README.md

metadata

dataset_info:
  features:
    - name: sentence
      dtype: string
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
  splits:
    - name: train
      num_bytes: 216779209.875
      num_examples: 2545
    - name: test
      num_bytes: 60513801
      num_examples: 604
  download_size: 275550950
  dataset_size: 277293010.875
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
language:
  - ms
license: cc-by-nc-nd-4.0
size_categories:
  - 1K<n<10K

Malay Conversational Speech Corpus

This dataset is contain 2.4 hours, processing train test split with test_size=0.2 from hugging face dataset (malaysia-ai/malay-conversational-speech-corpus) for further research purpose.

Everyone are welcome to download the dataset and see the conversion from iterable dataset to dataset in notebook !!

Mirror for Malay Conversational Speech Corpus