feedforward-mnist

a custom feedforward neural network for mnist handwriten data

explanation for each file

• main.py

  主文件

• processor.py

  正向传播的实现

• backward_processor.py

  反向传播的实现

• data_loader.py

  加载mnist数据集用

• layer.py

  各种函数的实现

• recorder.py

  输出准确率、梯度等，保存神经网络数据

• test.py

  测试神经网络对于单个和所有数据的预测

---

• main.py

  just main

• processor.py

  implementation of forward propagation

• backward_processor.py

  implementation of back propagation

• data_loader.py

  loading minst data sets

• layer.py

  implementation of functions (softmax, sigmoid and so on)

• recorder.py

  output accuracy, gradient, etc. save neural network data

• test.py

  test the prediction of the neural network for individual and total data