Deep Learning with Keras

Using third-party implementations of word2vec

We have covered word2vec extensively over the past few sections. At this point, you understand how the skip-gram and CBOW models work and how to build your own implementation of these models using Keras. However, third-party implementations of word2vec are readily available, and unless your use case is very complex or different, it makes sense to just use one such implementation instead of rolling your own.

The gensim library provides an implementation of word2vec. Even though this is a book about Keras and not gensim, we include a discussion on this because Keras does not provide any support for word2vec, and integrating the gensim implementation into Keras code is very common practice.

Installation ...