Accelerated Deep Learning at the Edge Training and Inference

Featuring breakthrough technology for training at fixed-point Int8 coupled with high sparsity ratios
To enable deep learning at a fraction of the cost /power of GPU systems for fast, secure, and scalable AI at the edge

Featuring breakthrough technology for training at fixed-point Int8 coupled with high sparsity ratios To enable deep learning at a fraction of the cost /power of GPU systems for fast, secure, and scalable AI at the edge

Background

Training and inference demand large compute resources that utilize expensive power-hungry GPUs and consequently most deep learning compute is performed in the cloud or in large on-prem data centers. Training new models take a long time – a single cycle can take days and weeks to complete, and inference queries suffer from long latencies of the round-trip delays to and from the cloud.Yet, the data which feeds into the cloud systems for updating the training models and inference…

Technology

At the core of our technology is the ability to train at 8bit fixed-point and achieve trained models that are highly sparse, as opposed to 32bit floating-point and no sparsity which is the norm today with GPUs. These two technological breakthroughs enable AI platforms that are superior in performance…

Edge Training and Inference

IOT and the data explosion they bring are driving the need for intelligent edge processing. Devices in stores, factories, terminals, office buildings, hospitals, city streets, 5G cell sites, vehicles, farms, homes and hand-held mobile devices generate massive amounts of data…

Solution

Deep-AI uniquely provides an integrated, holistic and accelerated training and inference deep learning solution for the edge. We deploy our technology on off-the-shelf FPGA cards, eliminating the need for GPUs, and providing a 10X gain in performance/power or performance/cost versus a GPU. The FPGA hardware is completely under-the-hood and transparent…

Training at 8-bit
is now possible