TRL - Transformer Reinforcement Learning
/All repositories
Overview
HEALTH SCORE
Excellent
94out of 100
Strong maintainer coverage and security posture. Development cadence is the only remaining gap.
LIFECYCLE
Active
Consistent commits, responsive maintainers, regular releases, and healthy issue triage.
Health breakdown
Excellent(94/100)
Maintainer availability, known risks and supply chain posture, and whether the project is actively developed or appropriately stable.
Maintainer responsiveness
Median response time is 2 days.
Bus factor
53 active maintainers currently have merge rights.
Organization Diversity
Contributors span 623 organizations.
Project Achievements
Leaderboards
About
TRL (Transformer Reinforcement Learning) is a library for training transformer language models using reinforcement learning. It provides implementations of algorithms like PPO (Proximal Policy Optimization) adapted for language models, enabling fine-tuning of models based on reward signals and human feedback.
First commit
Software value
$3.8M
License(s)
Apache-2.0
Data sources
Git
GitHub