Feature spotlights

Rapid prototyping and deployment

Start data science projects anywhere with a shared compute resource pool. Reduce training times and produce higher quality models. Scale-out, enterprise-class training and inference services with API support for batch, streaming and interactive deployment.

End-to-end information architecture

Deploy deep learning as part of data and AI services with support for popular frameworks. Aggregate open source and third-party tools in a unified, governed environment.

Containerized infrastructure management

Run machine learning and deep learning models natively in Red Hat® OpenShift®. Deploy containerized models inside a firewall while keeping data on premises and maintaining cloud portability.

High resolution, large model support

Increase the amount of memory available for deep learning models beyond the GPU footprint. Implement more complex models with larger, more high-resolution images.

Multitenant deployment

Allocate and share compute powers tuned to model demands in a multitenant architecture. Securely share your compute resources across tenants to maximize use.

Autoscaling, autosearch and load balancing

Enable dynamic scaling of resources, up or down, based on policies to ensure higher priority jobs run fast. Build real-time training visualization and runtime model monitoring. Automate hyperparameter search and optimization for faster development.

AI lifecycle management

Prepare, build, run and manage machine learning and deep learning models. Run through the training cycle with more data to improve the model continuously.

Deployment validation and optimization

Increase reliability and resiliency for model deployment with precompiled and validated machine learning and deep learning models. Accelerate performance with software optimized to run on target systems.

Explainable AI with model monitoring

Manage and monitor deep learning models from small to enterprise-wide deployment. Monitor model fairness and explainability while mitigating model drift and risk.

Technical details

Software requirements

  • Red Hat® OpenShift® 4.5
  • RHEL 7.7
  • CUDA Deep Neural Network (cuDNN) 7.6.5 library
  • NVIDIA CUDA 10.2
  • NVIDIA GPU driver 440.33.01
  • NVIDIA NCCL2 2.5.6

Hardware requirements

  • x86 64-bit server with NVIDIA Tesla T4, P100 or V100 GPUs