You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I notice that your model architecture is slightly different from ViT-S or DeIT-S. However, the paper doesn't provide the performance of EVA-02-S trained by prevalent training strategies (i.e. AdamW, 300 epochs, RandAug + RandErasing ).
I would be appreciate if you can provide these results.
The text was updated successfully, but these errors were encountered:
Thank you for your amazing work.
I notice that your model architecture is slightly different from ViT-S or DeIT-S. However, the paper doesn't provide the performance of EVA-02-S trained by prevalent training strategies (i.e. AdamW, 300 epochs, RandAug + RandErasing ).
I would be appreciate if you can provide these results.
The text was updated successfully, but these errors were encountered: