rosinality / vq-vae-2-pytorch

Implementation of Generating Diverse High-Fidelity Images with VQ-VAE-2 in PyTorch
Other
1.65k stars 275 forks source link

The hardware required. #16

Closed nobodypengium closed 5 years ago

nobodypengium commented 5 years ago

Thanks for your implementation. Could you tell me some details about the nums of GPUs you used, the training duration on your platform, and the memory occupied? Thanks!

rosinality commented 5 years ago

I have trained VQ-VAE with 1 P40. I don't remember training duration well, but it can be trained in short times. PixelSNAIL trained with 3-4 V100s with mixed precision training, and training duration was 3-5 days.

nobodypengium commented 5 years ago

Thanks a lot for your reply!