# train model
CUDA_VISIBLE_DEVICES=0 \
python -u synthesis.py \ --max_len=50 \ --transformer_step=160000 \
--vocoder_step=70000 \
--use_gpu=1 --checkpoint_path='./checkpoint' \ --log_dir='./log' \ --sample_path='./sample' \ --config_path='config/synthesis.yaml' \ if [ $? -ne 0 ]; then echo "Failed in training!" exit 1 fi
exit 0