# Toni-SM/skrl

Modular Reinforcement Learning (RL) library (implemented in PyTorch, JAX, and NVIDIA Warp) with support for Gymnasium/Gym, NVIDIA Isaac Lab, MuJoCo Playground and other environments

Repository: https://github.com/Toni-SM/skrl
Canonical: https://ross.abutalabs.com/products/skrl
Homepage: https://skrl.readthedocs.io/
Language: Python
License: MIT
License Family: permissive
Topics: reinforcement-learning, python, deep-learning, gym, gymnasium, machine-learning, robotics, jax, isaaclab, brax, flax, isaacsim, multi-agent, torch, warp
Last push: 2026-05-11T09:35:05+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 81, release rhythm 71, longevity 100
- inputs: {"age_days": 1780, "days_push": 114, "days_rel": 114, "gap_med": 40.5, "n_releases_24m": 7}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1089, forks 155 (observed 2026-08-28T04:03:32.763895+00:00)

## What it is
skrl is an open-source modular Reinforcement Learning library written in Python, implemented in PyTorch, JAX, and NVIDIA Warp. It supports Gymnasium/Gym, PettingZoo, NVIDIA Isaac Lab, MuJoCo Playground, and other environments, including multi-agent and simultaneous multi-environment training.

## Use cases
- train reinforcement learning agents in PyTorch or JAX
- train RL agents on NVIDIA Isaac Lab or MuJoCo Playground environments
- run multi-agent reinforcement learning experiments
- train agents across multiple environments simultaneously in one run
- learn and prototype RL algorithms with readable modular code
- train RL policies for robotics simulation

## When to choose
- you need a modular, readable RL library with multi-backend support (PyTorch, JAX, Warp)
- you work with Isaac Lab, MuJoCo, Gymnasium, or PettingZoo environments
- you need multi-agent RL or scoped multi-environment training
- you want an MIT-licensed library with active development and docs

## When to avoid
- you need a highly optimized single-algorithm implementation with maximum throughput
- you require backends outside PyTorch/JAX/Warp
- you need non-RL deep learning tooling like supervised training pipelines

## Facets
- artifact type: library
- maturity: active
- function: reinforcement-learning, machine-learning, deep-learning, simulation
- domain: reinforcement-learning, machine-learning, robotics, deep-learning
- platform: python, cross-platform
- tags: pytorch, jax, nvidia-warp, gymnasium, isaac-lab, mujoco, multi-agent, robotics, gpu

## Member repositories
- Toni-SM/skrl (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:32.763895+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:49:16.739598+00:00, confidence not recorded.
  - readme: https://github.com/Toni-SM/skrl (fetched 2026-08-28T04:03:32.763895+00:00, sha 74290eab773c)
  - registry_pypi: https://pypi.org/pypi/skrl/json (fetched 2026-08-29T12:52:17.741608+00:00, sha 3e08d768818c)
- Data as of 2026-08-30T08:39:29.467469+00:00.