Datasets:

casperdewith
/

sitata

license: cc0-1.0
language:
  - tok
tags:
  - toki-pona
pretty_name: all sentences of jan Sitata read aloud
size_categories:
  - 1K<n<10K
task_categories:
  - automatic-speech-recognition

This dataset contains an audio recording of each sentence in the book jan Sitata. The author pronoucned each sentence, including questions and exclamations, with the intonation as that of a declarative sentence. This makes the dataset quite homogeneous.

Every tenth fragment is in the test folder; the rest is in the train folder.

Each is a .wav file with a bit rate of 16 kHz. In total, there is 63 minutes of audio in this dataset.