
- 谁
- Hugging Face
- 引文
And most of all, share a fully reusable recipe, so you can reproduce this on your own task, harness, model combinations.
发生了什么
Hugging Face 在 Pi、Hermes Agent 和 OpenCode 中分别为 Qwen3.8-27B 训练了独立的 LoRA 适配器,随后将它们合并为单个适配器。在留出编程基准测试中,每个专用适配器在各自对应的框架内均优于基础模型;对于两个模型均能解决的任务,SWE-bench Pro 上的工具调用次数减少了 53–72%,Terminal-Bench 2.1 上减少了 53–70%。该文章分享了用于复现该方法的训练流程和相关产物。
推荐理由
在目标智能体框架内训练模型不仅提升了编程基准测试成绩,还减少了工具调用次数;公开的方案让团队能够为自己的模型、任务和框架构建适配器。
这件事到目前为止
Hugging Face and Harness Hub have released reinforcement-learning-adapted LoRA adapters for Qwen3.8-27B, along with open weights and a reusable training pipeline, to improve coding agent performance across Pi, Hermes Agent, and OpenCode. Harness Hub first released open weights for Qwen3.8-27B-pi-adapted, a version tuned for the pi coding agent. Clément Delangue reported that this model raised pass@1 rates from 29.4% to 62.7% on SWE-bench Pro (Hard 51) and from 39.5% to 55.8% on Terminal-Bench 2.1, while cutting tool calls per task by 55–62%. Hugging Face later detailed how it trained separate LoRA adapters for Qwen3.8-27B inside Pi, Hermes Agent, and OpenCode before merging them into a single adapter. According to Hugging Face, each specialist outperformed the base model in its own harness on held-out coding benchmarks, using 53–72% fewer tool calls on SWE-bench Pro and 53–70% fewer on Terminal-Bench 2.1 for tasks both models solved, and the company shared its training pipeline and artifacts for reproduction.
这件事的进展 · 2
查看完整事件- 11月20日 22:34
Hugging Face 如何为三种智能体框架训练 Qwen3.8-27B 适配器 本条
- 11月18日 04:13Qwen3.8-27B-pi-adapted 旨在优化 pi 编程智能体Clément Delangue · X
