# mini-dnn-cpp

**Repository Path**: aczz/mini-dnn-cpp

## Basic Information

- **Project Name**: mini-dnn-cpp
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-05-15
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# mini-dnn-cpp
**mini-dnn-cpp** is a C++ demo of deep neural networks. It is implemented purely in C++, whose only dependency, Eigen, is header-only. 

## Usage
Download and unzip [MNIST](http://yann.lecun.com/exdb/mnist/) dataset in `mini-dnn-cpp/data/mnist/`.

```shell
mkdir build
cd build
cmake ..
make
```

Run `./demo`.

Result: 
simple neural network with 3 FC layers can obtain `0.97+` accuracy on MNIST testset.
LeNet-like CNN can obtain `0.98+` accuracy on MNIST testset.