In this notebook, we explore the models proposed by Mikolov et al. in [1]. We build the Skipgram and CBOW models from scratch, train them on a relatively small corpus, implement an analogy function using the cosine similarity, and provide some examples that make use of the trained models and analogy function to perform the word analogy task. We look at three different number of dimensions for the word embeddings in order to get a better intuition how the number of dimensions influences the result.
-
Notifications
You must be signed in to change notification settings - Fork 0
nickvdw/word2vec-from-scratch
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
A word2vec implementation (for CBOW and Skipgram) demonstrated on the word analogy task