Mirror of Stable-Baselines: a fork of OpenAI Baselines, implementations of reinforcement learning algorithms

PyTorch version of Stable Baselines, reliable implementations of reinforcement learning algorithms.

Simple and concise implementation of DQN on toy environments.

Diagrams for visualizing neural network architecture

🌸Eau De Q-Network [RLC 25] is a pruning algorithm specifically designed for RL which discovers the final sparsity level of the networks🌸

✨iterated Q-Network [TMLR 25] learns several Bellman iterations in parallel instead of learning them sequentially✨

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

⚡ Flashbax: Accelerated Replay Buffers in JAX

Imitation learning benchmark focusing on complex locomotion tasks using MuJoCo.