Fix wrong dimension in Temporal Activation Regularization in RNNTrainer?

fastai / course-v3

The 3rd edition of course.fast.ai

https://course.fast.ai/

Apache License 2.0

4.9k stars 3.55k forks source link

Closed fg91 closed 4 years ago

fg91 commented 5 years ago

Current:

if len(h) > 1: ... apply TAR

However, len(h) gives the batch size and not the sequence length. Should this be changed to the following?

if h.size(1) > 1: ... apply TAR

review-notebook-app[bot] commented 5 years ago

Check out this pull request on

You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB.