We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
ICCV2021, Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet