Hugging Face Blog·· 22 d ago
Async GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL
Async GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL
AI summary
TRL v1.14's AsyncGRPOTrainer now supports training and synchronising only LoRA adapters to vLLM. A rank-1 adapter is just a few MB and can transfer through a Storage Bucket mounted to each job, without NCCL.
Selection record
Threshold 60Official, first-handFirst 47Second 34
Not admittedSum of both 81 < twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed：AsyncGRPO训练LoRA适配器与vLLM同步的AI技术实践
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Hugging Face Blog · huggingface.co