# train model CUDA_VISIBLE_DEVICES=0 \ python -u synthesis.py \ --use_gpu=1 \ --alpha=1.0 \ --checkpoint='./fastspeech_ljspeech_ckpt_1.0/fastspeech/step-162000' \ --config='fastspeech_ljspeech_ckpt_1.0/ljspeech.yaml' \ --output='./synthesis' \ --vocoder='waveflow' \ --config_vocoder='./waveflow_res128_ljspeech_ckpt_1.0/waveflow_ljspeech.yaml' \ --checkpoint_vocoder='./waveflow_res128_ljspeech_ckpt_1.0/step-2000000' \ if [ $? -ne 0 ]; then echo "Failed in synthesis!" exit 1 fi exit 0