caffe package analysis (linux)
①buildFolder for storing compilation results ②camke folder, will be used when using cmake ③Data folder, used to store the original data and the script to obtain new data, which contains cifar10, ilsvrc12, mnist data ④distribute folder, the location where the release package is generated after compilation ⑤ Docker folder, use Docker tool for migration ⑥ The docs folder contains documents with a lot of explanatory content, especially the tutorial ⑦examples folder,example ⑧includeFolders, header files ⑨Matlab folder, related to some operations of matlab ⑩modelsfolder, Installed models of deeplearning eg: each model's prototxt contains: deploy.prototxt describes the structure of this network, solver.prototxt describes hyperparameter configuration information train_val.prototxt is also the network structure described ⑪Python folder, some operations related to python ⑫scripts folder, used to store scripts ⑬src folder, Store CaffeSource codeThe place ⑭tools folder, Commonly used toolsSource code |
Focus:
① Source code: Mainly focus on the three folders src, tools and include;
②Example: Mainly use the three folders of build, models, examples
This article is used as a memo for "Deep Learning 21-Day Practical Caffe Notes", please use it in conjunction with the original book~ Foreword The Caffe framework provides a high degree of a...
helphelp Link: http: //pan.baidu.com/s/1hsnIPfe Password: xjpg brief introduction · · · · · · "Deep learning: 21 days combat Caffe" is a deep learni...
Summary of deep learning tools 1. Caffe【Convolutional Architecture for Fast Feature Embedding】 —Convolutional neural network based on C++/CUDA/Python implementation, providing command line, matl...
1. Deep learning DL: 1.1, Supervised learning, unsupervised learning, overfitting, training samples, generalization, training set, validation set, test set These knowledge related to deep learning nee...
Reference Bo Wen 1: http://blog.csdn.net/qq_32166627/article/details/52640730 Teacher Zhao's book The sixteenth day visualization method mainly used MATLAB interface method, and this book's blog post ...
Day 6 Run handwritten digit recognition routine 1. MNIST data set The MNIST data set includes 60,000 training sets and 10,000 test sets. Each image has been normalized by size and centered by numbers....
Day 8 data structure Caffe Rise: Net Drawings: Prototxt Blob: Brick layers: Layer SGD Method: Solver Jian Zhuangxiu: Train decoration: Finetune 8.1 Blob Blob represents a four-dimensional array (width...
The previous blog introduced how to use Faster-RCNN to train your own data set. After training, you will get a model. This blog describes how to use the trained model to test data. 1. The storage loca...
Programming for 5 minutes, environment 2 hours The specific steps refer to the following links: Then we proceed with the detailed process. The first step: we first download the source code: Step 2: Co...
Please mark the reprint URL: http://blog.csdn.net/frd2009041510 table of Contents(?)[+] Caffe study notes (1): Caffe architecture and model analysis Written in front: For how to quickly build the caff...