Machine Learning on FPGAs: Neural Networks

Machine learning is one of the fastest growing application models, and crosses every vertical market from the data center, to embedded vision applications in the Internet of Things (IoT) space, to medical and industrial applications. This short training video introduces the high-level concept of machine learning, focusing on Convolutional Neural Networks (CNNs). It also explains the benefits of using an FPGA in these applications. Watch the video to learn how to get started with CNNs on the FPGA, and contact your local salesperson to get more information.

CNN and GZIP implementation on FPGA by OpenCL

Convolutional Neural Network (CNN) is a Deep Learning algorithm used for various object classification. This AlexNet demo showcases the performance-per-watt advantage on a discrete Arria® 10 FPGA for an ImageNet scoring application. We are also showing a GZIP demonstration which compresses data at a high throughput and compression ratio. Both of these demos have been implemented in OpenCL™, which enables software programmers with limited experience in HDL. 

CNN Implementation on Intel FPGA Using OpenCL

Watch a short video on an introduction to machine learning and see a demo of the AlexNet CNN topology on Intel® FPGAs