Skip to content

Dataset · laubonghaudoi

Hong Kong Legislative Council speech audio data

laubonghaudoi/legco-speech

A large-scale audio dataset and transcripts of Legislative Council meetings in Cantonese, featuring raw and segmented recordings.

What's inside

Contains recordings of Hong Kong Legislative Council meetings in Cantonese (yue) with a total duration of over 20 thousand hours. Divided into raw and VAD-segmented subsets with text transcriptions.

What you can do with it

  • Train and fine-tune Automatic Speech Recognition (ASR) systems for Cantonese
  • Audio classification and speech processing tasks
Tasks
Automatic speech recognition, Audio → audio, Audio classification, Text generation
Languages
Cantonese
Size
1M–10M rows
Downloads this month
22.6K
Likes
3

How to load it

from datasets import load_dataset

ds = load_dataset("laubonghaudoi/legco-speech")