YAML Metadata
Error:
"datasets[0]" with value "https://github.com/egorsmkv/speech-recognition-uk" is not valid. If possible, use a dataset id from https://hf.co/datasets.
Online variant of pruned_transducer_stateless5
for Ukrainian: https://github.com/proger/icefall/tree/uk
Decoding demo using Sherpa: https://twitter.com/darkproger/status/1570733844114046976
Trained on pseudolabels generated by darkproger/pruned-transducer-stateless5-ukrainian-1 on the noisy 1200 hours training set. Common Voice data was used only for validation.
./pruned_transducer_stateless5/train.py \
--world-size 2 \
--num-epochs 31 \
--start-epoch 1 \
--full-libri 1 \
--exp-dir pruned_transducer_stateless5/exp-uk-filtered2 \
--max-duration 600 \
--use-fp16 1 \
--num-encoder-layers 18 \
--dim-feedforward 1024 \
--nhead 4 \
--encoder-dim 256 \
--decoder-dim 512 \
--joiner-dim 512 \
--bpe-model uk/data/lang_bpe_250/bpe.model \
--causal-convolution True \
--dynamic-chunk-training True
Unable to determine this model's library. Check the
docs
.
Evaluation results
- Validation WER on Common Voice 6.1 uktest set self-reported17.260