Metraipolis › archive › dc_601aa4438892

Context-Aware Interleaved Batching for WhisperX — arXiv:2608.31170

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.31170
License (as archived)
CC0-1.0
Archived
2026-10-17T08:07:57.697Z
SHA-256 of the archived text
bde6f47e7474f356e8fad358224bd3ef56deb2617c9b8ec341fae80490a148c1
Stable URL of this page
https://metraipolis.com/documento/dc_601aa4438892

How to cite

Context-Aware Interleaved Batching for WhisperX — arXiv:2608.31170 — arXiv (en), https://arxiv.org/abs/2608.31170 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_601aa4438892 (Metraipolis archive dc_601aa4438892, sha256 bde6f47e7474f356e8fad358224bd3ef56deb2617c9b8ec341fae80490a148c1)
BibTeX
@misc{dc601aa4438892,
  title = {Context-Aware Interleaved Batching for WhisperX — arXiv:2608.31170},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.31170},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_601aa4438892; sha256 bde6f47e7474f356e8fad358224bd3ef56deb2617c9b8ec341fae80490a148c1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.31170. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

While WhisperX accelerates speech transcription via intra-audio batching, it isolates audio segments, losing the historical context needed for coherent punctuation and terminology transcription. Conversely, standard Whisper retains context sequentially but suffers from slow inference and hallucination loops. To achieve the best of both worlds, we propose Context-Aware Interleaved Batching. By using VAD-derived segment boundaries, our algorithm stabilizes Whisper's text conditioning, allowing us to safely maintain continuous historical context across batched audio segments. As demonstrated on long-form audio benchmarks, this approach reduces Word Error Rate (WER) and improves proper noun transcription, all while maintaining high-throughput inference speeds.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.31170 — 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 2026-10-17T07:47:52.892Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).