Dear authors,
I understand that it takes time to craft and release the paper. As an alternative, could you release your pretraining hyperparameters here first?
· Sign up or log in to comment
