Reputation: 21
I would like to implement a distributed Topic Modeling Pipeline based on Gensim. Unfortunately, Gensim only supports distributed LSA and LDA. I read that Word2Vec and Doc2Vec tend to have better results than LSA and LDA. However, I also read about pretrained Word2Vec and Doc2Vec models. Is there a way to use Word2Vec or Doc2Vec in a distributed manner?
Upvotes: 0
Views: 118
Reputation: 54173
Gensim include no support for distributed Word2Vec
or Doc2Vec
.
But also, many people who think they need distributed versions of these algorithms can actually achieve their goals on a single beefy machine.
Further, you can probably confirm or deny whether Word2Vec
or Doc2Vec
work better on your specific data/goals, rather than trusting some other claims you've read, via some single-machines tests long before concerning yourself with scaling/distribution issues.
Upvotes: 0