Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

TripletLoss will produce only NaN when using mixed bfloat16 precision #364

Open
kbelenky opened this issue Oct 16, 2023 · 0 comments
Open

TripletLoss will produce only NaN when using mixed bfloat16 precision #364

kbelenky opened this issue Oct 16, 2023 · 0 comments

Comments

@kbelenky
Copy link

If you use the TripletLoss loss function with mixed bfloat16 precision:

mixed_precision.set_global_policy('mixed_bfloat16')

As described here:
https://www.tensorflow.org/guide/mixed_precision#setting_the_dtype_policy

Then the value for the loss will only ever be NaN.

The expected behavior is for the loss value to be a non-NaN value.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@kbelenky