This dataset consists of real-world recordings placed in several locations in Hangzhou, China. Researchers may find it useful for fine-tuning Speech-to-Text models. For privacy reasons, the file is encrypted, and anyone interested in the dataset should contact the uploader for the access key. However, access will only be granted to those who can provide a reasonable explanation of their intended use of the data and agree to respect the privacy of the individuals in the recordings.
