| lang=$1 |
| CVSS_ROOT=/data/zhangshaolei/datasets/cvss/cvss-c |
| COVOST2_ROOT=/data/zhangshaolei/datasets/covost2 |
| ROOT=/data/zhangshaolei/StreamSpeech |
| PRETRAIN_ROOT=$ROOT/pretrain_models |
|
|
| PYTHONPATH=$ROOT/fairseq python $ROOT/preprocess_scripts/prep_cvss_c_multilingual_data.py \ |
| --covost-data-root $COVOST2_ROOT/ --cvss-data-root $CVSS_ROOT/ \ |
| --output-root $CVSS_ROOT/$lang-en \ |
| --src-lang $lang \ |
| --target-type unit --unit-type km1000 --reduce-unit \ |
| --vocoder-checkpoint $PRETRAIN_ROOT/unit-based_HiFi-GAN_vocoder/mHuBERT.layer11.km1000.en/g_00500000 --vocoder-cfg $PRETRAIN_ROOT/unit-based_HiFi-GAN_vocoder/mHuBERT.layer11.km1000.en/config.json |
|
|
|
|
