Trying to run Mozilla TTS on AWS, sythesize problems

I have WaveRNN installed but when I try to sinthesize voice I get the following error:

python -m TTS.synthesize --use_cuda true --vocoder_config_path config.json --vocoder_path checkpoint_670000.pth.tar “Evil is Evil. Lesser, greater, middling… Makes no difference. The degree is arbitary. The definition’s blurred. If I’m to choose between one evil and another… I’d rather not choose at all.” config.json checkpoint_670000.pth.tar ~/TTS
Traceback (most recent call last):
File “/home/ubuntu/anaconda3/envs/python3/lib/python3.6/”, line 193, in _run_module_as_main
main”, mod_spec)
File “/home/ubuntu/anaconda3/envs/python3/lib/python3.6/”, line 85, in _run_code
exec(code, run_globals)
File “/home/ubuntu/anaconda3/envs/python3/lib/python3.6/site-packages/TTS-0.0.3+3366328-py3.6.egg/TTS/”, line 102, in
from WaveRNN.models.wavernn import Model as VocoderModel
ModuleNotFoundError: No module named ‘WaveRNN’

I haven’t tried this since last fall, but I found this colab notebook to be quite helpful - perhaps it will help you as well.

Thank you very much zephyr!