#201201st of 242 in ML frameworks and libraries
acme
Reinforcement learning library with reference agents in JAX and TensorFlow
- Stars
- 4.1k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Feb 2022
- Language
- Python
3,948 on 25 Apr 20264,087 on 14 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Acme is a DeepMind Python library of reinforcement learning building blocks and readable reference agents, meant as baselines and a starting point for new research. Agents run on JAX or TensorFlow depending on the agent, and the components are designed to run either single-stream or distributed. Example environments from gym, dm_control and bsuite install through an extra.
Who it is for: RL researchers needing reference agents and baselines
Strengths
- Reference agents written to be readable and usable as research baselines
- Same building blocks run single-process or distributed
- Quickstart and tutorial Colab notebooks plus an examples directory
- Pip extras for JAX, TensorFlow and example environments (gym, dm_control, bsuite)
Weaknesses
- README warns things may break occasionally; it is a research codebase
- Last tagged release is from 2022-02-10, so PyPI lags the repository
- Core install is minimal; JAX or TensorFlow must be installed separately
- Not aimed at LLM or agent apps; focused on classic RL
What it needs
- Needs JAX, TensorFlow, gym, dm_control, bsuite
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in ML frameworks and libraries
See all 242| Rank | Project | Score |
|---|---|---|
| 1 | transformersModel-definition library for pretrained text, vision, audio and multimodal models | 94 out of 100 |
| 2 | ultralyticsPython library and CLI for training and running YOLO vision models | 94 out of 100 |
| 3 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 94 out of 100 |
| #4 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |