import torch from diffusers import FluxPipeline from nunchaku import NunchakuFluxTransformer2dModel from nunchaku.utils import get_precision precision = get_precision() # auto-detect your precision is 'int4' or 'fp4' based on your GPU transformer = NunchakuFluxTransformer2dModel.from_pretrained( f"nunchaku-tech/nunchaku-flux.1-krea-dev/svdq-{precision}_r32-flux.1-krea-dev.safetensors" ) pipeline = FluxPipeline.from_pretrained( "black-forest-labs/FLUX.1-krea-dev", torch_dtype=torch.bfloat16, transformer=transformer ).to("cuda") prompt = ( "Tiny paper origami kingdom, a river flowing through a lush valley, bright saturated image," "a fox to the left, deer to the right, birds in the sky, bushes and tress all around" ) image = pipeline(prompt, height=1024, width=1024, guidance_scale=4.5).images[0] image.save("flux-krea-dev.png")