Giters

yitu-opensource / T2T-ViT

ICCV2021, Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

load_for_transfer_learning

lurong-xing opened this issue · comments

commented

Thank you for your work!
When I use "load_for_transfer_learning(model, /path/to/pretrained/weights, use_ema=True, strict=False, num_classes=1000)" to change the num_classes, the size of output is still 1000. So what should I do when I want to change the final layer of t2t-vit to deal with different datasets?