An algorithm that generalizes the paradigm of self-play reinforcement learning and search to imperfect-information games.

⚡ Flashbax: Accelerated Replay Buffers in JAX

Play, learn, solve, and analyze No-Limit Texas Hold Em. Implementation follows from Monte Carlo counter-factual regret minimization over with hierarchical K-means imperfect recall abstractions.

JAxtar is a project with a JAX-native implementation of parallelizeable A* & Q* solver for neural heuristic search research.

LLM training in simple, raw C/CUDA

Repository with QPSICE models dedicated to Power Electronics

Fast and modular sklearn replacement for generalized linear models

🕹️ A suite of diverse and challenging RL environments in JAX

Monte Carlo tree search in JAX

OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games.

🎲 Vectorized RL game environments in JAX

A Jax-based library for designing and training small transformers.

cfrx is a collection of algorithms and tools for hardware-accelerated Counterfactual Regret Minimization (CFR) algorithms in Jax.

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

Bringing Old Photo Back to Life (CVPR 2020 oral)

ALIEN is a CUDA-powered artificial life simulation program.

♟️ Vectorized RL game environments in JAX

🧬 Nucleotide Transformer: Building and Evaluating Robust Foundation Models for Human Genomics

A functional training loops library for JAX

Neuroevolution is a Competitive Alternative to Reinforcement Learning for Skill Discovery

Vectorization techniques for fast population-based training.

🕹️ A diverse suite of scalable reinforcement learning environments in JAX

Book Paris tennis court (include CAPTCHA bypass)

Accelerated Quality-Diversity

Accelerated Quality-Diversity

poker ai

Code for "Explore, Discover and Learn: Unsupervised Discovery of State-Covering Skills"

pytorch implement for the paper Few-Shot Adversarial Domain Adaptation