Sequence to Sequence modeling have seen great performance in building models where the input is a sequence of tokens (words for example) and output is also a sequence of tokens. The notebook provides an end-to-end training example of training the English word pronunciation model. - View it on GitHub