yandexdataschool / Practical_RL

A course in reinforcement learning in the wild
The Unlicense
5.92k stars 1.7k forks source link

Revert part of https://github.com/yandexdataschool/Practical_RL/pull/448 #462

Closed dniku closed 4 years ago

dniku commented 4 years ago

https://www.coursera.org/learn/practical-rl/discussions/all/threads/RRaAuhCbEeuoKgpcLmLqdw

I guess that either TF or Keras broke backwards compatibility, and __call__() no longer accepts Numpy tensors, contrary to what is said in https://stackoverflow.com/a/60871692.

We introduced this here: https://github.com/yandexdataschool/Practical_RL/pull/448/files#diff-61c5d5737d7f6d9831dd0d962e3e4feee9bc47dd25924cd56b16d8c499e3a43dR126

review-notebook-app[bot] commented 4 years ago

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.


Powered by ReviewNB

dniku commented 4 years ago

No review for a while, merging.