raoyongming / DynamicViT

[NeurIPS 2021] [T-PAMI] DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification
https://dynamicvit.ivg-research.xyz/
MIT License
551 stars 69 forks source link

pretrained model download #7

Closed dirtycomputer closed 2 years ago

dirtycomputer commented 2 years ago

!wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1iZc6d27EuEnlfUpJoNhsZEkt6GVgPy7-' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1iZc6d27EuEnlfUpJoNhsZEkt6GVgPy7-" -O lvvit_m-56M-224-84.0.tar && rm -rf /tmp/cookies.txt I can use the code above download lvvit_m-56M-224-84.0.tar But why I can't use !tar -xvf lvvit_m-56M-224-84.0.tar to extract the file It says

tar: This does not look like a tar archive tar: Skipping to next header tar: Exiting with failure status due to previous errors

raoyongming commented 2 years ago

Hi @dirtycomputer,

lvvit_m-56M-224-84.0.tar is the checkpoint file of the pretrained model. You can directly use `torch.load('lvvit_m-56M-224-84.0.tar')' to load the pretrained model. You don't need to untar the file.

dirtycomputer commented 2 years ago

Tks