app.py

# Standard library imports
import os
import threading

# Third-party imports
import gradio as gr
from peft import PeftModel
from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer

HF_TOKEN = os.getenv("HF_TOKEN")

tokenizer = AutoTokenizer.from_pretrained(
    "bunyaminergen/Qwen2.5-Coder-1.5B-Instruct-Reasoning",
    token=HF_TOKEN,
    trust_remote_code=True
)

base_model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2.5-Coder-1.5B-Instruct",
    device_map="auto",
    torch_dtype="auto",
    token=HF_TOKEN
)

base_model.resize_token_embeddings(len(tokenizer))

model = PeftModel.from_pretrained(
    base_model,
    "bunyaminergen/Qwen2.5-Coder-1.5B-Instruct-Reasoning",
    token=HF_TOKEN
)
model.eval()


def respond(
        message: str,
        history: list[tuple[str, str]],
        system_message: str,
        max_tokens: int,
        temperature: float,
        top_p: float,
):
    messages = [{"role": "system", "content": system_message}]
    for u, a in history:
        if u:
            messages.append({"role": "user", "content": u})
        if a:
            messages.append({"role": "assistant", "content": a})
    messages.append({"role": "user", "content": message})

    prompt = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

    streamer = TextIteratorStreamer(
        tokenizer,
        timeout=10.0,
        skip_prompt=True,
        skip_special_tokens=True
    )
    generation_kwargs = {
        **inputs,
        "max_new_tokens": max_tokens,
        "temperature": temperature,
        "top_p": top_p,
        "streamer": streamer,
    }
    thread = threading.Thread(target=model.generate, kwargs=generation_kwargs)
    thread.start()

    output = ""
    for chunk in streamer:
        output += chunk
        yield output


demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a helpful coding assistant.", label="System message"),
        gr.Slider(minimum=512, maximum=8192, value=2048, 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)"),
    ],
)

if __name__ == "__main__":
    demo.launch()