AI

Spell & Graphcore to Deliver Next-Generation AI Infrastructure

Offers Free On-demand Trial of Graphcore’s IPU-POD AI Platform

Spell, the leader in operationalizing AI for natural language processing (NLP), machine vision, and speech recognition applications, and Graphcore, maker of the Intelligence Processing Unit (IPU) for next-generation AI compute, have announced a partnership to deliver the next generation of AI infrastructure.

The new solution integrates Graphcore’s IPU-POD scale-out systems with Spell’s eponymous hardware-agnostic MLOps software platform for deep learning (DLOps) to make advanced AI development faster, easier, and less expensive.

With the emergence of AI based on large neural nets, transformers, and other advanced AI models, there is a growing demand for more specialized AI compute instances, like Graphcore IPU systems, to accelerate model training, lower cost and enable new breakthroughs in ML.  Conventional MLOps tools for traditional machine learning do not provide the abstraction and automation needed to enable and assure compliance, reproducibility, and resource management, both human and computational, for deep learning AI.

To overcome these challenges, Graphcore and Spell are delivering a comprehensive deep learning platform that is powerful, easy-to-use, and highly cost-effective. And they are offering a free trial of this powerful hardware and software combination to commercial AI practitioners and academic researchers.

Graphcore’s IPU has consistently demonstrated its ability to accelerate the most widely used AI models and workloads, while also unlocking new directions in artificial intelligence, made possible by its unique processor, system architecture and mature Poplar software stack.

Graphcore will utilize Spell to provide developers free access (limited to 6 hours of total compute time) to both interactive Jupyter notebooks and direct remote execution to run IPUs in the cloud, supported by  a range of quickstart code tutorials leveraging popular AI models, including natural language models (e.g., BERT), computer vision models (e.g., EfficientNet, ResNet) and Graph Neural Networks (e.g., temporal graph networks).

“Graphcore is delighted to be partnering with Spell in our shared mission to broaden access to artificial intelligence, while accelerating its progress. Spell Workspaces make it easier than ever before for developers to try out the IPU and experience the Graphcore advantage first-hand,” said Graphcore CEO, Nigel Toon, adding, “Our integration with the Spell platform more broadly means that Graphcore can simply and seamlessly deliver leading-edge AI compute as part of the rich and expanding Spell ecosystem.”

Spell has been recognized by Gartner, Forrester, and others as an AI infrastructure pioneer for its unique end-to-end software platform’s ability to orchestrate and automate the entire AI workflow, and provide users with a single, common interface that hides the complexity and differences of underlying hardware and systems software.

With integrated support for the Graphcore IPU, Spell’s AI infrastructure platform now gives Spell users transparent access to state-of-the-art AI acceleration for existing and new workloads.

“We are pleased to collaborate with Graphcore to produce powerful next-generation AI infrastructure that meets the needs of deep learning practitioners for quickly delivering the most advanced AI applications at scale,” said Serkan Piantino, co-founder & CEO of Spell. “And our self-guided free trial means it could not be easier for users to get hands-on with this solution to see what it can do for them.”

Previous ArticleNext Article

Leave a Reply

Your email address will not be published. Required fields are marked *