- Ubuntu or Linux
- Python3.8 or newer
- GPU (CUDA 11.8)
- Pytorch 2.1.0 or newer
python download_librispeech.py
python train_spm.py --librispeech-path ./corpus
python train.py --exp-dir ./exp --global-stats-path ./global_stats.json --librispeech-path ./corpus --sp-model-path ./spm_unigram_1023.model --nodes 1 --gpus 1 --epoch 20
python webapp.py --host 0.0.0.0 --port 8080