app.py

from fastapi import FastAPI
from transformers import AutoTokenizer, AutoModelForCausalLM

app = FastAPI()

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-4B")
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B")

@app.get("/")
def read_root():
    return {"message": "Model is ready"}