module must have its parameters and buffers on device cuda:0 (device_ids[0]) but found one of them on device: cpu

[AEP-WYK]

when I run the "python generate_samples.py --generator_file "../checkpoint/GAN_GEN_SHADOW_8.pth" --latent_size 512"

[Animesh Karnewar]

Are you trying to do generation of samples on the cpu? While you trained on the GPU. There is a quick fix for this with map_location argument of torch.load. More details here -> https://pytorch.org/docs/1.9.1/generated/torch.load.html.