Sacha Chernyavskiy
Sacha Chernyavskiy
Maybe, it is worth trying to add any simple procedural generation algorithm to test the concept, maybe it would be not that hard. Jax could be paired with recursive algorithms...
Hey! Do you seek to adapt any particular maml implementation or just redo one from the paper?
It means that it has an opportunity to be tested and improved
I think if there are not many reliable implementations, it doesn't sound too hard to test on gymnax-like envs. I am interested in trying MAML for MARL experiments too.
Hey! @FlynnDowey `jax` is a device-agnostic framework (run the same code on different devices: cpu, gpu, etc.), and thus, the implementation works on gpu, unless you have a corresponding version...
Hey! It should work with the games. You need to rewrite a [mcts bot](https://github.com/google-deepmind/open_spiel/blob/master/open_spiel/python/algorithms/mcts.py#L195). I can write in more detail, if you want.
Hey, dear @FlynnDowey. I am back and embarrassingly late. I did find a pipeline how to pair the library with `mctx` but found also some problems @lanctot may (or rather...
@lanctot, is it a third party game? Can look into that.
I have exactly the same device, so it'd be at least curious to try to understand what it's going on
Interesting, for me all the tests are passing. Did you disable the test?