nnizhang / VST

source code of “Visual Saliency Transformer” (ICCV2021)
123 stars 26 forks source link

Pretrained T2T-ViT model can't be opened. #3

Open wanghaitong-q opened 2 years ago

wanghaitong-q commented 2 years ago

Thanks for your hard work.

I find a problem in your project on github, that is the pretrained T2T-ViT_t-14 model couldn't be opened. There are always problems, regarsless of in windows or Linux.

nnizhang commented 2 years ago

Is the format of the downloaded model 80.7_T2T_ViT_t_14.pth.tar? You do not need to open it and you just need to change the model path in the codes.