SKIL Documentation

Skymind Intelligence Layer

The community edition of the Skymind Intelligence Layer (SKIL) is free. It takes data science projects from prototype to production quickly and easily. SKIL bridges the gap between the Python ecosystem and the JVM with a cross-team platform for Data Scientists, Data Engineers, and DevOps/IT. It is an automation tool for machine-learning workflows that enables easy training on Spark-GPU clusters, experiment tracking, one-click deployment of trained models, model performance monitoring and more.

Get Started

Version Information

The versions packaged with SKIL are based off a particular major release. They may contain additional forward-commits and/or patches that may not be present in the base release, or they may be lacking certain commits that are not yet stable enough to be released as part of SKIL.

SKIL v1.1 is distributed with DL4J components from v1.0.0-beta

DL4J Stack

Component
Version

deeplearning4j

1.0.0-beta

libnd4j

1.0.0-beta

nd4j

1.0.0-beta

datavec

1.0.0-beta

CUDA and MKL

Component
Version

CUDA

9.1

MKL

2018.3

Supported and Bundled Anaconda Libraries **

Component
Version

Python

2.7.15 :: Anaconda, Inc.

Jupyter

1.0.0

Keras

2.2.0

Pytorch CPU

0.4.0

TensorFlow CPU

1.8.0

Note

tensorflow_gpu is NOT installed by default. It can be installed with a

%sh 
/opt/skil/miniconda/bin/conda install tensorflow_gpu

paragraph in a notebook.

This is required for tensorflow-gpu as Keras backend as well.

Supported Spark and Scala **

Component
Version

Scala

2.10

Spark

1.6.3

** Contact Skymind for customized or optimized versions for your environment.