alloldman

Results 1 issues of alloldman

It seems that only the feature extracted by the ViT trained in e2e manner is provided. Would you release the e2e trained ViT model?