lightx2v

/

Qwen-Image-Lightning

like 394

IF I WANT TO USE LORA WEIGHT

pipe = DiffusionPipeline.from_pretrained(
it should be "Qwen/Qwen-Image-Edit" or "Qwen/Qwen-Image",

scheduler=scheduler, torch_dtype=torch.bfloat16
).to("cuda")
pipe.load_lora_weights(
"lightx2v/Qwen-Image-Lightning", weight_name="Qwen-Image-Lightning-4steps-V1.0.safetensors"
)