Extract Word Embeddings from GoogleNews-vectors-negative300.txt

[Lakshya Kumar]

Hello,

I want to extract word embeddings of particular words from the "GoogleNews-vectors-negative300.txt" file. Can anybody help me out by telling how i can do this efficiently?


Thanks,
Lakshya

[Andrey Kutuzov]

Hi Lakshya,

In the Linux shell:
grep WORD GoogleNews-vectors-negative300.txt

> --
> You received this message because you are subscribed to the Google
> Groups "gensim" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gensim+un...@googlegroups.com
> <mailto:gensim+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Solve et coagula!
Andrey

[Lev Konstantinovskiy]

Hi Lakshya,

Or you can load the corresponding 'bin' file into gensim and query for a word-vector as done in this tutorial https://github.com/RaRe-Technologies/gensim/blob/develop/docs/notebooks/word2vec.ipynb

Regards

Lev

On Wednesday, August 17, 2016 at 3:00:12 PM UTC+2, Andrey Kutuzov wrote:

Hi Lakshya,

In the Linux shell:
grep WORD GoogleNews-vectors-negative300.txt

17.08.2016 13:46, Lakshya Kumar wrote:
> Hello,
>
> I want to extract word embeddings of particular words from the
> "GoogleNews-vectors-negative300.txt" file. Can anybody help me out by
> telling how i can do this efficiently?
>
>
> Thanks,
> Lakshya
>
> --
> You received this message because you are subscribed to the Google
> Groups "gensim" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to gensim+unsubscribe@googlegroups.com
> <mailto:gensim+unsubscribe@googlegroups.com>.

[Radim Řehůřek]

Hmm, that tutorial is a little outdated (the blog post has been updated since).

It's not a good idea to deep the content duplicate... we should decide which one to keep, and scrap the other one. I'd probably keep the notebook, and then replace the blog post with a link to the notebook. Thoughts?

-rr