Update app.py

app.py

@@ -20,14 +20,30 @@ def describe_image(image, user_message):
     history.append(f"Assistant: {description}")
     
     full_conversation = "\n".join(history)
-    llama_result = llama_client.chat_completion(
-        messages=[{"role": "user", "content": full_conversation}],
-        max_tokens=512,
-        temperature=0.7,
-        top_p=0.95
-    )
+    return respond(user_message, history, "System: Describe the image.", 512, 0.7, 0.95)
+
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    messages = [{"role": "system", "content": system_message}]
+    
+    for val in history:
+        if val[0]:
+            messages.append({"role": "user", "content": val[0]})
+        if val[1]:
+            messages.append({"role": "assistant", "content": val[1]})
+    
+    messages.append({"role": "user", "content": message})
     
-    return description + "\n\nAssistant: " + llama_result['choices'][0]['message']['content']
+    response = ""
+    for message in llama_client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = message.choices[0].delta.content
+        response += token
+        yield response
 
 def chat_or_image(image, user_message):
     global history
@@ -35,14 +51,7 @@ def chat_or_image(image, user_message):
         return describe_image(image, user_message)
     else:
         history.append(f"User: {user_message}")
-        full_conversation = "\n".join(history)
-        llama_result = llama_client.chat_completion(
-            messages=[{"role": "user", "content": full_conversation}],
-            max_tokens=512,
-            temperature=0.7,
-            top_p=0.95
-        )
-        return llama_result['choices'][0]['message']['content']
+        return respond(user_message, history, "System: Answer the user's question.", 512, 0.7, 0.95)
 
 demo = gr.Interface(
     fn=chat_or_image,