Update app.py

app.py

@@ -14,29 +14,41 @@ def encode_image(image):
     return base64.b64encode(buffered.getvalue()).decode('utf-8')
 
 def chat_with_image(message, image, history):
-    # Encode the image
+    # Prepare the messages
+    messages = [{"role": "system", "content": "You are a helpful assistant that can analyze images and text."}]
+    
+    for human, assistant in history:
+        if isinstance(human, dict) and 'image' in human:
+            # This is an image message
+            encoded_image = encode_image(Image.open(human['image']))
+            messages.append({
+                "role": "user",
+                "content": [
+                    {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{encoded_image}"}},
+                    {"type": "text", "text": human['text']}
+                ]
+            })
+        else:
+            # This is a text-only message
+            messages.append({"role": "user", "content": human})
+        messages.append({"role": "assistant", "content": assistant})
+    
+    # Add the current message
     if image is not None:
         encoded_image = encode_image(Image.open(image))
-        image_message = {
+        messages.append({
             "role": "user",
             "content": [
                 {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{encoded_image}"}},
                 {"type": "text", "text": message}
             ]
-        }
+        })
     else:
-        image_message = {"role": "user", "content": message}
-    
-    # Prepare the messages
-    messages = [{"role": "system", "content": "You are a helpful assistant."}]
-    for human, assistant in history:
-        messages.append({"role": "user", "content": human})
-        messages.append({"role": "assistant", "content": assistant})
-    messages.append(image_message)
+        messages.append({"role": "user", "content": message})
     
     # Call the Together AI API
     response = client.chat.completions.create(
-        model="meta-llama/Llama-Vision-Free",
+        model="meta-llama/Llama-3.2-11B-Vision-Instruct-Turbo",
         messages=messages,
         max_tokens=512,
         temperature=0.7,
@@ -64,12 +76,18 @@ with gr.Blocks() as demo:
     clear = gr.Button("Clear")
 
     def user(user_message, image, history):
-        return "", image, history + [[user_message, None]]
+        if image is not None:
+            return "", None, history + [{"text": user_message, "image": image}, None]
+        else:
+            return "", None, history + [[user_message, None]]
 
     def bot(history):
-        user_message, image = history[-1][0], None
-        if len(history) > 1 and isinstance(history[-2][0], dict):
-            image = history[-2][0]['image']
+        user_message = history[-1][0]
+        image = None
+        if isinstance(user_message, dict):
+            image = user_message['image']
+            user_message = user_message['text']
+        
         bot_message = chat_with_image(user_message, image, history[:-1])
         history[-1][1] = ""
         for character in bot_message: