H2O is an open-source, in-memory, distributed platform designed for machine learning and predictive analytics on big data. Built with speed and scalability in mind, H2O enables you to effortlessly build and deploy high-quality models in an enterprise environment. H2O’s core code, written in Java, leverages a Distributed Key/Value store and a distributed Map/Reduce framework for seamless data access and processing across nodes. With intelligent data parsing and support for multiple formats, H2O simplifies data ingest from various sources.
This image is based on Ubuntu and contains H2O-3, the third incarnation of H2O. It is compatible with NVIDIA® GPUs available in Nebius Israel, which allows acceleration of machine learning and other compute-intensive applications.
-
Click the button in this card to go to VM creation. The image will be automatically selected under Image/boot disk selection.
-
Under Network settings, enable a public IP address for the VM (Public IP:
Auto
for a random address orList
if you have a reserved static address). -
Under Access, paste the public key from the pair into the SSH key field.
-
Create the VM.
-
Connect to the VM via SSH using local forwarding for TCP port 54321. For example:
ssh -i <path_to_public_SSH_key> -L 54321:localhost:54321 <username>:<VM's_public_IP_address>
The
ufw
firewall in this product only allows incoming traffic to port 22 (SSH). This is why you need local port forwarding when connecting. -
To use H2O Flow, the H2O-3 user interface, go to
http://localhost:54321
in your web browser.
H2O-3 is started as a systemd
daemon named h2o.service
, listening to incoming connections on port 54321. If you want to set H2O-3 up to start in another way (see Starting H2O in the H2O-3 documentation), you need to stop this service first:
sudo systemctl stop h2o.service
sudo systemctl disable h2o.service
H2O-3 binary files are located in /usr/local/h2o
.
- Financial services: credit scoring, personalized offers, customer churn, anti-money laundering, trading strategies.
- Healthcare: clinical workflow, capacity simulation, diagnosis assistance, infectious disease forecasting, pricing, claims assessment.
- Insurance: fraud detection and mitigation, claims management, personalized rate management and product bundling.
- Manufacturing: predictive design and maintenance, transportation optimization.
- Marketing: audience segmentation, content personalization, next best offer and next best action.
- Retail: assortment and pricing optimization.
- Telecom: customer support, fleet maintenance.
Nebius Israel does not provide technical support for the product. If you have any issues, please refer to the developer’s information resources.
Software | Version |
---|---|
Ubuntu | 22.04 LTS |
Docker | 24.0.2 |
NVIDIA CUDA Toolkit | 12.1.1 |
NVIDIA Container Toolkit | 1.13.3-1 |
NVIDIA Data Center Driver | 535.54.03 |
default-jdk | 2:1.11-72build2 |
default-jre | 2:1.11-72build2 |
h2o | 3.43.0 |