The evolution of IBM PowerAI Enterprise !

Source: Deep Learning on Medium


Go to the profile of Chuck Calio

IBM® PowerAI Enterprise has been rebranded as IBM Watson® Machine Learning (ML) Accelerator ! IBM Watson™ ML Accelerator V1.2 targets organizations moving from exploration and single-node development to scale-out production environments for machine learning and deep learning technologies. This unique offering helps to shorten the time to accuracy and can greatly improve the efficient use of resources across multiple data scientists, maximizing their productivity.

IBM Watson ML Accelerator V1.2 is “enterprise-ready” with regards to performance, dynamic scaling, multi-tenancy, resiliency, and security. Version 1.2 includes key value capabilities such as:

  • Multi-tenancy
  • Elastic distributed training
  • Parallel hyperparameter optimization
  • Training visualization
  • Transparent scaling from a single server to many servers

Version 1.2 also includes the following enhancements:

  • IBM Watson Studio integration enhancements
  • IBM Watson ML integration
  • Support for x86-based servers
  • Updated to IBM PowerAI V1.6
  • Full integrated support for PyTorch
  • Updated installer
  • Hyperparameter optimization update
  • Elastic Distributed Training Transparency for PyTorch
  • SnapML support
  • Python 3 support across all supported deep learning frameworks

Key Requirements to run this software include:

  • IBM Power® System AC922 servers or x86-based servers
  • Red Hat Enterprise Linux® 7.6

The planned availability date is March 22, 2019. The IBM announce letter is located at https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/5/897/ENUS219-055/index.html&request_locale=en