Skip to content

Dataset · BAAI

Mandarin Chinese child speech (3-5 years)

BAAI/ChildMandarin

Open audio dataset of speech from 3-5 year old children in Mandarin Chinese for ASR and speaker verification tasks.

What's inside

The dataset from BAAI contains 10 to 100 thousand audio and text pairs in Mandarin Chinese. Recordings were collected specifically to work with the voices of young children aged 3 to 5 years, for whom public speech data is critically lacking.

What you can do with it

  • Training and adapting ASR systems for children's speech recognition
  • Developing speaker verification models for children's voice interfaces
Tasks
Automatic speech recognition
Languages
Chinese
Size
10K–100K rows
Downloads this month
481
Likes
50
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("BAAI/ChildMandarin")