Update app.py

app.py

@@ -125,6 +125,7 @@ def generate_response_non_streaming(instruction, model_name, temperature=0.7, ma
     try:
         # 直接使用简单的提示格式，不使用模型的聊天模板
         prompt = f"User:{instruction}\nAssistant:"
+        print("prompt:",prompt)
         chat_input = current_tokenizer.encode(prompt, return_tensors="pt").to(current_model.device)
         
         # 生成响应
@@ -137,7 +138,8 @@ def generate_response_non_streaming(instruction, model_name, temperature=0.7, ma
         )
         
         # 解码并返回生成的文本
-        generated_text = current_tokenizer.decode(output[0][len(chat_input[0]):], skip_special_tokens=True)
+        generated_text = current_tokenizer.decode(output[0], skip_special_tokens=True)#[len(chat_input[0]):]
+        print("generated_text:",generated_text)
         return generated_text
     except Exception as e:
         return f"生成响应时出错: {str(e)}"