Overview

The CEVA Deep Neural Network (CDNN) is a comprehensive compiler technology that creates fully-optimized runtime software for CEVA-XM Vision DSPs and NeuPro AI processors. Targeted for mass-market embedded devices, CDNN incorporates a broad range of network optimizations, advanced quantization algorithms, data flow management and fully-optimized compute CNN and RNN libraries into a holistic solution that enables cloud-trained AI models to be deployed on edge devices for inference processing.

The CDNN compiler enable an extremely simple and streamlined transition of existing deep neural networks to an embedded environment. NeuPro AI processor ensure superior performance with minimal power consumption. Separately, each component of the CDNN compiler is a powerful enabler of embedded imaging and vision applications. Combined, these pieces deliver an ultimate toolkit to support new network structures and changing layer types of deep neural networks.

CEVA supplies a full development platform for partners and developers based on the CEVA-XM and NeuPro architectures to enable the development of deep learning applications using the CDNN, targeting any advanced network.

Benefits

The CDNN compiler streamlines implementations of deep learning in embedded systems by automatically optimize offline pre-trained neural networks to real-time embedded-ready networks for CEVA-XM cores and CEVA-NeuPro AI processors, using the CEVA Network Optimizer. This enables real-time, high-quality image classification, object recognition, and vision analytics.

Automatic conversion to embedded-ready networks
Supports a wide variety of neural network structures, including any number and type of layers of deep neural networks
Powerful hardware accelerators IP to maximize processing throughput

Main Features

  • The CEVA Network Optimizer converts pre-trained neural network models and weights from offline training frameworks (such as Caffe or TensorFlow, ONNX and others) to a real-time network model.
  • The CDNN runtime software framework accelerates deployment of machine learning in low-power embedded systems.
  • Coupled with the CEVA-XM family of intelligent vision processors and CEVA-NeuPro AI processors, the CDNN offers significant time-to-market and power advantages for implementing machine learning in embedded systems.

CEVA CDNN live demonstration

The CEVA Deep Neural Network (CDNN) is a comprehensive compiler technology that creates fully-optimized runtime software for CEVA-XM Vision DSPs and NeuPro AI processors. CDNN offers significant time-to-market and power advantages for implementing machine learning in embedded systems for smartphones, advanced driver assistance systems (ADAS), surveillance equipment, drones, robots and other camera-enabled smart devices.
In this video clip we demonstrate how a pre-trained neural network (“Age Classification”, can be found here: http://www.openu.ac.il/home/hassner/p…) is downloaded from the internet, converted to an embedded-friendly network, and runs on the CEVA-XM4 FPGA platform in less than 10 mins!