open_spiel

open_spiel copied to clipboard

Any methods to use a pytorch model to generate trajectory in C++ side?

1

Hi there, suppose I have a pytorch model, and I want to use it to generate trajectories for reinforcement learning, for example REINFORCE. The code in python should be like...

zizhang-qiu

On the acceleration of convergence process in neural replicator dynamics

1

Hi,guys, i was wondering whether it is possible to accelerate the convergence process in Neural Replicator Dynamics. In this paper,https://arxiv.org/abs/1906.00190, the accumulated values **y**t at time t is approximated by...

linshuxi

on neural replicator dynamics

2

Hi guys, I am trying to implement the neural replicator dynamics algorithm in the paper "Neural Replicator Dynamics: Multiagent Learning via Hedging Policy Gradients", and the link of the article...

linshuxi

non-marginal strategy selectors in psro_v2

1

Hello. In the current implementation of psro_v2, right before it is going to do best response, it will first select a subset of strategies from the current strategy pool and...

rezunli96

Bug with nox

1

Hello, When I run: `nox -s tests` I encounter the following error: `FileNotFoundError: [Errno 2] No such file or directory: 'build/temp.linux-x86_64-3.11'` I'm unsure whether this is a bug or if...

morLev

Update version of abseil to 20240116.1

2

lanctot

Adding game german_whist_foregame

6

German Whist is a 2 player imperfect information trick-taking card game. The rules are outlined here https://en.wikipedia.org/wiki/German_whist#:~:text=German%20whist%20is%20a%20variant,is%20probably%20of%20British%20origin. Modified from previous pull request, no longer requires x86intrin.h or bmi2 support.

willmcgowan

lanctot

Adding game TwixT

7

TwixT is a 2-player, deterministic, perfect information game, played on a 24x24 grid. See [Wikipedia](https://en.wikipedia.org/wiki/TwixT). This PR adds it to open_spiel. The board size can be adjusted from 5x5 to...

stevens68

Neurd Clip Instability [RNaD]

1

In the paper and open spiel implementation, the neurd clip value is set to 10k. https://github.com/google-deepmind/open_spiel/blob/931e39a99ee73412500def0227925f8f19f033fe/open_spiel/python/algorithms/rnad/rnad.py#L604 From my testing, this is a source of major instability since the vtrace operator...

spktrm