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
{{ message }}
This repository has been archived by the owner on May 1, 2024. It is now read-only.
Hi,
Is there comparison with the scenario where the model is trained with cropsize of 384 directly?
Is training with cropsize=384 from scratch better or worse than the proposed 224-pretrain-and-384-finetune?
The text was updated successfully, but these errors were encountered: