Название: MATLAB Deep Learning HDL Toolbox User's Guide (R2021a)
Автор: MathWorks
Издательство: The MathWorks, Inc.
Год: 2021
Страниц: 278
Язык: английский
Формат: pdf (true)
Размер: 10.1 MB

Deep Learning HDL Toolbox provides functions and tools to prototype and implement Deep Learning networks on FPGAs and SoCs. It provides pre-built bitstreams for running a variety of Deep Learning networks on supported Xilinx and Intel FPGA and SoC devices. Profiling and estimation tools let you customize a Deep Learning network by exploring design, performance, and resource utilization tradeoffs.

Deep Learning HDL Toolbox enables you to customize the hardware implementation of your Deep Learning network and generate portable, synthesizable Verilog and VHDL code for deployment on any FPGA (with HDL Coder and Simulink).

Deep Learning is a branch of Machine Learning that teaches computers to do what comes naturally to humans: learn from experience. The learning algorithms use computational methods to “learn” information directly from data without relying on a predetermined equation as model. Deep Learning uses neural networks to learn useful representations of data directly from images. It is a specialized form of machine learning that can be used for applications such as classifying images, detecting objects, recognizing speech, and describing the content. The relevant features are automatically extracted from the images. The Deep Learning algorithms can be applied to supervised and unsupervised learning. These algorithms scale with data, that is, the performance of the network improves with size of the data.

You can train deep learning neural networks for classification tasks by using methods such as training
from scratch, or by transfer learning, or by feature extraction. Training a deep learning neural network from scratch requires a large amount of labeled data. To create the network architecture by using Neural Network ToolboxTM, you can use the built-in layers, define your own layers, or import layers from Caffe models. The neural network is then trained by using the large amounts of labeled data. Use trained network for predicting or classifying the unlabeled data. These networks can take few days or couple of weeks to train. Therefore, it is not a commonly used method for training networks.