inesc-id/longlibriheavy
收藏Hugging Face2025-07-30 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/inesc-id/longlibriheavy
下载链接
链接失效反馈官方服务：
资源简介：
LongLibriHeavy是一个新颖的长形式自动语音识别（ASR）基准数据集，旨在评估最先进模型在长形式自动语音识别方面的能力。该数据集由LibriHeavy语料库派生而来，特别挑选用于挑战ASR系统处理扩展语音记录的能力。当前版本包含开发集（dev）和测试集（llh_test_16, llh_test_30, llh_test_60, llh_test_180, 和 llh_test）。
LongLibriHeavy is a novel benchmark dataset designed to evaluate the capabilities of state-of-the-art models in long-form automatic speech recognition (ASR). Derived from the LibriHeavy corpus, this dataset is specifically curated to challenge ASR systems with extended speech recordings. The current release includes only the dev and test (llh_test_16, llh_test_30, llh_test_60, llh_test_180, and llh_test) splits.
提供机构：
inesc-id
