Intel and Microsoft team up to empower AI on Edge
Last week, we saw this post from Moe Tanabian, Vice President and General Manager of Azure Edge Devices at Microsoft, on the Microsoft Tech Community Internet of Things blog: Enhancing AI Across the Intelligent Edge Ecosystem with Intel.
The text mentions a future collaboration between Microsoft and Intel for the creation of a framework that will integrate Azure Artificial Intelligence and Machine Learning services with the Movidius Vision Processing Units and the OpenVINO toolkit.
While we wait for this collaboration to materialize, from the Bravent Innovation Lab, we are preparing a new Idiwork Experiment to test some features of the Intel Distribution of OpenVINO toolkit using models trained with Microsoft Azure Custom Vision.
Recently, we received an Intel Neural Compute Stick 2 (NCS2) device with a 4GB Intel Movidius Myriad X Vision Processing Unit (VPU) inside. This is a well-known USB plug and play AI device for hardware acceleration of Deep Neural Networks inferences at the Edge.
We will test the performance of this device connected to a Raspberry Pi or a Windows PC in a real application. We will also simulate workloads on the Intel DevCloud platform to compare the inference times of our Deep Learning models with different Edge devices such as:
- Intel CPUs
- Intel Integrated GPUs
- Intel FPGAs
- Intel Movidius VPUs
Stay tuned for the Idiwork and Bravent posts so you don’t miss a thing!