Hi Philip,
sorry for the late reply. We made some test with Deeplearning4J,
in order to build a Word2Vec model in a Spark environment, three
months ago.
At the begininning library looked quite promising.. However,
because of some memory limitations on our side, that required some
fine tuning on job parallelization, we finally preferred to move
to Spark MLlib implementation, which was also more familiar for
us, and a bit more straightforward to configure. ;)
Cheers,
Michele.
Il 14/07/2016 10:14, Philip Wenig ha scritto:
Hi
folks,
has anyone used these Java libraries before?
=== Deeplearning4j ===
http://deeplearning4j.org
https://en.wikipedia.org/wiki/Deeplearning4j
=== ND4J ===
https://github.com/deeplearning4j/nd4j
https://en.wikipedia.org/wiki/ND4J_(software)
Are they worth to give 'em a try?
Best,
Philip
--
Michele Gabusi
Data Mining and Business Analytics
Big Data Competency Center
michele.gabusi@xxxxxx
Engineering Ingegneria Informatica
Corso Stati Uniti, 23/C - 35127 Padova - Italy
Tel. +39-049.8283549
Fax +39-049.8283569
www.eng.it