T2T-ViT
T2T-ViT copied to clipboard
training T2T-ViT is slower than training resnet?
Have you compaired the inference speed of T2T-ViT and resnet? At the same accuracy, which famaliy of models achieve higher FPS?
You can use 128x8 for 8 GPU if you want to train T2T-ViT faster, cause I only had 2080Ti before, so I use batch size=64x8