search

yitu-opensource / T2T-ViT

ICCV2021, Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet
Other
1.15k stars 177 forks source link

The released models seem broken and cannot be opened? #2

Closed QiushiYang closed 3 years ago

QiushiYang commented 3 years ago

Thanks a lot for sharing your codes. But it seems the T2T-ViT Models are broken and I can't open them. Would you like to upload them again? Many thanks :-)

yuanli2333 commented 3 years ago

Hi, do you mean you can't load the models after downloading from google drive? What's type of error you have, could you provide more details about it?

GuideWsp commented 3 years ago

The pth.tar is not tar file, you can load it by torch.load('*.pth.tar')

Thanks a lot for sharing your codes. But it seems the T2T-ViT Models are broken and I can't open them. Would you like to upload them again? Many thanks :-)

QiushiYang commented 3 years ago

Thank you all so much! I tried to use torch.load() and could load the model successfully.