Re:Open-sourcing PyTorch-BigGraph for faster embeddings of e...
by mark_l_watson on Tuesday April 02, 2019 @03:29PM
(#19557748)
binarymax in another comment is correct about what PBG does. Still, I am interested in combining deep learning and Neo4J (and possible RDF data stores in the future) and I am experimenting with some code for a book project.
[ Parent
| on HN ]
Re:Open-sourcing PyTorch-BigGraph for faster embeddings of e...
by mark_l_watson on Tuesday April 02, 2019 @04:20PM
(#19558181)
BTW, I am running their code that produced the article results (on small freebase data set) right now and also studying their code.
BTW, part 2: I know that Facebook and Google get a lot of criticism over privacy issues but the flip side of the privacy coin is how generous they are in sharing results, pre-trained models, etc. (even given that publishing great results, data, and code is a great way to attract more potential employees).
Re:Open-sourcing PyTorch-BigGraph for faster embeddings of e...
by rahulkulhari on Wednesday April 03, 2019 @12:51AM
(#19560853)
look at word octavian(https://www.octavian.ai/) is doing
Re:Open-sourcing PyTorch-BigGraph for faster embeddings of e...
by mark_l_watson on Wednesday April 03, 2019 @08:47AM
(#19562856)
Thanks! Interesting stuff.