openpecha/garchen_rinpoche_inference_transcripts
收藏Hugging Face2025-06-26 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/openpecha/garchen_rinpoche_inference_transcripts
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含552个藏传佛教高僧Garchen Rinpoche的教诲音频段，总时长约1.21小时。数据集用于评估三种不同的藏语自动语音识别模型：通用STT模型、Situ Rinpoche模型和Dilgo Khyentse模型。性能指标包括字符错误率、单词错误率和音节错误率。
The dataset consists of 552 audio segments of Tibetan teachings by the esteemed Buddhist master Garchen Rinpoche, with a total duration of approximately 1.21 hours. It is used to evaluate three different Tibetan automatic speech recognition models: the General STT model, the Situ Rinpoche model, and the Dilgo Khyentse model. Performance metrics include Character Error Rate, Word Error Rate, and Syllable Error Rate.
提供机构：
openpecha
