总览 / 概念 / ASR Benchmarks
↗ 源 Markdown
概念
asr benchmark speech-model stub
创建2026-05-10
更新2026-05-10
阅读量级1 分钟
概念导读

ASR benchmarks evaluate automatic speech recognition quality across languages, accents, noise conditions, and domain-specific speech.

  1. Related

ASR Benchmarks

ASR benchmarks evaluate automatic speech recognition quality across languages, accents, noise conditions, and domain-specific speech.

They are used to compare speech frontends, LLM-ASR systems, and omni-modal speech models.

  • nim4-asr — Production-oriented LLM-ASR framework.
  • qwen3-asr — Qwen ASR model family.
  • funasr — Open-source ASR toolkit.