Sign in to view source links and access this dataset
Description
asmr-zh-r18 is a collection of Chinese R18 ASMR audio works for fine-tuning text-to-speech and voice cloning models. The dataset contains 5,876 works, sourced from the asmr.one API, with a total raw size of approximately 1.1 terabytes. It was uploaded by jasonfan and last updated on March 20, 2026.
Use Cases
Fine-tuning text-to-speech models based on Chinese ASMR audio samples.
Training voice cloning models based on the provided MP3, WAV, and FLAC audio tracks.
Researching audio feature extraction and speech patterns in Chinese-language ASMR content.
Strengths
Contains 5,876 distinct audio works, providing a substantial number of samples.
Total raw data size is approximately 1.1 terabytes, indicating significant audio content.
Audio is provided in multiple formats (MP3, WAV, FLAC) across 3 tracks per work.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Data may reflect source bias inherent to the asmr.one API and the R18 category.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
asmr.one API, Chinese R18 category
Collection Method
Data was gathered via the asmr.one API.
Freshness
Last updated 2026-03-20 23:57:06; freshness should be verified.
Geography
China (implied by Chinese-language content)
License is unknown; users must verify usage rights before downloading and applying the data.