You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Remove State::UndoAction from catch and cliff_walking. Properly imple…
…menting this function requires significant changes to the implementation (extra nontrivial book-keeping).
PiperOrigin-RevId: 397732772
Change-Id: I7f5b5d4d1bcb94cf39e7f2b699f2b21259f6b0a9
Add catch to SARSA and Q-learning tests and fix a bug by which the al…
…gorithms didn't handle games with chance start nodes.
PiperOrigin-RevId: 359502727
Change-Id: Iaf8f9ce9d640d33d2765c7e9b41a13c9aef23fb7
Add catch to SARSA and Q-learning tests and fix a bug by which the al…
…gorithms didn't handle games with chance start nodes.
PiperOrigin-RevId: 359502727
Change-Id: Iaf8f9ce9d640d33d2765c7e9b41a13c9aef23fb7
Copybara import of the project:
--
10a55ff by Asugawara <asgasw@gmail.com>:
add nfsp
--
7908ccc by Asugawara <asgasw@gmail.com>:
add Sonnet Linear Module
--
5671c9f by Asugawara <asgasw@gmail.com>:
action_probs: LongTensor to Tensor
--
b6b9d7d by Asugawara <asgasw@gmail.com>:
remove image and progress
COPYBARA_INTEGRATE_REVIEW=google-deepmind#450 from Asugawara:nfsp_pytorch b6b9d7d
PiperOrigin-RevId: 345889227
Change-Id: Ib5558b3e05f4cfe96c1a9854a6956100b03ee2d4