Update app.py

app.py

@@ -1,68 +1,112 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 
+import os
+from typing import List, Tuple
 
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    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="Balab2021/qwen-workflow-planner-qwen2p5-lora")
-
-    messages = [{"role": "system", "content": system_message}]
-
-    messages.extend(history)
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    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
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-chatbot = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", 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)",
-        ),
-    ],
-)
-
-with gr.Blocks() as demo:
-    with gr.Sidebar():
-        gr.LoginButton()
-    chatbot.render()
+import gradio as gr
+import torch
+from dotenv import load_dotenv
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+load_dotenv()
+
+MODEL_ID = "Balab2021/qwen-workflow-planner-qwen2p5-lora"
+
+def get_hf_token() -> str:
+	if not HF_TOKEN_KEYS:
+		raise RuntimeError(
+			"Missing HF_TOKEN_KEYS environment variable. "
+			"Set it to one or more token env var names (comma-separated), "
+			"for example: HF_TOKEN_KEYS=HF_TOKEN"
+		)
+
+	raw_value = HF_TOKEN_KEYS.strip().strip("\"'")
+
+	# Allow HF_TOKEN_KEYS to hold a direct Hugging Face token.
+	if raw_value.startswith("hf_"):
+		return raw_value
+
+	keys = [key.strip() for key in raw_value.split(",") if key.strip()]
+
+	if not keys:
+		raise RuntimeError(
+			"HF_TOKEN_KEYS is empty. "
+			"Set it to one or more token env var names, for example: HF_TOKEN"
+		)
+
+	for key in keys:
+		token = os.getenv(key)
+		if token:
+			return token.strip().strip("\"'")
+	raise RuntimeError(
+		"Missing Hugging Face token. None of the env vars listed in "
+		f"HF_TOKEN_KEYS contain a token value. Checked keys: {', '.join(keys)}"
+	)
+
+
+def build_messages(history: List[Tuple[str, str]], user_message: str):
+	messages = []
+	for user_text, assistant_text in history:
+		if user_text:
+			messages.append({"role": "user", "content": user_text})
+		if assistant_text:
+			messages.append({"role": "assistant", "content": assistant_text})
+	messages.append({"role": "user", "content": user_message})
+	return messages
+
+
+def create_app():
+	load_dotenv()
+	token = get_hf_token()
+
+	tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, token=token)
+	model = AutoModelForCausalLM.from_pretrained(
+		MODEL_ID,
+		token=token,
+		torch_dtype="auto",
+		device_map="auto",
+	)
+
+	def chat_fn(
+		message: str,
+		history: List[Tuple[str, str]],
+		temperature: float,
+		max_new_tokens: int,
+	) -> str:
+		messages = build_messages(history, message)
+		prompt = tokenizer.apply_chat_template(
+			messages,
+			tokenize=False,
+			add_generation_prompt=True,
+		)
+
+		inputs = tokenizer(prompt, return_tensors="pt")
+		inputs = {k: v.to(model.device) for k, v in inputs.items()}
+
+		with torch.no_grad():
+			output_ids = model.generate(
+				**inputs,
+				max_new_tokens=max_new_tokens,
+				temperature=temperature,
+				do_sample=temperature > 0,
+				pad_token_id=tokenizer.eos_token_id,
+			)
+
+		generated_ids = output_ids[0][inputs["input_ids"].shape[-1] :]
+		response = tokenizer.decode(generated_ids, skip_special_tokens=True).strip()
+		return response
+
+	demo = gr.ChatInterface(
+		fn=chat_fn,
+		additional_inputs=[
+			gr.Slider(0.0, 1.5, value=0.2, step=0.05, label="Temperature"),
+			gr.Slider(32, 2048, value=512, step=32, label="Max New Tokens"),
+		],
+		title="Qwen Workflow Planner Chat",
+		description=f"Model: {MODEL_ID}",
+	)
+	return demo
 
 
 if __name__ == "__main__":