Visualize the word embeddings using t-SNE

Let's visualize the word embeddings that we generated in the previous section. The t-SNE is the most popular method to display high-dimensional data in two-dimensional spaces. We shall use the method from the scikit-learn library and reuse the code given in TensorFlow documentation to draw a graph of the word embeddings we just learned.

Note

The original code from the TensorFlow documentation is available at the following link: https://github.com/tensorflow/tensorflow/blob/r1.3/tensorflow/examples/tutorials/word2vec/word2vec_basic.py.

Here is how we implement the procedure:

1. Create the tsne model:

tsne = TSNE(perplexity=30, n_components=2,
            init='pca', n_iter=5000, method='exact')

2. Limit the number of embeddings to display to 500, otherwise, the graph becomes very unreadable:

n_embeddings = 500

3. Create the low-dimensional representation by calling the fit_transform() method on the tsne model and passing the first n_embeddings of final_embeddings...