---
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](https://huggingface.co/datasets/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](https://huggingface.co/datasets/clt013/malay-speech-3k-rows-split/blob/main/malay_speech_3k_preparation.ipynb) !!

Mirror for [Malay Conversational Speech Corpus](https://magichub.com/datasets/malay-conversational-speech-corpus/)