Skip to content

Fix not evaluating the optimizer state after each step#381

Open
waldheinz wants to merge 1 commit intofilipstrand:mainfrom
waldheinz:fix-train-eval
Open

Fix not evaluating the optimizer state after each step#381
waldheinz wants to merge 1 commit intofilipstrand:mainfrom
waldheinz:fix-train-eval

Conversation

@waldheinz
Copy link
Contributor

Not doing this causes huge memory use if neither "should_plot_loss", "should_generate_image" or "should_save" force an evaluation in the code below.

I found this when setting plot_frequency to 32 because evaluating the loss using the validation set is not cheap at all. But this is a separate issue.

Not doing this causes huge memory use if neither "should_plot_loss",
"should_generate_image" or "should_save" force an evaluation in the
code below.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant