app.py

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

# Загружаем модель Qwen локально
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-7B", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-7B", device_map="auto", torch_dtype=torch.float16, trust_remote_code=True)

def respond(message):
    inputs = tokenizer(message, return_tensors="pt").to(model.device)
    outputs = model.generate(**inputs, max_new_tokens=200)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response

# Создаём интерфейс
gr.Interface(
    fn=respond,
    inputs=gr.Textbox(label="Ваше сообщение"),
    outputs=gr.Textbox(label="Qwen отвечает"),
    title="Qwen Прокси",
    description="Это API-прокси для Janotaro.ai"
).launch()