This is a token bi-gram trained on LibriSpeech 960h text. It is used for LODR decoding in icefall
.
Please refer to https://github.com/k2-fsa/icefall/pull/678 for more details.
This is a token bi-gram trained on LibriSpeech 960h text. It is used for LODR decoding in icefall
.
Please refer to https://github.com/k2-fsa/icefall/pull/678 for more details.