Interesting links, 19/09/2025

A Visual Guide to Quantization

NetEase-FuXi/EETQ — Easy and Efficient Quantization for Transformers

mmBERT: ModernBERT goes Multilingual

jhu-clsp/mmBERT

Finetuning ByT5 for GED

NeMo T5

spring-media/DeepPhonemizer

honzas83/t5s

PolyIPA – Multilingual Phoneme-to-Grapheme Conversion Model

Mitigating the Exposure Bias in Sentence-Level Grapheme-to-Phoneme (G2P) Transduction

@inproceedings{yoon23d_interspeech,
  title     = {Mitigating the Exposure Bias in Sentence-Level {G}rapheme-to-{P}honeme ({G2P}) Transduction},
  author    = {Eunseop Yoon and Hee Suk Yoon and Dhananjaya Gowda and SooHwan Eom and Daehyeok Kim and John Harvill and Heting Gao and Mark Hasegawa-Johnson and Chanwoo Kim and Chang D. Yoo},
  year      = {2023},
  booktitle = {Interspeech 2023},
  pages     = {2028--2032},
  doi       = {10.21437/Interspeech.2023-2336},
  issn      = {2958-1796},
}

T5G2P: Using Text-to-Text Transfer Transformer for Grapheme-to-Phoneme Conversion

@inproceedings{rezackova21_interspeech,
  title     = {T5G2P: Using Text-to-Text Transfer Transformer for Grapheme-to-Phoneme Conversion},
  author    = {Markéta Řezáčková and Jan Švec and Daniel Tihelka},
  year      = {2021},
  booktitle = {Interspeech 2021},
  pages     = {6--10},
  doi       = {10.21437/Interspeech.2021-546},
  issn      = {2958-1796},
}

T5G2P: Text-to-Text Transfer Transformer Based Grapheme-to-Phoneme Conversion

@ARTICLE{10592637,
  author={Řezáčková, Markéta and Tihelka, Daniel and Matoušek, Jindřich},
  journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing}, 
  title={T5G2P: Text-to-Text Transfer Transformer Based Grapheme-to-Phoneme Conversion}, 
  year={2024},
  volume={32},
  number={},
  pages={3466-3476},
  doi={10.1109/TASLP.2024.3426332}
}