DQLoRA: A Lightweight Domain-Aware Denoising ASR via Adapter-guided Distillation
Fuente:
arXiv
Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909688866537472 |
|---|---|
| author | Yang, Yiru |
| author_facet | Yang, Yiru |
| contents | We present a demo of DQLoRA, an Adapter-Guided Distillation framework for robust speech recognition under low-resource and noisy conditions. Our method employs a frozen Whisper model as the teacher to provide semantic supervision, and a lightweight Wav2Vec2 student equipped with QLoRA-based Adapters. Training is conducted on the FLEURS dataset augmented with DNS-style noise. The student is optimized by jointly minimizing CTC loss and KL-based distillation loss, enabling efficient adaptation while preserving recognition accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_10313 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DQLoRA: A Lightweight Domain-Aware Denoising ASR via Adapter-guided Distillation Yang, Yiru Sound Audio and Speech Processing We present a demo of DQLoRA, an Adapter-Guided Distillation framework for robust speech recognition under low-resource and noisy conditions. Our method employs a frozen Whisper model as the teacher to provide semantic supervision, and a lightweight Wav2Vec2 student equipped with QLoRA-based Adapters. Training is conducted on the FLEURS dataset augmented with DNS-style noise. The student is optimized by jointly minimizing CTC loss and KL-based distillation loss, enabling efficient adaptation while preserving recognition accuracy. |
| title | DQLoRA: A Lightweight Domain-Aware Denoising ASR via Adapter-guided Distillation |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2507.10313 |