Update README.md

Author: r0f1

README.md

@@ -185,6 +185,7 @@ Why the default feature importance for random forests is wrong: [link](http://ex
 [Text classification Intro](https://mlwhiz.com/blog/2018/12/17/text_classification/), [Preprocessing blog post](https://mlwhiz.com/blog/2019/01/17/deeplearning_nlp_preprocess/).  
 [gensim](https://radimrehurek.com/gensim/) - NLP, doc2vec, word2vec, text processing, topic modelling (LSA, LDA), [Example](https://markroxor.github.io/gensim/static/notebooks/gensim_news_classification.html), [Coherence Model](https://radimrehurek.com/gensim/models/coherencemodel.html) for evaluation.  
 Embeddings - [GloVe](https://nlp.stanford.edu/projects/glove/) ([[1](https://www.kaggle.com/jhoward/improved-lstm-baseline-glove-dropout)], [[2](https://www.kaggle.com/sbongo/do-pretrained-embeddings-give-you-the-extra-edge)]), [StarSpace](https://github.com/facebookresearch/StarSpace), [wikipedia2vec](https://wikipedia2vec.github.io/wikipedia2vec/pretrained/).  
+[magnitude](https://github.com/plasticityai/magnitude) - Vector embedding utility package.  
 [pyldavis](https://github.com/bmabey/pyLDAvis) - Visualization for topic modelling.  
 [spaCy](https://spacy.io/) - NLP.  
 [NTLK](https://www.nltk.org/) - NLP, helpful `KMeansClusterer` with `cosine_distance`.