0

I am working with word2vec. I have to load model (GoogleNews vector). But each time run project, It wastes a lot of time to load the model (about 3-4 min). Is there any way to load faster? Below is my code to load model:

print("Loading the model file...")
model = KeyedVectors.load_word2vec_format('~/Downloads/GoogleNews-vectors-negative300.bin.gz', binary=True)
print("Successfully loaded GoogleNews bin file!")
Huy Nguyen
  • 1,032
  • 14
  • 38
  • GoogleNews word2vec is about 1.5GB if I remember correctly. I think 3-4 minute to load such a huge file is okay. – Sreeram TP Sep 30 '17 at 12:33
  • I think you can use the 100 size vector without any appreciable loss in accuracy. – cs95 Sep 30 '17 at 12:35
  • 1
    Possible duplicate of [How to speed up Gensim Word2vec model load time?](https://stackoverflow.com/questions/42986405/how-to-speed-up-gensim-word2vec-model-load-time) – Sreeram TP Sep 30 '17 at 12:35
  • Yeah, But I am searching a solution that can "cache" it after the first loading time – Huy Nguyen Sep 30 '17 at 12:36

0 Answers0