cfrx: Counterfactual Regret Minimization in Jax.

cfrx is an open-source library designed for efficient implementation of counterfactual regret minimization (CFR) algorithms using JAX. It focuses on computational speed and easy parallelization on hardware accelerators like GPUs and TPUs.

Key Features:

• JIT Compilation for Speed: cfrx makes the most out of JAX's just-in-time (JIT) compilation to minimize runtime overhead and maximize computational speed.

• Hardware Accelerator Support: It supports parallelization on GPUs and TPUs, enabling efficient scaling of computations for large-scale problems.

• Python/JAX Ease of Use: cfrx provides a Pythonic interface built on JAX, offering simplicity and accessibility compared to traditional C++ implementations or prohibitively slow pure-Python code.

Installation

pip install cfrx

Getting started

An example notebook is available here.

Snippet for training a MCCFR-outcome sampling on the Kuhn Poker game.

import jax

from cfrx.envs.kuhn_poker.env import KuhnPoker
from cfrx.policy import TabularPolicy
from cfrx.trainers.mccfr import MCCFRTrainer

env = KuhnPoker()

policy = TabularPolicy(
    n_actions=env.n_actions,
    exploration_factor=0.6,
    info_state_idx_fn=env.info_state_idx,
)

random_key = jax.random.PRNGKey(0)

trainer = MCCFRTrainer(env=env, policy=policy)

training_state, metrics = trainer.train(
    random_key=random_key, n_iterations=100_000, metrics_period=5_000
)

Implemented features and upcoming features

Algorithms
MCCFR (outcome-sampling)	✅
MCCFR (other variants)	❌
Vanilla CFR	✅
Deep CFR	❌

Metrics
Exploitability	✅
Local Best Response	❌

Environments
Kuhn Poker	✅
Leduc Poker	✅
Larger games	❌

Performance

Below is a small benchmark against open_spiel for MCCFR-outcome-sampling on Kuhn Poker and Leduc Poker. Compared to the Python API of open_spiel, cfrx has faster runtime and demonstrates similar convergence.

See also

cfrx is heavily inspired by the amazing google-deepmind/open_spiel library as well as by many projects from the Jax ecosystem and especially sotetsuk/pgx and google-deepmind/mctx.

Contributing

Contributions are welcome, refer to the contributions guidelines.