import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
import gradio as gr

model_id = "Qwen/Qwen2.5-7B-Instruct-GPTQ"

tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    trust_remote_code=True
)

def chat_with_qwen(prompt):
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    outputs = model.generate(
        **inputs,
        max_new_tokens=512,
        do_sample=True,
        temperature=0.7,
        top_p=0.9
    )
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

gr.Interface(
    fn=chat_with_qwen,
    inputs=gr.Textbox(lines=4, placeholder="Tulis pertanyaanmu di sini..."),
    outputs="text",
    title="Qwen 2.5 7B INT4 Chatbot",
    description="Ngobrol yuk sama Qwen 2.5 7B versi INT4~ 🧠✨"
).launch()