Unity-Technologies/ml-agents
The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement learning and imitation learning. observed · 2026-08-28
Health v2 · maintenance only
69/100
- Activity 99
- Release rhythm 13
- Longevity 100
Flags: no_license
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.
- gap_med: 332
- age_days: 3281
- days_rel: 365
- days_push: 8
- n_releases_24m: 2
Adoption not part of the score
19647 stars · 4484 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded
The Unity Machine Learning Agents Toolkit (ML-Agents) lets Unity games and simulations serve as environments for training intelligent agents with deep reinforcement learning, imitation learning, and neuroevolution via a PyTorch-based Python API. It ships with 17+ example environments, multi-agent and self-play support, curriculum learning, and cross-platform inference for trained models.
Use cases
- train NPC behavior for Unity games with reinforcement learning
- train multi-agent cooperative or competitive agents in a 3D simulator
- learn policies from human demonstrations via imitation learning
- run RL research experiments in Unity environments from Python
- wrap Unity environments as gymnasium environments for existing RL libraries
- use curriculum learning and environment randomization to train robust agents
- automated game build testing with trained agents
When to choose
- you develop games or simulations in Unity and want intelligent agents
- you need multi-agent, self-play, or imitation learning in a 3D engine
- you want a free, well-documented RL environment platform with example scenes
- you need cross-platform inference of trained models inside Unity
When to avoid
- your project does not use Unity as the simulation environment
- you need general-purpose RL without a game engine (use Gymnasium or similar)
- you require cutting-edge research algorithms not among PPO, SAC, MA-POCA, BC, or GAIL
- you cannot use Python and PyTorch for training
Facets
library · maturity active
machine-learning reinforcement-learning deep-learning simulation sdk reinforcement-learning machine-learning simulation artificial-intelligence game-engine python cross-platform unity pytorch imitation-learning multi-agent npc-behavior curriculum-learning gym-environments game-development
10 sources
- readme: https://github.com/Unity-Technologies/ml-agents · fetched 2026-08-28 · abd28bbe573d
- homepage: https://unity.com/products/machine-learning-agents · fetched 2026-08-29 · aa029ba48245
- site_page: https://unity.com/features/collaboration · fetched 2026-08-29 · cad621e9760e
- site_page: https://unity.com/features/multiplatform · fetched 2026-08-29 · 2449aa9cdd6b
- site_page: https://unity.com/features/liveops · fetched 2026-08-29 · 08535bc492a2
- site_page: https://unity.com/features/multiplayer · fetched 2026-08-29 · a1d8686f7179
- site_page: https://docs.unity.com · fetched 2026-08-29 · 3ef51e67a898
- site_page: https://unity.com/features/ai · fetched 2026-08-29 · a552b25d5458
- site_page: https://unity.com/roadmap · fetched 2026-08-29 · e647dc2f4252
- site_page: https://unity.com/faq · fetched 2026-08-29 · 3ecf5af1eaae
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| Unity-Technologies/ml-agents | main | 69 |
For agents
markdown · JSON · MCP: product_card(name="Unity-Technologies/ml-agents")
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem