Developer Kits with Intel® Core™ Processors
Preinstalled Software
Get the most out of your hardware performance with the Intel® Distribution of OpenVINO™ toolkit, Ubuntu* Desktop LTS, and libraries.
Pretrained Models for Acceleration
Choose from a variety of optimized detection and recognition models for developing deep learning applications.
Training Extensions for Deep Learning
Modify, customize, train, and extend computer vision models for deep learning and inference optimization.
Overview
Develop and deploy solutions on Intel's latest high-performance platforms.
- Up to 46% better deep learning inference performance
- Preinstalled tools and SDKs included on kits
- Quick setup with preloaded samples
- Accelerate workloads on CPUs and GPUs
- Offload multiple workloads to optional Intel® Vision Accelerator products
Who Needs This Product
System integrators, independent software vendors (ISV), and IoT developers who create solutions using the following processes and applications:
- Computer vision
- Media encoding and decoding
- Signal and data processing
- High-performance deep learning applications across Intel® platforms
- Applications that run across all Intel® Vision products
- Ability to offload multiple workloads
Reference Implementations
These solutions have been prebuilt and validated for developers to test and deploy industrial, retail, and digital security applications enabled for computer vision.
Hardware
Prevalidated developer kits with an optional vision accelerator.
IEI Flex-BX200 AIoT Developer Kit
Optional: IEI* Mustang-V100-MX8 with Intel® Vision Accelerator Design
Intel® Vision Accelerator Design
This option enables the following capabilities:
- Delivers high-performance machine vision at ultra-low power.
- Offloads workloads to increase available processing efficiency.
- Is engineered for performance and inferencing at the edge.
Review the supported pretrained models for the Intel Distribution of OpenVINO toolkit. Learn More
Software
Intel® Distribution of OpenVINO™ Toolkit
- Enable convolutional neural network-based deep-learning inference on the edge.
- Support heterogeneous running across various accelerators—CPUs, GPUs, Intel® Movidius™ Neural Compute Sticks (NCS), and Intel® Vision Accelerator Design products—using a common API.
- Speed up time to market via a library of functions and preoptimized kernels.
Overview | Training | Documentation | Get Started | Forum
Intel® Developer Cloud
Explore this cloud-hosted AI development platform with access to Intel® hardware to find a solution that meets your needs.
- Try a variety of hosted hardware including Intel Atom®, Intel Core, and Intel® Xeon® processors or the Intel® Movidius™ VPU.
- Upload your own AI model or choose from a library of prebuilt sample applications.
- Benchmark hardware solutions to optimize for your performance needs.
Intel® Media SDK
- Accelerate rich media performance. Speed up video playback, encoding, processing, and media formatting conversion.
- Prototype, optimize, and productize your media pipelines with a comprehensive, convenient API.
- Debug and customize your products quickly.
Overview | Get Started | Support
Intel® oneAPI Base & IoT Toolkit
The combination of Intel® oneAPI Base Toolkit and Intel® oneAPI IoT Toolkit provides developers what they need to implement efficient, reliable, cross-architecture IoT solutions that run at the network edge. It delivers a core set of high-performance build tools and libraries and analysis tools to simplify IoT system design, development, and deployment across CPUs, GPUs, FPGAs, and other accelerator architectures.
Product and Performance Information
As measured by a ResNet-50 network with a batch size of 1 and int8 precision on Intel® Core™ i7-9700E vs. Intel® Core™ i7-7700 processor Intel® Core™ i7-9700E processor, PL1= 65 W TDP, 8C8T, Turbo up to 4.4 GHz, Intel® UHD Graphics 630, Motherboard: Asus* Prime Q370M-C, Memory: 2 times 8 GB DDR4-2666, Storage: 512 GB Intel® SSD 545s Series, OS: Ubuntu* 18.04 LTS (Bionic Beaver). Intel® Core™ i7-7700 processor, PL1=65W TDP, 4C8T, Turbo up to 4.2 GHz, Intel® HD Graphics 630, Motherboard: Asus* Prime Q270M-C, Memory: 2 times 8 GB DDR4-2400, Storage: 512 GB Intel® SSD 545s Series, OS: Ubuntu 18.04 LTS (Bionic Beaver). ResNet-50 is a popular CNN (Convoluted Neural Network) architecture that uses inception modules for computer vision recognition tasks (such as image classification). Image classification is the task of classifying a given image into one of the predefined categories. Reported metrics: Performance = Images/Second Performance results are based on testing as of May 24, 2019 and may not reflect all publicly available security updates. No product can be absolutely secure.