ART
ART copied to clipboard
OpenPipe
Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen2.5, Qwen3, Llama, and more!
Add a 10-second timeout to results_queue.join() to prevent indefinite hangs when lingering results aren't properly consumed. If a timeout occurs, drain any remaining items from the queue to allow training...
[train.py] _calculate_logprobs - Processing chunk 2168 to 2176 [train.py] _calculate_logprobs - Processing chunk 2176 to 2184 [train.py] _calculate_logprobs - Processing chunk 2184 to 2192 [train.py] _calculate_logprobs - Processing chunk 2192...
Hi folks! Would you be interested in adding experiment tracking support using `trackio`, a lightweight, free experiment tracking library from Hugging Face? Trackio docs: https://huggingface.co/docs/trackio/en/index It uses the same syntax...
I'm training the qwen3 model using the tau-bench datasets follow to the example source code in dev folder. Unfortunatedly, i've got the error as belowed: ``` import os from dotenv...
Why does a model trained via MCP show a significant discrepancy in results when tested using generate_benchmarks.py, compared to the outcomes from train.py? A preliminary investigation indicates that the root...
Description ## Summary When training a LangGraph agent with `openpipe-art[backend,langgraph]`, the process fails at model initialization with the following error: RuntimeError: torch.cuda.MemPool doesn't currently support expandable_segments. The error occurs inside...
**Description** I am experiencing an AttributeError when attempting to run a script that uses Unsloth with vLLM. The traceback indicates that an LLM object from vLLM is being accessed for...
I imagine this boils down to some change in a dependency - but raising here in case others encounter. I have not been able to identify a fix on my...
