arXiv cs.CL· Khoa D. Nguyen, Thai Bao Huynh, Binh T. Nguyen·· 11 小时前AI 评分18
PhoPunct：面向越南语标点恢复的 Transformer 架构大规模基准
PhoPunct: A Large-Scale Benchmark of Transformer Architectures for Vietnamese Punctuation Restoration
AI 导读
研究团队构建了包含 2210 万个标注 token（1990 万新闻、220 万小说）的越南语标点恢复基准，评测 4 种预训练 Transformer 骨干搭配可选 Bi-LSTM 与 CRF 层组成的 17 种架构及 PhoBERT-large-Bi-LSTM-CRF 模型 PhoPunct。
正文 · 原文
Abstract:Vietnamese punctuation restoration is often evaluated on relatively small benchmarks, limiting conclusions about how architectural choices scale with data volume. We present a controlled benchmark on News and Novels corpora containing 22.1 million labeled tokens (19.9M News and 2.2M Novels). We evaluate 17 architectures formed from four pretrained Transformer backbones and optional bidirectional long short-term memory (Bi-LSTM) and conditional random field (CRF) layers, together with a PhoBERT-large-Bi-LSTM-CRF model, PhoPunct. Every model is trained for eight epochs with the same learning rate and without architecture-specific tuning. Under this fixed budget, adding Bi-LSTM and CRF layers does not reliably improve macro F1: in seven of eight matched domain-backbone comparisons, a plain Transformer classifier matches or outperforms its Bi-LSTM-CRF counterpart. PhoPunct obtains the best News result (61.35), whereas XLM-R with a CRF obtains the best Novels result (50.18), exceeding PhoPunct (48.97). These results show that the value of recurrent and structured-prediction heads depends on their optimization regime and motivate explicit separation between architectural improvements and improvements due to architecture-specific tuning. Our source code is available at: this https URL
| Comments: | Accepted at PACLIC 40 (2026) |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.09107 [cs.CL] |
| (or arXiv:2611.09107v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.09107 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Khoa D. Nguyen [view email]
[v1]
Fri, 6 Nov 2026 07:39:32 UTC (511 KB)
来源：arXiv cs.CL · arxiv.org