how to set offload_model True with diffusers
#5
by zhaooooooo - opened
A100-SXM-80GB
720*1280
error:OOM
error happens below:
video = self.vae.decode(latents, return_dict=False)[0]
you can do this
pipe = WanPipeline.from_pretrained("Wan-AI/Wan2.2-T2V-A14B-Diffusers", vae=vae, torch_dtype=dtype)
pipe.enable_model_cpu_offload(device)
Thanks! This trick works!
Is that running in Zerogpu?