See https://github.com/k2-fsa/icefall/pull/1664

icefall-asr-ksponspeech-zipformer-2024-06-24

KsponSpeech is a large-scale spontaneous speech corpus of Korean. This corpus contains 969 hours of open-domain dialog utterances, spoken by about 2,000 native Korean speakers in a clean environment.

All data were constructed by recording the dialogue of two people freely conversing on a variety of topics and manually transcribing the utterances.

The transcription provides a dual transcription consisting of orthography and pronunciation, and disfluency tags for spontaneity of speech, such as filler words, repeated words, and word fragments.

The original audio data has a pcm extension. During preprocessing, it is converted into a file in the flac extension and saved anew.

KsponSpeech is publicly available on an open data hub site of the Korea government. The dataset must be downloaded manually.

For more details, please visit:

zipformer (Zipformer + pruned statelss transducer)

zipformer

Number of model parameters: 74,778,511, i.e., 74.78 M

Training on KsponSpeech (with MUSAN)

The CERs are:

decoding method	eval_clean	eval_other	comment
greedy search	10.60	11.56	--epoch 30 --avg 9
fast beam search	10.59	11.54	--epoch 30 --avg 9
modified beam search	10.35	11.35	--epoch 30 --avg 9