Intel FPGAs power Microsoft Project Brainwave AI
Today, Microsoft* opened a public preview of Azure Machine Learning Hardware Accelerated Models* powered by Project Brainwave*. This new AI inferencing service leverages Intel® Arria® 10 FPGAs to provide data scientists and developers with industry-leading inference performance in the cloud and on the edge. The open source community can now deploy new and existing models based on the ResNet 50 dataset to on-premises systems, and then take advantage of the Azure* accelerated cloud to access hardware that is specifically configured for their workloads.
AI developers can take full advantage of Intel FPGAs, combined with Intel® Xeon® processors, using familiar Python and TensorFlow* SDKs. The operating system-agnostic Azure cloud can precisely configure the hardware for peak performance without requiring developers to program the FPGAs themselves. In addition, developers can deploy image-recognition models to Microsoft-designed Intel FPGA-based systems on-premises that act as Azure IOT Edge devices and connect into the Azure IoT Hub. Leverage the performance of Intel FPGAs and Microsoft Project Brainwave to do real-time inference on your workload.
You can read more at Intel Developer Zone.
Image credit: Copyright ©2018 Microsoft Corporation. All rights reserved.