OpenPipe/ART

View on GitHub

Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.6, GPT-OSS, Llama, and more!

Stars: 10.8k
Forks: 989
Open beginner issues: 0
Indexed issues: 72
Avg merge: 6h 29m
Merged PRs (30d): 85

Dominant language: Python
License: Apache-2.0
Last GitHub push: Sep 19, 2026
Latest indexed: Sep 20, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

72 open issues indexed

Add @art.rollout decorator to gather trajectories Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

OpenPipe/ART#347 · 6 comments · 1 reaction · Opened Aug 19, 2025
Training freezes at 0% "train" after "gather" stage is complete Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

OpenPipe/ART#346 · 20 comments · Opened Aug 19, 2025
RuntimeError: CUDA error: device-side assert triggered when running 2048.ipynb Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

OpenPipe/ART#344 · 1 comment · Opened Aug 18, 2025
try mcp-14b-alpha-001, it block by below error, how can i know where the trigger it? Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 25/100

OpenPipe/ART#343 · 1 comment · Opened Aug 18, 2025
About asynchronous generation and training like AReal Open

discussion

Difficulty 5/5 Over a week Newbie friendliness 15/100

OpenPipe/ART#337 · 2 comments · Opened Aug 14, 2025
Feature request - load best checkpoint Open

enhancement

Difficulty 3/5 1-2 days Newbie friendliness 45/100

OpenPipe/ART#299 · 2 comments · Opened Aug 5, 2025
closing Local backend - bug Open

bug

Difficulty 3/5 1-2 days Newbie friendliness 35/100

OpenPipe/ART#288 · 2 comments · Opened Aug 1, 2025
Gradient reliability with sample-by-sample vs batch processing Open

question

Difficulty 4/5 3-5 days Newbie friendliness 25/100

OpenPipe/ART#285 · 2 comments · Opened Aug 1, 2025
langgraph create_react_agent Open

question

Difficulty 4/5 3-5 days Newbie friendliness 25/100

OpenPipe/ART#275 · 4 comments · Opened Jul 29, 2025
GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning - Research Discussion Open

discussion

Difficulty 5/5 Over a week Newbie friendliness 20/100

OpenPipe/ART#274 · 1 comment · 3 reactions · Opened Jul 28, 2025
Cannot train the tau-bench datasets on qwen3 model Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 25/100

OpenPipe/ART#270 · 10 comments · Opened Jul 24, 2025
Any video tutorial of this project ? Open

documentation

Difficulty 5/5 Over a week Newbie friendliness 25/100

OpenPipe/ART#269 · 1 comment · Opened Jul 23, 2025
enforce_eager=True not being respected in test scripts Open

bug

OpenPipe/ART#262 · 2 comments · 1 assignee · Opened Jul 21, 2025
VLM Support Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 20/100

OpenPipe/ART#236 · 3 comments · 1 reaction · Opened Jul 14, 2025
Decouple vLLM & Unsloth Trainer Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

OpenPipe/ART#210 · 5 comments · 2 reactions · Opened Jul 8, 2025
RuntimeError: Engine core initialization failed. Failed core proc(s): {} Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 25/100

OpenPipe/ART#206 · 3 comments · Opened Jul 7, 2025
Worker proc VllmWorker-1 died unexpectedly Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 30/100

OpenPipe/ART#201 · 1 comment · Opened Jul 4, 2025
Add Arctic Inference Suffix Decoding Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

OpenPipe/ART#182 · 1 comment · Opened Jun 30, 2025
Multimodal support Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

OpenPipe/ART#100 · 4 comments · 3 reactions · Opened May 2, 2025
Gemma 3 compatibility Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

OpenPipe/ART#90 · 1 comment · 1 reaction · Opened Apr 30, 2025
Optimizer state isn't preserved across runs Open

enhancement

Difficulty 4/5 3-5 days Newbie friendliness 35/100

OpenPipe/ART#72 · 3 comments · Opened Apr 21, 2025
Docs on decision-making process for choosing batch size, learning rate, etc? Open

documentation

Difficulty 4/5 3-5 days Newbie friendliness 30/100

OpenPipe/ART#26 · 1 comment · 2 reactions · Opened Apr 10, 2025

OpenPipe/ART

Get new issues in your inbox