---
id: 20260305-T0-06
title: "RxnNano：分层课程训练化学LLM"
title_en: "RxnNano Trains Compact LLMs for Chemistry"
url: https://ai.daily.yangsir.net/daily/20260305-T0-06
issue_date: 2026-03-05
publish_date: 2026-03-04T05:00:00.000Z
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2603.02215
---

# RxnNano：分层课程训练化学LLM

arXiv论文《RxnNano》提出通过分层课程学习训练紧凑型LLM，用于化学反应和逆合成预测。该模型在药物发现任务中表现优于现有参数膨胀方法，参数量减少30%。

## English Version

**RxnNano Trains Compact LLMs for Chemistry**

arXiv paper 'RxnNano' proposes training compact LLMs through curriculum learning for chemical reaction and retrosynthesis prediction. The model outperforms existing parameter-heavy methods in drug discovery tasks while reducing parameters by 30%.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2603.02215)

**详情页**：https://ai.daily.yangsir.net/daily/20260305-T0-06

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*