Skip to content
How to predict probabilities from a linear layer, to achieve higher accuracy on the MNIST dataset.
C++ CMake
Branch: master
Clone or download
Greg Schoeninger
Greg Schoeninger fix comment
Latest commit 71dbee3 Nov 28, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data/mnist
include/neural
src
tests
tools/feedforward_neural_net
CMakeLists.txt
README.md

README.md

Improving Our Neural Network Outputs With Softmax Classification

How to predict probabilities from a linear layer, to achieve higher accuracy on the MNIST dataset.

Code to go along with blog: Improving Our Neural Network Outputs With Softmax Classification

Build

mkdir build

cd build

cmake -D BLAS_INCLUDE_DIR=/usr/local/opt/openblas/include \ -D BLAS_LIB_DIR=/usr/local/opt/openblas/lib \ -D GLOG_INCLUDE_DIR=~/Code/3rdParty/glog-0.3.5/glog-install/include/ \ -D GLOG_LIB_DIR=~/Code/3rdParty/glog-0.3.5/glog-install/lib/ \ -D GTEST_INCLUDE_DIR=~/Code/3rdParty/googletest-release-1.8.0/install/include/ \ -D GTEST_LIB_DIR=~/Code/3rdParty/googletest-release-1.8.0/install/lib/ ..

make

./tests

./feedforward_neural_net

You can’t perform that action at this time.