Update app.py

app.py

@@ -5,7 +5,7 @@ from transformers import AutoTokenizer, pipeline
 
 # Load the model and tokenizer
 model_id = "hsuwill000/DeepSeek-R1-Distill-Qwen-1.5B-openvino"
-model = OVModelForCausalLM.from_pretrained(model_id, device="CPU")  # 明確指定设备
+model = OVModelForCausalLM.from_pretrained(model_id, device="CPU")  # 明確指定設備
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 
 # Create generation pipeline
@@ -15,17 +15,17 @@ def respond(message):
     try:
         start_time = time.time()
         
-        # 強化 Prompt 讓模型輸出更合理
+        # 強化 prompt 讓模型輸出更合理
         instruction = (
-            "請用簡單、繁體中文、準確的語言回答問題，避免冗長和重複內容。\n"
+            "請用簡單、準確的語言回答問題，避免冗長和重複內容。\n"
             "User: " + message + "\n"
             "Assistant: "
         )
 
-        # Generate response with improved settings
+        # 生成回應
         response = pipe(
             instruction,
-            max_length=2048,  # 限制最大輸出長度
+            max_length=200,  # 限制最大輸出長度
             truncation=True,
             num_return_sequences=1,
             temperature=0.3,
@@ -43,24 +43,48 @@ def respond(message):
         inference_time = time.time() - start_time
         print(f"Inference time: {inference_time:.4f} seconds")
 
-        # 返回對話記錄更新結果
-        return [(message, reply)]
+        # 返回對話記錄更新結果（採用 openai-style 格式，避免未來版本警告）
+        return [{"role": "user", "content": message}, {"role": "assistant", "content": reply}]
     
     except Exception as e:
         print(f"Error: {e}")
-        return [(message, "Sorry, something went wrong. Please try again.")]
+        return [{"role": "user", "content": message}, {"role": "assistant", "content": "Sorry, something went wrong. Please try again."}]
 
-# Set up Gradio chat interface
 with gr.Blocks() as demo:
     gr.Markdown("# DeepSeek-R1-Distill-Qwen-1.5B-openvino Chat")
     gr.Markdown("Chat with DeepSeek-R1-Distill-Qwen-1.5B-openvino model.")
     
-    chatbot = gr.Chatbot()
-    # 設置 clear_on_submit=True，使得訊息送出後立即清空輸入框
-    msg = gr.Textbox(label="Your Message", clear_on_submit=True)
-    
-    # 提交後更新聊天記錄，輸入框會由 clear_on_submit 自動清空
+    # 設定 elem_id 以便後續在前端取得元素
+    chatbot = gr.Chatbot(type="messages")
+    msg = gr.Textbox(label="Your Message", elem_id="input_box")
+
+    # 點擊按鈕或按 Enter 都會送出訊息
+    send_button = gr.Button("Send")
+    # 兩種送出方式皆觸發 respond 函數
     msg.submit(respond, inputs=msg, outputs=chatbot)
+    send_button.click(respond, inputs=msg, outputs=chatbot)
+
+    # 注入 JavaScript，在按下 Enter 鍵時清空輸入框
+    gr.HTML(
+        """
+        <script>
+        // 當頁面載入完成後
+        window.addEventListener("load", function() {
+            const input_box = document.getElementById("input_box");
+            // 監聽鍵盤按下事件
+            input_box.addEventListener("keydown", function(e) {
+                // 如果按下 Enter 鍵
+                if (e.key === "Enter") {
+                    // 延遲一點時間以確保送出事件被觸發
+                    setTimeout(() => {
+                        input_box.value = "";
+                    }, 0);
+                }
+            });
+        });
+        </script>
+        """
+    )
 
 if __name__ == "__main__":
     demo.launch(share=True)