Marc Lanctot

Marked R-NaD code as unmaintained. https://github.com/google-deepmind/open_spiel/commit/d99705de2cca7075e12fbbd76443fcc123249d6f

Hi @Atan03, Ok so there are some examples of how to use those subgames here: https://github.com/google-deepmind/open_spiel/blob/d8a3babc2703d58dfd7a86974c91a28b9593b90a/open_spiel/games/universal_poker/universal_poker_test.cc#L662 My guess is this was never exposed to Python.. yeah doesn't look like it....

Closing due to inactivity, but feel free to use this thread if you have any more questions.

Yes, it should work. However the python AlphaZero based has been deprecated because we no longer support tensorflow. We have a small team of external contributors porting it to JAX....

The old version did, yes.. but like I said, it's been deprecated because it no longer works. See https://github.com/google-deepmind/open_spiel/issues/1326. The files have been moved here: https://github.com/google-deepmind/open_spiel/tree/lanctot-patch-67/open_spiel/python/algorithms/tf We'll see if the...

New effort to support AlphaZero in JAX: https://github.com/google-deepmind/open_spiel/pull/1362

After looking through the AlphaZero code quickly, it wasn't clear .. maybe/possibly? @tewalds, do you remember if you supported this case?

> Does this python version AlphaZero have the speed of c++ version? The Python AlphaZero is not as fast as the C++ version.

> > The MCTS implementation does, but I'm pretty sure you can't in AlphaZero since you add noise to the first layer of nodes, which distorts the shape of the...

Thanks, guys. Just linking the submitted PR for completion: https://github.com/deepmind/open_spiel/pull/1035