Arabictagger

Padding tags don't get masked by the transformer, to fix this problem I edited the embedding function to give padding tags zero embeddings.

first non-pre-released version

A new Layer Encoder-Transformer was built to solve tagging problems.

New beta version b2 adds a new module TextMatcher designed to address a common problem in Arabic NLP called orthographic or spelling variation which is a challenge for information retrieval in Arabic. Many approaches exist to address this, and in the ArabicTagger package, I approached the problem with a learnable weighted similarity(cosine similarity). For more information see the GitHub readme