# facebookresearch/ELF

An End-To-End, Lightweight and Flexible Platform for Game Research

Repository: https://github.com/facebookresearch/ELF
Canonical: https://ross.abutalabs.com/products/facebookresearch-elf
Language: C++
License: NOASSERTION
License Family: other
Topics: gaming, cpp, python, artificial-intelligence, deep-learning, neural-network, platform, reinforcement-learning
Archived: true
Last push: 2021-08-30T20:13:52+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3376, "days_push": 1829, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2090, forks 281 (observed 2026-08-28T04:06:12.350441+00:00)

## What it is
ELF is an end-to-end, lightweight and flexible C++/Python platform for game research, focused on real-time strategy games. It hosts multiple game instances concurrently in C++, exposes batched game states to Python, and ships with a fast RTS engine, MiniRTS, Capture the Flag, Tower Defense, and a PyTorch-based reinforcement learning backend.

## Use cases
- train RL agents to play real-time strategy games
- run thousands of concurrent game environments for reinforcement learning
- research self-play and multi-agent RL algorithms
- build custom RTS-like game environments with full access to game internals
- benchmark RL algorithms on miniature RTS games like MiniRTS
- plug existing C/C++ games into an RL training pipeline

## When to choose
- you need high-throughput concurrent game simulation for RL research
- you want a fast, lightweight RTS environment with modifiable game dynamics
- you are studying self-play, MCTS, or batched actor-learner RL algorithms

## When to avoid
- you need a maintained project with recent updates and active support
- you want turn-based or 3D game environments rather than RTS-style games
- you need a simple single-environment Gym-style interface without concurrency

## Facets
- artifact type: framework
- maturity: maintenance
- function: reinforcement-learning, simulation, game-engine, machine-learning, deep-learning
- domain: reinforcement-learning, artificial-intelligence, machine-learning
- platform: cpp, python, cross-platform
- tags: rts-games, game-research, rl-environments, concurrent-simulation, self-play, pytorch, facebook-research, game-development, linux, macos

## Member repositories
- facebookresearch/ELF (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:12.350441+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 2026-08-30T02:55:36.043361+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/ELF (fetched 2026-08-28T04:06:12.350441+00:00, sha cfb7e5fa476b)
- Data as of 2026-08-30T08:39:29.467469+00:00.