Deep Learning: Natural Language Processing in Python with Word2Vec: Word2Vec and Word Embeddings in Python and Theano (Deep Learning and Natural Language Processing Book 1) by LazyProgrammer

Author:LazyProgrammer
Language: eng
Format: azw3
Published: 2016-08-19T04:00:00+00:00

So in total that’s D(V + 1) floats we need to update.

What if we used Theano?

The thing with Theano is that we need to take the entire weight matrix as one variable, and the updates are presented to Theano as tuples in the form:

(W, W - lr*g)

So it’s not really possible (yet) to update only parts of the weight.

If we were to do this in Theano, that would be 2xVxD floats to update (in other words, every single parameter), which is even more.

Now you might think you could create 2xV separate D-dimensional vectors, but that’s probably not a good idea since creating objects will incur additional overhead.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.