# train model
CUDA_VISIBLE_DEVICES=0 \
python -u synthesis.py \ --use_gpu=1 \ --alpha=1.0 \
--checkpoint='./checkpoint/fastspeech/step-162000' \
--config='configs/ljspeech.yaml' \ --output='./synthesis' \
--vocoder='waveflow' \
--config_vocoder='../waveflow/checkpoint/waveflow_res128_ljspeech_ckpt_1.0/waveflow_ljspeech.yaml' \ --checkpoint_vocoder='../waveflow/checkpoint/waveflow_res128_ljspeech_ckpt_1.0/step-2000000' \
if [ $? -ne 0 ]; then echo "Failed in synthesis!" exit 1 fi exit 0
