Update app.py

app.py

@@ -224,6 +224,40 @@ def on_model_series_change(model_series):
         return gr.update(choices=APOLLO_MODELS[model_series], value=APOLLO_MODELS[model_series][0])
     return gr.update(choices=[], value=None)
 
+def process_message(message, chat_history, model_series_value, model_name_value, temperature_value, max_tokens_value):
+    """Process user message and generate response"""
+    if message.strip() == "":
+        return "", chat_history
+    
+    # 打印用户提交的消息，用于调试
+    print("instruction:", message)
+    
+    # Add user message to chat history
+    chat_history = list(chat_history)
+    chat_history.append((message, None))
+    
+    # 自动加载模型（如果需要）
+    global current_model, current_tokenizer, current_model_path
+    if current_model_path != model_name_value or current_model is None:
+        try:
+            load_result = load_model(model_name_value)
+            if "failed" in load_result.lower():
+                chat_history[-1] = (message, f"模型加载失败: {load_result}")
+                return "", chat_history
+        except Exception as e:
+            chat_history[-1] = (message, f"模型加载出错: {str(e)}")
+            return "", chat_history
+    
+    # Generate response
+    try:
+        response = generate_response_non_streaming(message, model_name_value, temperature_value, max_tokens_value)
+        # Add response to chat history
+        chat_history[-1] = (message, response)
+    except Exception as e:
+        chat_history[-1] = (message, f"生成响应时出错: {str(e)}")
+    
+    return "", chat_history
+
 # Create Gradio interface
 with gr.Blocks(css=css) as demo:
     # Title and description
@@ -269,9 +303,9 @@ with gr.Blocks(css=css) as demo:
                     label="Maximum Tokens"
                 )
             
-            # Load model button
-            load_button = gr.Button("Load Model")
-            model_status = gr.Textbox(label="Model Status", value="No model loaded yet")
+            # 移除Load Model按钮和状态显示
+            # load_button = gr.Button("Load Model")
+            # model_status = gr.Textbox(label="Model Status", value="No model loaded yet")
         
         with gr.Column(scale=2):
             # Chat interface
@@ -292,40 +326,16 @@ with gr.Blocks(css=css) as demo:
         outputs=model_name
     )
     
-    # Load model
-    load_button.click(
-        fn=load_model,
-        inputs=model_name,
-        outputs=model_status
-    )
-    
-    # Bind message submission
-    def process_message(message, chat_history):
-        """Process user message and generate response"""
-        if message.strip() == "":
-            return "", chat_history
-        
-        # Add user message to chat history
-        chat_history = list(chat_history)
-        chat_history.append((message, None))
-        
-        # Generate response
-        response = generate_response_non_streaming(message, model_name.value, temperature.value, max_tokens.value)
-        
-        # Add response to chat history
-        chat_history[-1] = (message, response)
-        
-        return "", chat_history
-
+    # 修改提交事件绑定
     submit_event = user_input.submit(
         fn=process_message,
-        inputs=[user_input, chatbot],
+        inputs=[user_input, chatbot, model_series, model_name, temperature, max_tokens],
         outputs=[user_input, chatbot]
     )
 
     submit_button.click(
         fn=process_message,
-        inputs=[user_input, chatbot],
+        inputs=[user_input, chatbot, model_series, model_name, temperature, max_tokens],
         outputs=[user_input, chatbot]
     )