Datasets:

formospeech
/

hakka_elearning

dataset_info:
  config_name: hak_sx
  features:
    - name: id
      dtype: string
    - name: stem
      dtype: string
    - name: audio
      dtype: audio
    - name: duration
      dtype: float64
    - name: hanzi
      dtype: string
    - name: mandarin
      dtype: string
    - name: lang_code
      dtype: string
    - name: lang_name
      dtype: string
  splits:
    - name: train
      num_bytes: 1257276978
      num_examples: 15783
  download_size: 1256330348
  dataset_size: 1257276978
configs:
  - config_name: hak_sx
    data_files:
      - split: train
        path: hak_sx/train-*

TRAIN

Subset	lang_name	hours	n_utts	n_chars_in_utts	secs/utt	chars/sec	n_sents	n_chars_in_sents
hak_sx	Hakka_Sixian	10.88	15,783	144,031	2.48	3.68	0	0
Total	-	10.88	15,783	144,031	2.48	3.68	0	0