app.py

from fastapi import FastAPI
from qwen_classifier.predict import predict_single  # Your existing function
import torch

app = FastAPI(title="Qwen Classifier")

@app.on_event("startup")
async def load_model():
    # Warm up GPU
    torch.zeros(1).cuda() 

@app.post("/predict")
async def predict(text: str):
    return predict_single(text, backend="local")