Update app.py

app.py

@@ -1,6 +1,77 @@
+import os
+import threading
 import gradio as gr
-from huggingface_hub import InferenceClient
+import torch
+from transformers import (
+    AutoTokenizer,
+    AutoModelForCausalLM,
+    TextIteratorStreamer,
+)
+
+MODEL_ID = "Amey9766/qwen-0.6b-hospitality-housekeeping"
+
+# --- Load once (global) so it doesn't reload every message ---
+tokenizer = None
+model = None
+device = None
+
+def load_model(hf_token: str | None = None):
+    global tokenizer, model, device
+
+    if model is not None and tokenizer is not None:
+        return
+
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    dtype = torch.float16 if device == "cuda" else torch.float32
+
+    # If your repo is private/gated, you must provide a token.
+    # Priority: Gradio OAuth token -> Space secret HF_TOKEN -> None
+    use_token = hf_token or os.getenv("HF_TOKEN")
+
+    tokenizer = AutoTokenizer.from_pretrained(
+        MODEL_ID,
+        token=use_token,
+        trust_remote_code=True,
+        use_fast=True,
+    )
+
+    model = AutoModelForCausalLM.from_pretrained(
+        MODEL_ID,
+        token=use_token,
+        torch_dtype=dtype,
+        device_map="auto" if device == "cuda" else None,
+        trust_remote_code=True,
+    )
+
+    if device == "cpu":
+        model.to(device)
+
+def build_prompt(system_message: str, history: list[dict[str, str]], user_message: str) -> str:
+    """
+    Universal prompt builder.
+    Works even if the model doesn't have a strict chat template.
+    If your tokenizer supports apply_chat_template, we use it.
+    """
+    messages = [{"role": "system", "content": system_message}]
+    messages.extend(history)
+    messages.append({"role": "user", "content": user_message})
+
+    if hasattr(tokenizer, "apply_chat_template"):
+        try:
+            return tokenizer.apply_chat_template(
+                messages,
+                tokenize=False,
+                add_generation_prompt=True
+            )
+        except Exception:
+            pass
 
+    # Fallback plain prompt
+    prompt = f"System: {system_message}\n"
+    for m in history:
+        prompt += f"{m['role'].capitalize()}: {m['content']}\n"
+    prompt += f"User: {user_message}\nAssistant:"
+    return prompt
 
 def respond(
     message,
@@ -11,60 +82,51 @@ def respond(
     top_p,
     hf_token: gr.OAuthToken,
 ):
-    """
-    For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-    """
-    client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
+    # Load model (once)
+    load_model(hf_token.token if hf_token else None)
 
-    messages = [{"role": "system", "content": system_message}]
+    prompt = build_prompt(system_message, history, message)
 
-    messages.extend(history)
+    inputs = tokenizer(prompt, return_tensors="pt")
+    inputs = {k: v.to(device) for k, v in inputs.items()}
 
-    messages.append({"role": "user", "content": message})
+    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
 
-    response = ""
+    gen_kwargs = dict(
+        **inputs,
+        max_new_tokens=int(max_tokens),
+        do_sample=True,
+        temperature=float(temperature),
+        top_p=float(top_p),
+        streamer=streamer,
+    )
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        choices = message.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
+    # Run generation in background thread so streamer can yield tokens
+    thread = threading.Thread(target=model.generate, kwargs=gen_kwargs)
+    thread.start()
 
-        response += token
-        yield response
+    partial = ""
+    for token in streamer:
+        partial += token
+        yield partial
 
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
 chatbot = gr.ChatInterface(
     respond,
     type="messages",
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        gr.Textbox(value="You are a helpful housekeeping assistant for hotels.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
+        gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
     ],
 )
 
 with gr.Blocks() as demo:
     with gr.Sidebar():
+        gr.Markdown("### Login (only needed if the model repo is private/gated)")
         gr.LoginButton()
+        gr.Markdown(f"**Model:** `{MODEL_ID}`")
     chatbot.render()
 
-
 if __name__ == "__main__":
     demo.launch()