V2XL weights #776
Unanswered
benihime91
asked this question in
General
Replies: 1 comment 1 reply
-
@benihime91 I have converted them, but there is a fairly significant drop validating in PyTorch with standard imagenet eval preprocessing compared to their numbers in TF. The other weights finetuned from 21k were also showing bigger drops than normal but the XL is to the point where it ends up worse than the L. I was going to look into it more.... |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I noticed that efficientnetv2_XL model was added but no weights ... When can we expect the weights release ? 😳
Beta Was this translation helpful? Give feedback.
All reactions