arXiv Computation and Language September 25, 2026 By Phannet Pov, Hyun Woo Park, Voneat Pen, Sovandara Chhoun, Wan-Sup Cho, Saksonita Khoeurn

Closing the Quality Gap in Low-Resource Text-to-Speech: LoRA Fine-Tuning of VoxCPM2 for Khmer and Korean

Read the original on arXiv Computation and Language →

The paper investigates how to close the quality gap in low‑resource text‑to‑speech for Khmer and Korean using the VoxCPM2 model. By training a single low‑rank adaptation (LoRA) adapter on a shared 25.5‑hour corpus, the authors improve Khmer’s mean opinion score from 3.85 to 4.23 with a rank‑64 adapter, while Korean shows no significant gain. The study highlights that adaptation benefits mainly when the base model is weak and that training loss does not always align with human ratings.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

arXiv AI

Aug 14

Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition

arXiv:2608. 12327v1 Announce Type: cross Abstract: Multilingual pretrained models nominally support Nepali, yet no controlled benchmark has compared them under a single fine-tuning protocol.

By Suman Paudel, Sarbin Sayami

fine-tuning multimodal benchmarks

arXiv AI

Jun 9

Subtitle-Aligned Fine-Tuning of Whisper for Swiss German ASR: Benchmark Contamination, Convention Mismatch, and an Honest Baseline at 25.6% WER (13.8% cWER)

arXiv:2606. 07608v1 Announce Type: cross Abstract: We present a systematic study of fine-tuning OpenAI's Whisper large-v3 for Swiss German ASR, using 1,367 hours of broadcast speech paired with Standard German subtitles as weak supervision.

By Felix Akeret

fine-tuning multimodal benchmarks safety data-engineering

arXiv AI

Aug 6

Breaking the Curse ofMultilinguality inMany-to-Many Speech-to-Text Translation via a Resource-AwareMixture of Speech Encoders

arXiv:2608. 04586v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT).

By Yexing Du, Kaiyuan Liu, Youcheng Pan, Bo Yang, Chengpeng Fu, Yu Wang, Ming Liu

llms multimodal benchmarks safety

arXiv Machine Learning

Jul 21

Robust Assamese Speech Recognition through Controlled Fine-Tuning of Whisper Models

arXiv:2607. 17164v1 Announce Type: new Abstract: Developing Automatic Speech Recognition (ASR) for morphologically rich, low-resource languages such as Assamese is challenging due to insufficient annotated speech data.

By Ganapati Das, Dwipen Laskar, Hasin Afzal Ahmed, Sanjib Kr Kalita, Kshirod Sarmah, Hem Chandra Das, Manjula Kalita

fine-tuning multimodal safety

arXiv AI

Aug 7

Breaking the Curse of Multilinguality in Many-to-Many Speech-to-Text Translation via a Resource-Aware Mixture of Speech Encoders

arXiv:2608. 04586v2 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT).

By Yexing Du, Kaiyuan Liu, Youcheng Pan, Bo Yang, Chengpeng Fu, Yu Wang, Ming Liu

llms multimodal benchmarks safety

arXiv Computation and Language

Sep 15

Reducing the Output-Mode Gap in Speech Language Models via Joint-Output On-Policy Distillation

arXiv:2609.15313v1 Announce Type: cross Abstract: Autoregressive generation of interleaved text and acoustic tokens is a common approach to spoken-response generation in speech large language models....

By Daxin Tan, Dehua Tao, Chengxi Deng, Hanlin Zhang, Xiao Chen

llms fine-tuning efficiency multimodal

One email a morning, machine-written

Related stories

Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition

Subtitle-Aligned Fine-Tuning of Whisper for Swiss German ASR: Benchmark Contamination, Convention Mismatch, and an Honest Baseline at 25.6% WER (13.8% cWER)

Breaking the Curse ofMultilinguality inMany-to-Many Speech-to-Text Translation via a Resource-AwareMixture of Speech Encoders

Robust Assamese Speech Recognition through Controlled Fine-Tuning of Whisper Models

Breaking the Curse of Multilinguality in Many-to-Many Speech-to-Text Translation via a Resource-Aware Mixture of Speech Encoders

Reducing the Output-Mode Gap in Speech Language Models via Joint-Output On-Policy Distillation