本篇案例研究不仅介绍了 LeNet*（一种进行手写数字识别的重要图像识别拓扑），还展示了如何利用训练工具在面向英特尔® 架构优化的 Caffe* 上对混合国家标准技术研究所 (MNIST) 数据集进行可视化设置、调试和训练。目标受众是数据科学家。
Caffe* is a deep learning framework developed by the Berkeley Vision and Learning Center (BVLC). Caffe optimized for Intel architecture is currently integrated with the latest release of Intel® Math Kernel Library (Intel® MKL) 2017 optimized for Advanced Vector Extensions (AVX)-2 and AVX-512 instructions which are supported in Intel® Xeon® and Intel® Xeon Phi™ processors (among others). This...
The main algorithm of speech recognition has changed to DNN (Deep Neural Network). Without internet, the speech recognition service in your mobile devices nearly useless, very few times it can listen to what you said and work.With support for the SSSE3 instruction set on Intel’s CPU, we could easy run a DNN based speech recognition application without the internet. Adding direct SSSE3 support...
Jack (Java* Android* Compiler Kit) is a new Google* tool that includes a compiler from Java source code to the Android dex file format. Jack has its own .jack library and provides most toolchain features as part of a single tool: repackaging, shrinking, obfuscation and multidex. There is also a tool that translates existing .jar files to the .jack library format.
This series of two articles discusses how data and memory layout affect performance and suggests specific steps to improve software performance. The basic steps shown in these two articles can yield significant performance gains. These two articles are designed at an intermediate level. It is assumed the reader desires to optimize software performance using common C, C++ and Fortran* programming...