Skip to main content

Welcome to H2O-3 Secure

H2O-3 Secure is a commercially licensed, in-memory, distributed, fast, and scalable machine learning and predictive analytics platform. It lets you build machine learning models on big data and provides easy productionalization of those models in an enterprise environment.

Basic framework​

H2O-3 Secure's core code is written in Java. A distributed key-value store is used to access and reference data, models, objects, etc. across all nodes and machines. The algorithms are implemented on top of H2O-3 Secure's distributed map-reduce framework and utilize the Java fork/join framework for multi-threading. The data is read in parallel and is distributed across the cluster. It is stored in-memory in a columnar format in a compressed way. H2O's data parser has built-in intelligence to guess the schema of the incoming dataset and supports data ingest from multiple sources in various formats.

REST API​

H2O-3 Secure's REST API allow access to all the capabilities of H2O-3 Secure from an external program or script through JSON over HTTP. The REST API is used by the R binding (H2O-R) and the Python binding (H2O-Python).

The speed, quality, ease-of-use, and model-deployment for our various supervised and unsupervised algorithms (such as Deep Learning, GLRM, or our tree ensembles) make H2O-3 Secure a highly sought after API for big data data science.

Available algorithms​

H2O-3 Secure supports the following algorithms:

Requirements​

At a minimum, we recommend the following for compatibility with H2O-3 Secure:

  • Operating Systems:

    • Windows 7 or later
    • OS X 10.9 or later
    • Ubuntu 12.04
    • RHEL/CentOS 6 or later

  • Languages: R and Python are not required to use H2O-3 Secure unless you want to use H2O in those environments, but Java is always required (see Java requirements).

    • R version 3 or later
    • Python 3.7.x, 3.8.x, 3.9.x, 3.10.x, 3.11.x, 3.12.x, 3.13.x, 3.14.x

  • numpy: H2O-3 Secure supports numpy<2 on Python 3.7-3.11 and numpy>=2 on Python 3.12+. The Python client tolerates numpy 2.x's new scalar repr at the Rapids-AST boundary, so passing numpy 2.x scalars to H2OFrame indexing works without conversion. Other numpy 2.x behaviour changes (deprecated aliases, integer-overflow semantics, np.bool8 removal) still apply to your own code — consult numpy's release notes. If a cascading dependency forces numpy 2 onto a Python 3.7-3.11 environment, recover with:

    pip install --force-reinstall 'numpy<2'

Java requirements​

H2O-3 Secure runs on Java. The 64-bit JDK is required to build H2O-3 Secure or run H2O-3 Secure tests. Only the 64-bit JRE is required to run the H2O-3 Secure binary using either the command line, R, or Python packages.

Java support​

H2O-3 Secure supports the following versions of Java:

  • Java SE 17
  • Java SE 16
  • Java SE 15
  • Java SE 14
  • Java SE 13
  • Java SE 12
  • Java SE 11
  • Java SE 10
  • Java SE 9
  • Java SE 8

Download the latest supported version of Java.

Unsupported Java versions​

We recommend that only power users force an unsupported Java version. Unsupported Java versions can only be used for experiments. For production versions, we only guarantee the Java versions from the supported list.

How to force an unsupported Java version​

The following code forces an unsupported Java version:

java -jar -Dsys.ai.h2o.debug.allowJavaVersions=19 h2o.jar

Optional requirements​

This section outlines requirements for optional ways you can run H2O-3 Secure.

Optional Conda requirements​

Conda is only required if you want to run H2O-3 Secure on the Anaconda cloud:

  • Conda packages target Python 3.7 - 3.14, matching the versions supported by the pip package.

User support​

H2O-3 Secure supports many different types of users.

Feedback