English(EN) In-Place Indic Tokenizer Expansion and Continual Pretraining of Nemotron-3-Nano-30B-A3B

Nemotron-3-Nano-30B-A3B 通过词汇扩展适配印度语言

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-15 04:00

研究人员通过扩展词汇量并进行持续预训练，将 Nemotron-3-Nano-30B-A3B 模型适配到八种印度语言。该过程涉及持续的字节对编码和 AdaptMix 再平衡，将每词的 token 数量减少了 38.6%。虽然退火检查点在 Indic MILU 基准测试上显示出显著改进，但其在英语 MMLU 上的性能有所下降。 AI

影响这项研究展示了一种将大型语言模型适配到新语言的方法，有可能扩大其全球适用性和可访问性。

排序理由该集群包含一篇详细介绍模型适配和训练方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 16 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍模型适配和训练方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Avinash Benki, Prasanjith Pasunuti, Sutanshu Saksena Raj, Parijat Chatterjee, Nayanjit Sarkar, Daksh Maan · 2026-10-15 04:00

Nemotron-3-Nano-30B-A3B 的就地指示符标记器扩展和持续预训练

arXiv:2610.15373v1 Announce Type: cross Abstract: We adapted Nemotron-3-Nano-30B-A3B for eight Indian languages without changing its backbone architecture. We expanded the vocabulary in place from 131,072 to 163,072 entries using continued byte-pair encoding and AdaptMix rebalanc…

本文如何被排名

报道来源 [1]

Nemotron-3-Nano-30B-A3B 的就地指示符标记器扩展和持续预训练

相关实体

相关话题