d18n/jp-tts

Japanese Text-to-Speech using espnet

jp-tts

Japanese Text-to-Speech using espnet

Currently working on some tooling for language learning, and wanted the ability to generate relatively realistic audio files for example sentences

This is a flask app that allows you to send a request to /synthesize?q=YOUR_TEXT_HERE and will return a .wav file with the synthesized audio

This runs using pre-trained models from espnet, and I would like to add support for more voices and train my own

To get this up and running locally, you can pull the image from dockerhub:

docker pull d18n/jp-tts

Or, you can build the image locally

docker build . -t d18n/jp-tts

To run the docker image after it's available locally:

docker run -it -p 1337:1337 d18n/jp-tts

Or, if you have a gpu and have nvidia-docker installed/configured on your system, you can run

docker run --gpus all -it -p 1337:1337 d18n/jp-tts

You can test that it's working by making a request to:

http://localhost:1337synthesize?q=これはテストです

P.S. This is the first time I've pushed anything to Dockerhub, and really the first project I've done with docker in general so if you have any suggestions, feel free to make an issue!