Closed nicofirst1 closed 3 years ago
Thanks, except minor comments it looks nice!
Nit: Do you know if rich is faster than tqdm or did you do any test? I mean rich-callback vs tqdm-callback or plain tqdm.
For a comparison between tqdm and rich you can refer to this issue
yep I read those already but it mentions comparable performance for quick iterations, it'd be nice to check for long runs like your coco use case but I think it is not so important now, we can leave it as is
Is everything ok with this PR? I have some updates but would like to create a new PR based on this one
Except these (very) minor comments it looks good to me! If @eugene-kharitonov agrees we could merge Meanwhile you could branch out of your own branch and commit the updates there
Adding progress bar for epochs both in training and testing.