Welcome to Deeplearning4j
Deeplearning4j aims to be cutting-edge plug and play, more convention than configuration, which allows for fast prototyping for non-researchers. DL4J is customizable at scale. Released under the Apache 2.0 license, all derivatives of DL4J belong to their authors. DL4J can import neural net models from most major frameworks via Keras, including TensorFlow, Caffe, Torch and Theano, bridging the gap between the Python ecosystem and the JVM with a cross-team toolkit for data scientists, data engineers and DevOps. Keras is employed as Deeplearning4j's Python API.
Deep neural nets are capable of record-breaking accuracy. For a quick neural net introduction, please visit our overview page. In a nutshell, Deeplearning4j lets you compose deep neural nets from various shallow nets, each of which form a so-called `layer`. This flexibility lets you combine restricted Boltzmann machines, other autoencoders, convolutional nets or recurrent nets as needed in a distributed, production-grade framework that works with Spark and Hadoop on top of distributed CPUs or GPUs.
There are a lot of parameters to adjust when you're training a deep-learning network. We've done our best to explain them, so that Deeplearning4j can serve as a DIY tool for Java, Scala and Clojure programmers.
- Distributed CPUs and GPUs
- Java, Scala and Python APIs
- Adapted for micro-service architecture
- Parallel training via iterative reduce
- Scalable on Hadoop
- GPU support for scaling on AWS
- Deeplearning4J: Neural Net Platform
- ND4J: Numpy for the JVM
- DataVec: Tool for Machine Learning ETL Operations
- JavaCPP: The Bridge Between Java and Native C++
- Arbiter: Evaluation Tool for Machine Learning Algorithms
- RL4J: Deep Reinforcement Learning for the JVM
If you have any questions, please join us on Gitter; for premium support, contact us at Skymind. ND4J is the Java-based scientific computing engine powering our matrix operations. On large matrices, our benchmarks show it runs roughly twice as fast as Numpy.
Use cases?DL4J is applied for Fraud detection, network intrusion detection, Recommender Systems (CRM, adtech, churn prevention), Regression and predictive analytics, Face/image recognition, Voice search, Speech-to-text (transcription), and preventative hardware monitoring (anomaly detection).
Why Deeplearning4j?With a versatile n-dimensional array class for Java and Scala, DL4J is Scalable on Hadoop, utlizes GPU support for scaling on AWS, includes a general vectorization tool for machine-learning libs, and most of all relies on ND4J: A matrix library much faster than Numpy and largely written in C++. We also built RL4J: Reinforcement Learning for Java with Deep Q learning and A3C.
How can I contribute?Developers who would like to contribute to Deeplearning4j can get started by reading our Developer's Guide.
Is DL4J parallelized?Deeplearning4j includes both a distributed, multi-threaded deep-learning framework and a normal single-threaded deep-learning framework. Training takes place in the cluster, which means it can process massive amounts of data quickly. Nets are trained in parallel via iterative reduce, and they are equally compatible with Java, Scala and Clojure. Deeplearning4j's role as a modular component in an open stack makes it the first deep-learning framework adapted for a micro-service architecture.
Can I program in Python with Deeplearning4j?Yes! Deeplearning4j's Python API employs Keras, a high-level, intuitive abstraction layer that also takes TensorFlow and Theano as a backend. Teams that have trained models on other Python frameworks can import them to the JVM and Deeplearning4j using Keras model import. And developers just starting to train configure and train a model can do so via our Python interface.
Deep Learning TutorialsIntroduction to Neural Networks Word2vec: Extracting Relations From Raw Text Eigenvectors, PCA and Entropy Regression, Predictive Analytics and Neural Nets Long Short-Term Memory Networks and RNNs DL4J Tutorials Index How to Choose a Neural Net
DL4J's Neural NetworksRestricted Boltzmann machines Convolutional Nets (images) Recurrent Nets LSTMs (time series and sensor data) Deep-belief networks Deep Autoencoders (Question-Answer/data compression) Stacked Denoising Autoencoders
Recent ResearchUniversity of Massachusets "RandomOut: Using a convolutional gradient norm to win The Filter Lottery" Stanford NLP: "Large-Scale Language Classification" Like2Vec: A New Technique for Recommender Algorithms