yoom618/librispeech_pc

Librispeech-PC数据集是基于Librispeech ASR数据集的一个变体，主要特点是恢复了标点符号和大小写。该数据集包含多种类型的转录文本，包括原始文本、标准化文本和恢复标点符号及大小写后的文本。数据集的使用方法与Librispeech ASR数据集类似，提供了详细的代码示例来加载和使用数据集。数据集的样本数量统计信息包括训练集、验证集和测试集的样本数量，并指出了部分样本被丢弃的情况。

The Librispeech-PC dataset is a variant of the Librispeech ASR dataset, with the main feature being the restoration of punctuation and capitalization. The dataset includes multiple types of transcriptions, including raw text, normalized text, and text with restored punctuation and capitalization. The usage of the dataset is similar to the Librispeech ASR dataset, with detailed code examples provided for loading and using the dataset. The datasets sample count statistics include the number of samples in the training, validation, and test sets, and it notes that some samples were dropped.