alibaba/ROLL
An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models
Build with Backblaze B2
SDKs, agent skills, IDE extensions, and reference pipelines from Backblaze Labs. All open source.
Loading star history...
Where it stands
Rank, growth pace and release cadence relative to everything GStars tracks.
Ranks #346 of 3,103 repositories GStars tracks this week (top 11%).
1 release in the last 90 days, the latest 23 days ago, with 60 contributors on GitHub.
Gained 49% fewer stars in the last 90 days than in the 90 days before.
Use Cases & Benefits
- Provides an efficient and user-friendly reinforcement learning scaling library tailored for large language models using large-scale GPU resources.
- Enables flexible multi-role distributed training with advanced RL algorithms and heterogeneous task scheduling for improved LLM performance and scalability.
- Use for training multi-task reinforcement learning models covering reasoning, coding, and instruction following with asynchronous rollout and reward calculation.
- Use for developing agentic RL applications involving multi-turn interactions, games, dialogues, and tool use with asynchronous training support.
- Use for integrating and optimizing large-scale LLM training pipelines leveraging DeepSpeed, Megatron-LM, vLLM, and advanced device mapping for resource management.
Should you depend on it?
License terms, project hygiene and risks, from GitHub metadata and GStars' snapshots.
License
Apache 2.0
Permissive: commercial use, modification and redistribution are allowed with attribution; includes an explicit patent grant from contributors.
Community files
- README: present
Watch-outs
No watch-outs from GStars' checks.
General guidance from repository metadata, not legal advice. How we score
From the README
🚀 An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models 🚀
ROLL is an efficient and user-friendly RL library designed for Large Language Models (LLMs) utilizing Large Scale GPU resources. It significantly enhances LLM performance in key areas such as human preference alignment, complex reasoning, and multi-turn agentic interaction scenarios.
Leveraging a multi-role distributed architecture with Ray for flexible resource allocation and heterogeneous task scheduling, ROLL integrates cutting-edge technologies like Megatron-Core, SGLang and vLLM to accelerate model training and inference.
Daily snapshots
Stars gained per day, and how forks and open issues moved, from GStars' daily snapshots
Forks
322+20
Open issues
140+22
Compared with peers in AI Agents
Fastest-growing repositories GStars tracks in this cohort for the week of Oct 18, by 7-day star growth
| Repository | Stars | +7d | Health | Last release |
|---|---|---|---|---|
| morluto/rea | 373.3K | +88.2K | 60 | 5 days ago |
| DietrichGebert/ponytail | 174.8K | +9.9K | 75 | 7 days ago |
| stablyai/orca | 103.9K | +9.5K | 40 | today |
| affaan-m/ECC | 287.3K | +6.4K | 64 | 22 days ago |
| Leonxlnx/taste-skill | 101.6K | +3.8K | 75 | 3 days ago |
| alibaba/ROLL | 3.4K | +0 | 75 | 23 days ago |
| Category median | 174.8K | +9.5K | 64 | — |
Also by alibaba
Other repositories from this maintainer that GStars tracks
- Go48.1K stars
open-code-review
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
- TypeScript29.5K stars
page-agent
JavaScript in-page GUI agent. Control web interfaces with natural language.
- C++16.1K stars
zvec
A lightweight, lightning-fast, in-process vector database
- TypeScript12.6K stars
formily
📱🚀 🧩 Cross Device & High Performance Normal Form/Dynamic(JSON Schema) Form/Form Builder -- Support React/React Native/Vue 2/Vue 3
- Java11.0K stars
spring-ai-alibaba
Agentic AI Framework for Java Developers
Discover Repositories
Search across tracked repositories by name or description