# Simple-Efficient/RL-Factory

Train your Agent model via our easy and efficient framework

Repository: https://github.com/Simple-Efficient/RL-Factory
Canonical: https://ross.abutalabs.com/products/rl-factory
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-12-05T03:25:56+00:00

## Health v2 (maintenance only)
Score: 34/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 55, release rhythm 8, longevity 33
- inputs: {"age_days": 467, "days_push": 271, "days_rel": 467, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1780, forks 165 (observed 2026-08-28T04:05:35.391203+00:00)

## What it is
RLFactory is a reinforcement learning post-training framework for training agentic LLM models, decoupling the environment from RL training via tool configs and reward functions. It supports multi-turn tool-calling, async execution for 2x faster training, model-judge rewards, and one-click DeepSearch training with models like Qwen3.

## Use cases
- train an agent model with reinforcement learning
- rl post-training for tool-calling llm
- train a deepsearch agent with grpo
- build an agent that calls mcp tools
- fine-tune qwen3 for multi-turn tool use
- speed up agentic rl training with async tool calls

## When to choose
- you want to RL post-train an LLM agent with custom tools and reward functions
- you need fast, async multi-turn tool-calling training
- you want one-click DeepSearch agent training on Qwen3

## When to avoid
- you only need supervised fine-tuning without RL
- you need a general-purpose RL library unrelated to LLM agents
- you require multimodal or Android agent training before those features ship

## Facets
- artifact type: framework
- maturity: active
- function: llm-training, agent-framework, machine-learning, mcp
- domain: large-language-models, reinforcement-learning, machine-learning
- platform: python
- tags: reinforcement-learning, post-training, grpo, tool-calling, qwen3, deepsearch, agentic-learning, ai-agents, gpu, linux

## Member repositories
- Simple-Efficient/RL-Factory (main) score 34

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:35.391203+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 2026-08-30T03:24:26.767341+00:00, confidence not recorded.
  - readme: https://github.com/Simple-Efficient/RL-Factory (fetched 2026-08-28T04:05:35.391203+00:00, sha d107141540b4)
- Data as of 2026-08-30T08:39:29.467469+00:00.