January 2018
Beginner to intermediate
284 pages
8h 35m
English
The following table lists a few commonly used pre-trained word embeddings:
|
Name |
Year |
URL |
Comments |
|
Word2Vec |
2013 |
A multilingual pre-trained vector is available at https://github.com/Kyubyong/wordvectors. |
|
|
GloVe |
2014 |
Developed by Stanford, it is claimed to be better than Word2Vec. GloVe is essentially a count-based model that combines global matrix decomposition and local context window. |
|
|
FastText |
2016 |
In FastText, the atomic unit is n-gram characters, and a word vector is represented by the aggregation of the n-gram characters. ... |
Read now
Unlock full access