Datasets:

formospeech
/

fsr25_train

Modalities:

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

TRAIN

Subset	lang_group	hours	n_utts	n_chars	secs/utt	chars/sec
Hakka_Dapu	客語_大埔	20.01	12,197	179,935	5.91	2.50
Hakka_Zhaoan	客語_詔安	20.20	15,143	199,508	4.80	2.74
Total	-	40.21	27,340	379,443	5.29	2.62

Downloads last month: 1

Collection including formospeech/fsr25_train

Taiwanese Hakka for ASR (Kaldi Init.)

Speech for ASR (Kaldi Init.) training • 14 items • Updated Dec 19, 2025