kengz/SLM-Lab

Modular Deep Reinforcement Learning framework in PyTorch. Companion library of the book "Foundations of Deep Reinforcement Learning".

What it solves

SLM Lab provides a modular framework for Deep Reinforcement Learning (DRL) research and application. It eliminates the need to write repetitive boilerplate code for every experiment by offering a standardized way to configure, run, and analyze RL agents across various environments.

How it works

The framework is built on PyTorch and integrates with Gymnasium (the maintained fork of OpenAI Gym). Users define their experiments using JSON specification files, which allow them to swap algorithms, hyperparameters, and environments without changing the underlying code. It supports both local training and remote cloud GPU training via dstack, with results automatically synced to HuggingFace.

Who it’s for

It is designed for RL researchers, students (specifically those reading the companion book Foundations of Deep Reinforcement Learning), and developers looking to implement and validate RL algorithms on standard benchmarks.

Highlights

  • Broad Algorithm Support: Includes validated implementations of PPO, SAC, DQN, A2C, REINFORCE, and CrossQ.
  • Configuration-Driven: Experiments are fully defined via JSON specs for high reproducibility.
  • Environment Flexibility: Compatible with any Gymnasium environment, including Classic Control, Box2D, MuJoCo, and Atari.
  • Cloud Integration: Built-in support for dstack for GPU training and HuggingFace for result sharing.
  • Automatic Analysis: Integrated TensorBoard logging and training curve generation.

Related

  • Project
    rlcode/reinforcement-learning

    A collection of easy-to-read, standalone PyTorch implementations of core reinforcement learning algorithms, from basic Grid World to deep RL for Atari games.

  • Project
    open-mmlab/mmengine

    A foundational PyTorch library that serves as a training engine for deep learning models, providing standardized components for training loops, configuration, and monitoring.

  • Project
    neuralhydrology/neuralhydrology

    A PyTorch-based Python library designed for deep learning research in hydrology, allowing users to train neural networks via configuration files.

  • Project
    galilai-group/stable-pretraining

    stable-pretraining is a PyTorch Lightning-based framework for foundation-model research that simplifies self-supervised, supervised, and multi-modal pretraining (SimCLR, DINO, MAE, CLIP, etc.) via a dict-based design, live evaluation callbacks, GPU-side augmentation, and SLURM-grade orchestration, with an experimental JAX backend.

  • Project
    pytorch/pytorch

    A Python-based machine learning framework that provides GPU-accelerated tensor computation and a dynamic autograd system for building deep neural networks.