Open Nav

Fire up your cnvrg.io Metacloud training pipelines with Habana Gaudi AI processors

Share on linkedin
Share on twitter
Share on facebook
Share on whatsapp
Share on pocket

Gaudi AI processors deliver cost-efficient AI training. cnvrg.io Metacloud makes them easy to use.  

Enterprises wanting more value from their data science initiatives face increasing complexity of datasets, available algorithms, computing and storage resources, and implementation diversity, whether on-premises, across clouds, or in hybrid deployments. The cnvrg Metacloud managed service helps enterprises to tame this complexity and to promote MLOps best practices for continuous training and deployment of models into production.  This helps organizations realize more value from their data science investments more quickly.  

A key feature of Metacloud is a marketplace of best-in-class computing resources it offers developers for all types of AI or ML workloads, based on each hardware architecture’s cost and performance trade-offs.  For example, specialized AI processors are well-suited for training deep learning models in parallel.  By contrast, CPUs can be more cost-effective for real-time inferencing.

Now cnvrg.io is making it simple for developers to deploy Habana Gaudi AI processors in their training pipelines.  Within cnvrg Metacloud, developers can now choose the Gaudi-accelerated Amazon EC2 DL1.24 xlarge instance, which is optimized for training neural networks.  Gaudi-accelerated instances yield up to 40% better price-performance than existing GPU-based EC2 instances.

Putting that power to use quickly can be another matter.  For many data scientists, the sad reality is that they spend too much time on chores like deploying and configuring hardware, and less time on the data science work they’d rather be doing.  It can take months to re-configure software stacks for a new environment. Habana’s SynapseAI(R) software suite enables an easy transition of existing TensorFlow or PyTorch models to Gaudi HPUs (Habana Processing Units), whether those models are running in your own data center, in the cloud, or both.  Developers can also leverage Metacloud’s Git integration to add the Habana reference models repository.  This features over twenty popular AI models optimized for Gaudi training processors that developers can add to their Metacloud projects. 

Watch this demonstration shows how easy it is for developers to deploy the new Gaudi-accelerated DL1 instance in their Metacloud training pipelines within minutes!

Start building high-impact models today!

The cnvrg.io Metacloud managed platform offers developers the ability to take advantage of their own compute and storage, while also offering native integrations to major cloud providers and leading OEMs. It removes the operational complexity of training and deploying models into production. With the newly available Gaudi-based AWS DL1 instances now enabled in Metacloud, you can immediately take advantage of the combined cost-performance benefits of Gaudi and the MLOps capabilities of cnvrg.io. 

 Learn more about cnvrg.io Metacloud, and sign up to get started in minutes!  You can also visit the Habana developer site for documentation, examples, and tutorials to help you build new or migrate existing AI models and optimize their performance.

Top MLOps guides and news in your inbox every month

Share on facebook
Share on twitter
Share on linkedin
Share on whatsapp
Share on pocket