Metraipolis › archive › dc_936e7bf86eb3

Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition — arXiv:2608.12327

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.12327
License (as archived)
CC0-1.0
Archived
2026-10-19T04:17:48.093Z
SHA-256 of the archived text
17814d539ae2f9125ccf42dc052fd9f395cb77a0180ae8a9b9aa89afc0ea4728
Stable URL of this page
https://metraipolis.com/documento/dc_936e7bf86eb3

How to cite

Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition — arXiv:2608.12327 — arXiv (en), https://arxiv.org/abs/2608.12327 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_936e7bf86eb3 (Metraipolis archive dc_936e7bf86eb3, sha256 17814d539ae2f9125ccf42dc052fd9f395cb77a0180ae8a9b9aa89afc0ea4728)
BibTeX
@misc{dc936e7bf86eb3,
  title = {Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition — arXiv:2608.12327},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.12327},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_936e7bf86eb3; sha256 17814d539ae2f9125ccf42dc052fd9f395cb77a0180ae8a9b9aa89afc0ea4728},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12327. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multilingual pretrained models nominally support Nepali, yet no controlled benchmark has compared them under a single fine-tuning protocol. We fine-tune six pretrained models (XLSR-53, IndicWav2Vec, MMS-1B, Whisper-Medium, Whisper-Large-v3-Turbo, and Conformer-Hi) spanning CTC self-supervised, autoregressive encoder-decoder, and hybrid Conformer-CTC architectures, on the OpenSLR SLR54 Nepali corpus (~165 hours) using identical preprocessing, splits, optimizer, and family-matched learning-rate schedules. We evaluate Word Error Rate (WER), Character Error Rate (CER), and Real-Time Factor (RTF) on three independent test sets (OpenSLR, FLEURS, Common Voice). Whisper-Large-v3-Turbo (14.76% WER) and IndicWav2Vec (14.89% WER) tie at the top despite a 9x parameter gap and 40x pretraining-data gap, providing direct empirical evidence that language-family proximity in pretraining can substitute for raw scale for in-domain Nepali. CTC decoders run up to 29x faster than autoregressive Whisper at the same accuracy, flipping the practical deployment preference toward CTC under any latency budget. Massively multilingual pretraining (MMS-1B) yields the smallest out-of-domain degradation on FLEURS (+12.55 pp), indicating that scale buys robustness rather than peak in-domain accuracy. The resulting benchmark provides the first standardized, multi-model, efficiency-aware reference numbers for Nepali ASR.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.12327 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-19T04:13:06.636Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).