inputs)
         short_answer = processor.decode(out[0], skip_special_tokens=True)
+        
+        context = "\n".join([f"Q: {q}\nA: {a}" for q, a in history])
+        full_prompt = f"""Context of previous conversation:
+{context}
 
-        # Step 2: Construct prompt for Groq API
-        full_prompt = f"{context}\nUser: {question}\nBLIP: {short_answer}\nAssistant:"
+Current Image Description: {short_answer}
+Question: {question}
+Please provide a detailed answer based on the image and previous context."""
         
-        # Step 3: Send prompt to Groq API for a detailed answer
         headers = {
             "Authorization": f"Bearer {groq_api_key}",
             "Content-Type": "application/json"
         }
         
         data = {
-            "model": "llama3-8b-8192",  # Specify the model to use
+            "model": "llama3-8b-8192",
             "messages": [{"role": "user", "content": full_prompt}]
         }
         
         response = requests.post(groq_api_url, headers=headers, json=data)
-
-        # Check if the response is successful
+        
         if response.status_code == 200:
-            detailed_answer = response.json().get('choices', [])[0].get('message', {}).get('content', '').strip()
-            # Update the context with the latest question and answer
-            updated_context = f"{context}\nUser: {question}\nAssistant: {detailed_answer}"
-            return updated_context, updated_context  # Return updated context for display
+            detailed_answer = response.json()['choices'][0]['message']['content'].strip()
+            history.append((question, detailed_answer))
+            return history, history
         else:
-            return f"Error {response.status_code}: {response.text}", context
-
+            error_msg = f"Error {response.status_code}: {response.text}"
+            history.append((question, error_msg))
+            return history, history
+            
     except Exception as e:
-        return f"An error occurred: {str(e)}", context
+        error_msg = f"An error occurred: {str(e)}"
+        history.append((question, error_msg))
+        return history, history
 
-# Create Gradio interface with context management
-def chatbot_interface(image, question, context=""):
-    # Initialize context if image is uploaded
-    if context == "" and image is not None:
-        context = ""  # Reset context when the image is first uploaded
+def clear_history():
+    return [], []
 
-    # Get the answer from the model
-    answer, updated_context = qna(image, question, context)
-
-    # Return the updated context for display
-    return updated_context
-
-# Define the Gradio interface
-interf = gr.Interface(
-    fn=chatbot_interface,
-    inputs=[
-        gr.Image(type="pil", label="Upload Image"), 
-        gr.Textbox(label="Ask a question")
-    ],
-    outputs="text",  # Output the full conversation context
-    title="Interactive Image Chatbot",
-    description="Upload an image and have a conversation about it. Ask multiple questions about the image."
-)
+with gr.Blocks() as demo:
+    gr.Markdown("# Interactive Image Chatbot")
+    
+    with gr.Row():
+        image_input = gr.Image(type="pil")
+        
+    with gr.Row():
+        with gr.Column():
+            chatbot = gr.Chatbot()
+            question = gr.Textbox(label="Ask a question about the image")
+            clear = gr.Button("Clear Conversation")
+    
+    state = gr.State([])
+    
+    question.submit(
+        qna,
+        inputs=[image_input, question, state],
+        outputs=[chatbot, state]
+    )
+    
+    clear.click(
+        clear_history,
+        outputs=[chatbot, state]
+    )
 
-# Launch the interface
 if __name__ == "__main__":
-    interf.launch()
+    demo.launch()