# train model CUDA_VISIBLE_DEVICES=0 \ python -u synthesis.py \ --use_gpu=0 \ --output='./synthesis' \ --config='configs/ljspeech.yaml' \ --checkpoint_transformer='./checkpoint/transformer/step-120000' \ --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 training!" exit 1 fi exit 0