Agent Reinforcement Trainer Files

Train multi-step agents for real-world tasks using GRPO

This is an exact mirror of the Agent Reinforcement Trainer project, hosted at https://github.com/OpenPipe/ART. SourceForge is not affiliated with Agent Reinforcement Trainer.

The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Name	Modified	Size	InfoDownloads / Week
Parent folder
openpipe_art-0.5.13-py3-none-any.whl	2026-03-03	23.3 kB	0
openpipe_art-0.5.13.tar.gz	2026-03-03	8.2 MB	0
README.md	2026-03-03	1.4 kB	0
v0.5.13 source code.tar.gz	2026-03-03	8.3 MB	0
v0.5.13 source code.zip	2026-03-03	8.5 MB	0
Totals: 5 Items		25.1 MB	0

Release Highlights

What's Changed

fix: Use uv build in release workflow (#592)
Release v0.5.12 (#591)
Flex Attention for Local Megatron Backend (#586)
Add fix art issues skill (#585)
Remove batch-level padding from tokenize_sft_batch (#582)
Add mamba and megatron related build packages to no-build-isolation. (#579)
fix: resolve pre-existing CI failures (#578)
feat: Add dedicated mode for UnslothService (#577)
Update transformers to v5.x, unsloth, and add MoE LoRA conversion (#576)
Release v0.5.11 (#572)
Add SFT notebooks (#570)
Add KL-penalized advantage adjustment (#562)
Use uv for Megatron Setup (#560)

Full Changelog: https://github.com/OpenPipe/ART/compare/v0.5.11...v0.5.13

Source: README.md, updated 2026-03-03

Other Useful Business Software

Go From AI Idea to AI App Fast Icon

Go From AI Idea to AI App Fast

One platform to build, fine-tune, and deploy ML models. No MLOps team required.

Access Gemini 3 and 200+ models. Build chatbots, agents, or custom models with built-in monitoring and scaling.

Try Free

Our Free Plans just got better! | Auth0 Icon

Our Free Plans just got better! | Auth0

With up to 25k MAUs and unlimited Okta connections, our Free Plan lets you focus on what you do best—building great apps.

You asked, we delivered! Auth0 is excited to expand our Free and Paid plans to include more options so you can focus on building, deploying, and scaling applications without having to worry about your security. Auth0 now, thank yourself later.

Try free now

Go From AI Idea to AI App Fast

One platform to build, fine-tune, and deploy ML models. No MLOps team required.

Access Gemini 3 and 200+ models. Build chatbots, agents, or custom models with built-in monitoring and scaling.

Try Free

Recommended Projects

Habitat-Lab
A modular high-level library to train embodied AI agents
AWorld
Build, evaluate and train General Multi-Agent Assistance with ease
verl-agent
Designed for training LLM/VLM agents via RL
Acme
A library of reinforcement learning components and agents
OpenClaw-RL
Train any agents simply by 'talking'