Hi folks, thanks for you to release the repo. I am working on the efficiency training of ViT and your paper mentions that "it bears the flexibility to be extended to various types of layers, including convolution and self-attention layers." I am wondering if you intend to release such version repo for ViT or if you have some advice regarding extending the repo to ViT? Thanks for your help!
Very sorry that I mis-open an issue and I have closed it. Please never mind about that one.
Hi folks, thanks for you to release the repo. I am working on the efficiency training of ViT and your paper mentions that "it bears the flexibility to be extended to various types of layers, including convolution and self-attention layers." I am wondering if you intend to release such version repo for ViT or if you have some advice regarding extending the repo to ViT? Thanks for your help!
Very sorry that I mis-open an issue and I have closed it. Please never mind about that one.