node-word2vec
node-word2vec copied to clipboard
Why is 1GB of ram required to load a 160mb word vector file?
Are there some pre-computations done which cause this lib to eat up RAM? Because I'm using a 160mb text file of word vectors and the node process is taking up 900mb+ of RAM.
Just wondering whether there is a good reason for this, or whether I should dig about looking for some inefficiencies somewhere.
Thanks