|
|
--- |
|
|
license: cc-by-nc-4.0 |
|
|
task_categories: |
|
|
- automatic-speech-recognition |
|
|
language: |
|
|
- hi |
|
|
- mr |
|
|
--- |
|
|
# ReadNet |
|
|
## Dataset Description |
|
|
ReadNet is an audio dataset collected from more than 2,00,000 children in the age group of 5-16 years in Hindi and Marathi language. |
|
|
The dataset consists of audio files in the wav format where children read out [ASER Samples](https://asercentre.org/wp-content/uploads/2022/12/Hindi_ASER-2018.pdf) which consists of letters, words, stories and paragraphs in their native language. |
|
|
This dataset is a subset of the larger dataset which conssits of 2200 hours of data. This dataset consists of ~87 hours of audio data in Hindi and Marathi, |
|
|
### Languages |
|
|
Hindi and Marathi |
|
|
```python |
|
|
from datasets import load_dataset |
|
|
|
|
|
dataset = load_dataset("PrathamOrgAI/ReadNet") |
|
|
``` |
|
|
|
