import gradio as gr
import torch
from threading import Thread
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer, BitsAndBytesConfig

MODEL_ID = "huihui-ai/Huihui-Qwen3.5-35B-A3B-abliterated"

print("Loading tokenizer...")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)

print("Loading model (4-bit quantized)...")
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    quantization_config=BitsAndBytesConfig(
        load_in_4bit=True,
        bnb_4bit_compute_dtype=torch.bfloat16,
        bnb_4bit_quant_type="nf4",
        bnb_4bit_use_double_quant=True,
    ),
    device_map="auto",
    dtype=torch.bfloat16,
)
print("Model loaded!")


def chat(message, history):
    messages = [{"role": "system", "content": "You are a helpful assistant."}]
    for user_msg, bot_msg in history:
        messages.append({"role": "user", "content": user_msg})
        if bot_msg:
            messages.append({"role": "assistant", "content": bot_msg})
    messages.append({"role": "user", "content": message})

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

    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
    thread = Thread(
        target=model.generate,
        kwargs=dict(
            **inputs,
            max_new_tokens=2048,
            temperature=0.7,
            top_k=20,
            top_p=0.95,
            do_sample=True,
            streamer=streamer,
        ),
    )
    thread.start()

    partial = ""
    for token in streamer:
        partial += token
        yield partial


demo = gr.ChatInterface(
    chat,
    title="Huihui-Qwen3.5-35B-A3B Abliterated",
    description="Chat with the abliterated Qwen3.5-35B-A3B model (4-bit quantized, uncensored)",
)
demo.launch(server_name="0.0.0.0", server_port=7860)