Now here: IBM Spectrum Deep Learning Impact for Power Systems 1.1
Helena. 270005V76W Visits (8272)
IBM Spectrum Deep Learning Impact for Power Systems 1.1 is now here; a complete enterprise-ready end-to-end solution for deep learning on IBM Power Systems. It leverages IBM PowerAI and combines IBM Spectrum Conductor with Spark with IBM Spectrum Conductor Deep Learning Impact, to deliver ready-to-use deep learning frameworks and a distributed deep learning infrastructure.
Data scientists now have a single offering that offers a highly available and resilient multitenant distributed framework, providing Apache Spark and deep learning application lifecycle support, centralized management and monitoring, end-to-end security and support from IBM.
The complete deep learning lifecycle supports installation and configuration, data ingest and transformation, hyperparameter search and optimization, and training a model over a multi-GPU multi-node distributed infrastructure which allows for faster training and quicker movement from training to inference and production.
While training over distributed infrastructure, data scientists can visualize results and get real-time insight on how training is progressing. This provides added flexibility in stopping training early, adjusting parameters and restarting. More importantly, the use of elastic resource allocation, distributes training across many GPUs and Power Systems, meaning that training runs can share the same resources at execution time. More efficient training runs, leads to quicker time to production.
Choose IBM Spectrum Deep Learning Impact for Power Systems 1.1 today! IBM Spectrum Deep Learning Impact for Power Systems 1.1 is now available on IBM Advanced Administrative System (AAS). To get started, see the offering release notes or publ