This project only houses a Dockerfile which encapsulates all (or, most of) the dependencies to get started with the Cmusphinx Grapheme-to-Phoneme (G2P) conversion library called g2p-seq2seq.
The Dockerfile extends from a Tensorflow image and adds additional dependencies.
It shares an Apache License, like the g2p-seq2seq project.
For a more detailed usage, see the g2p-seq2seq project page.
In summary, you can quickly get started with the following
Build the docker image
docker build -t deontaljaard/g2p .Run it
docker run -it deontaljaard/g2pYou can then play with the g2p-seq2seq lib. Here is a quick example.
docker exec -it deontaljaard/g2p bash
echo "test" >> words.txt
g2p-seq2seq --decode words.txt --model_dir /g2p-seq2seq-model/g2p-seq2seq-model-6.2-cmudict-nostressHappy G2P!