Errors in data labelling

Hello! Thank you for the public dataset, but I found some errors in the data annotation, for example in zh2en-10-art.json, the text and the corresponding audio cut out correspond incorrectly.