app.py

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from multiprocessing import Process, Queue
from diffusers import FluxPipeline
import torch
import io
from fastapi.responses import StreamingResponse
import uvicorn

app = FastAPI()

pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16, revision="main")
pipe.enable_model_cpu_offload()

class ImageRequest(BaseModel):
    prompt: str

def generate_image_response(request, queue):
    try:
        image = pipe(
            request.prompt,
            guidance_scale=0.0,
            num_inference_steps=4,
            max_sequence_length=256,
            generator=torch.Generator("cpu").manual_seed(0)
        ).images[0]

        img_io = io.BytesIO()
        image.save(img_io, 'PNG')
        img_io.seek(0)
        queue.put(img_io.getvalue())
    except Exception as e:
        queue.put(f"Error: {str(e)}")

@app.post("/generate_image")
async def generate_image(request: ImageRequest):
    queue = Queue()
    p = Process(target=generate_image_response, args=(request, queue))
    p.start()
    p.join()
    response = queue.get()
    if "Error" in response:
        raise HTTPException(status_code=500, detail=response)
    return StreamingResponse(io.BytesIO(response), media_type="image/png")

if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8002)