Deeplearning4j

from Wikipedia, the free encyclopedia
Deeplearning4j
Basic data

Maintainer various
developer Adam Gibson
Current  version 0.9.1
( August 12, 2017 )
operating system Linux , macOS , Windows
programming language C , C ++ , Java
category Machine learning
License Apache 2.0
deeplearning4j.org

Deeplearning4j is a free , cross-platform program library for artificial intelligence and machine learning . Deeplearning4j is written in the Java programming language and is therefore compatible with any other language that can run on a JVM , such as Scala or Clojure . The underlying calculations are implemented in C and C ++ . Commercial support is provided by Skymind.

With the help of Keras , Deeplearning4j has a programming interface (API) for the Python language . This bridges the gap between widespread, but Python-based program libraries and Java. With Keras an import of the models of artificial neural networks of the systems TensorFlow , Caffe, Torch and Theano is possible. Deeplearning4j is one of the top 10 deep learning projects on Github.

Different types of artificial neural networks such as e.g. B. Restricted Boltzmann machine , convolutional NNs or feedback neural networks are supported. GPUs can take over part of the processing load to increase speed.

Individual evidence

  1. Release 0.9.1 . August 12, 2017 (accessed July 20, 2018).
  2. ^ Cade Metz: The Mission to Bring Google's AI to the Rest of the World . In: WIRED . February 6, 2014 (English, wired.com [accessed March 3, 2017]).
  3. ^ Deep Learning for (Some of) the People . In: Bloomberg.com . June 3, 2014 (English, bloomberg.com [accessed March 3, 2017]).
  4. ^ Keras Documentation. In: Keras: Deep Learning library for Theano and TensorFlow. Retrieved March 6, 2017 (English).
  5. ^ Importing Models From Keras to Deeplearning4j. Retrieved March 3, 2017 (English).
  6. Top 10 Deep Learning Projects on Github. KDnuggets News, January 2016, accessed March 3, 2017 (American English).

Web links