How to run pretrained ESPnet TTS in docker. See also https://espnet.github.io/espnet/docker.html

espnet_tts_docker.sh

git clone https://github.com/espnet/espnet
cd espnet
git checkout v.0.9.7

# Prepare TTS input
echo "HELLO WORLD" > egs/ljspeech/tts1/text

# Run TTS in docker. If you have GPUs, use "espnet/espnet:gpu-cuda10.1-cudnn7-u18"
# https://hub.docker.com/r/espnet/espnet/tags
docker run --rm -v $(pwd)/utils:/espnet/utils -v $(pwd)/espnet:/espnet/espnet -v $(pwd)/egs:/espnet/egs \
  espnet/espnet:cpu-u18 bash -c "cd espnet/egs/ljspeech/tts1; /espnet/utils/synth_wav.sh text"

# Play sound
ffplay egs/ljspeech/tts1/decode/text/wav_wnv/text_1_gen.wav