search

yitu-opensource / T2T-ViT

ICCV2021, Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet
Other
1.15k stars 176 forks source link

data preprocessing #74

Open abcsimple opened 1 year ago

abcsimple commented 1 year ago

hi, did you normolize the training data by mean and std values when you training the pretrained models?