Neural Machine Translation Deutsch to English citing the point that a word based model is far more efficient than a character based model for translation based applications. Maybe text generation applications can be better for character based models Dataset: ANKI LSTM Encoder-Decoder Schematic: Input Examples: Output Examples:
