Utilizing Character and Word Embeddings for Text Normalization with Sequence-to-Sequence Models | Read Paper on Bytez