fix

app.py

@@ -8,6 +8,37 @@ HF_MODEL_ID = "rieon/DeepCoder-14B-Preview-Suger"
 client = InferenceClient(model=HF_MODEL_ID)
 
 def respond(
+    message: str,
+    history: list[dict],       # [{"role":"user"/"assistant","content":…}, …]
+    system_message: str,
+    max_tokens: int,
+    temperature: float,
+    top_p: float,
+):
+    # 1️⃣ Build one raw-text prompt from system + chat history + new user turn
+    prompt = system_message.strip() + "\n"
+    for msg in history:
+        role = msg["role"]
+        content = msg["content"]
+        if role == "user":
+            prompt += f"User: {content}\n"
+        elif role == "assistant":
+            prompt += f"Assistant: {content}\n"
+    prompt += f"User: {message}\nAssistant:"
+
+    # 2️⃣ Stream tokens from the text-generation endpoint
+    generated = ""
+    for chunk in client.text_generation(
+        prompt,                     # first positional arg
+        max_new_tokens=max_tokens,
+        temperature=temperature,
+        top_p=top_p,
+        stream=True,
+    ):
+        generated += chunk.generated_text
+        yield generated
+
+def respond2(
     message: str,
     history: list[tuple[str, str]],
     system_message: str,