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

which Label Smoothing is better? : tf's categoricalcrossentropy vs custom function #13

Open

MokkeMeguru opened this issue Mar 26, 2020 · 1 comment

MokkeMeguru commented Mar 26, 2020

Hello, I found label smoothing can lead us better performance for the transformer.

But I wonder why you implement label smoothing despite there is the argument of label smoothing in tf.keras.losses.CategoricalCrossEntropy (https://www.tensorflow.org/api_docs/python/tf/keras/losses/CategoricalCrossentropy?version=nightly)

So, I want to get your answer to this question.
Why you implement the custom function for label smoothing?

github-actions bot commented Mar 26, 2020

Message that will be displayed on users' first issue

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment