Qualcomm releases deep learning SDK for mobile platforms
Qualcomm Technologies announced the immediate availability of the Qualcomm Snapdragon Neural Processing Engine (NPE) software development kit (SDK) on Qualcomm Developer Network.
The Snapdragon NPE marks the first deep learning software framework designed for Snapdragon Mobile Platforms, providing developers with tools to accelerate deep neural network workloads on mobile and other edge devices powered by Snapdragon processors.
Developers can select any optimal Snapdragon core for the desired user experience, including Qualcomm Kryo CPU, Qualcomm Adreno GPU or Qualcomm Hexagon DSP.
The SDK is designed to offer developers in multiple industries — including mobile, automotive, healthcare, security and imaging — the tools they require to deliver on-device neural network-driven user experiences.
Developers can benefit from deep learning user experiences like style transfers and filters (augmented reality), scene detection, facial recognition, natural language understanding, object tracking and avoidance, gesturing, and text recognition.
A few developers have already begun using the Snapdragon NPE for their mobile app experiences, such as Facebook, which unveiled plans to integrate the Snapdragon NPE into the camera of the Facebook app to accelerate Caffe2-powered AR features.
By using the Snapdragon NPE, Facebook can achieve 5-times better performance on the Adreno GPU, compared to a generic CPU implementation, delivering a more fluid, seamless and realistic application of AR features when capturing photos and live videos.
“We developed the Snapdragon NPE SDK to help meet the increasing demand for running deep neural networks efficiently on mobile and other edge devices powered by Snapdragon,” said Gary Brotman, director of product management at Qualcomm Technologies, Inc. “The Snapdragon NPE was created to give developers the tools to easily migrate intelligence from the cloud to edge devices, and affords them the flexibility to target the appropriate Snapdragon core with the power and performance profile for a given user experience.”
The Snapdragon NPE, which is compatible with Snapdragon 600 and 800 Series Mobile Platforms, is designed to support common deep learning frameworks such as Caffe, Caffe2 and Tensorflow, as well as offers support for custom layers.
The SDK includes runtime software, libraries, APIs, offline model conversion tools, sample code, documentation, and debugging and benchmarking tools.
The Snapdragon NPE can now be downloaded on the Qualcomm Developer Network.
Previously reported, Qualcomm recently integrated Precise BioMatch Mobile, an algorithm solution for secure fingerprint recognition in smartphones and tablets, into its Fingerprint Sensors, the company’s next-generation ultrasonic fingerprint solutions.