Open tienanh28122000 opened 1 year ago
Our models have RNNLms like English one. You just add RNNLM folder, no?
As for Kaldi recipe, most of them work, something like https://github.com/kaldi-asr/kaldi/blob/master/egs/tedlium/s5_r2/local/rnnlm/tuning/run_lstm_tdnn.sh can be a base probably.
Hi everyone, is there a recipe for RNNLM n-best rescoring in VOSK. If not, how can I implement it, can you refer some Kaldi recipe to do it. Thank you very much!