1
votes

Using H2O is mandatory, Perform Embedding on the text file(dataset as input) using word2vec model to produce an output as Vectors(arrays). The word2Vec model will be using is below-https://code.google.com/archive/p/word2vec/ which is written in C/C++. So, how can one implement using H2O?

2
By "implement" do you mean you want to load pre-generated word-embeddings into H2O, but then use, say, findSynonyms from inside h2o?Darren Cook
Yes, that will be helpful.roshan_ray

2 Answers

0
votes

[ If using H2O is "mandatory" for you, are you perhaps using it in a course? If so, I'm curious where that is... ]

H2O-3 is written in Java, and has an internal word2vec implementation written from scratch in Java. You wouldn't actually use the referenced C++ implementation inside of H2O-3.

Here is a pointer to the relevant word2vec documentation for H2O-3:

0
votes

You have to generate the previous model as csv, import the csv as a frame and construct the w2oModel and specify the data frame as pretrained parameter.