Scaling BERT Models for Turkish Automatic Punctuation and Capitalization Correction PubID: pubid: 979-8-3503-7943-3/24/$31.00 ©2024 IEEE
Abdulkader Saoud, Mahmut Alomeyr, Himmet Toprak Kesgin, Mehmet Fatih Amasyali · 2024
Read Scaling BERT Models for Turkish Automatic Punctuation and Capitalization Correction PubID: pubid: 979-8-3503-7943-3/24/$31.00 ©2024 IEEE by Abdulkader Saoud, Mahmut Alomeyr, Himmet Toprak Kesgin, Mehmet Fatih Amasyali (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-02698.txt.
Description
Abdulkader Saoud and co-authors scale in 2024 BERT models for Turkish punctuation and capitalization fixes across five sizes. Clean marks aid reading, and Turkish poses its own text quirks. Each model size trades accuracy against compute cost in a distinct way.
Grades for precision, recall, and balanced score rise with model size in the tests. The base size earns the top correction precision of the set. Tiny forms still serve where compute is scarce, with lower grades as the price. The study gives deployers a size guide by need and budget for Turkish text tools.
Catalog record
| CREATOR | Abdulkader Saoud, Mahmut Alomeyr, Himmet Toprak Kesgin, Mehmet Fatih Amasyali |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-02698