中文版 | English
Codes are based on the book author's codes and DeepLearningPython35 for adjustments to run in python3. There are few differences🌚.(My codes must be worse🙈)
I don't want to use other people's codes directly. So I wrote these codes by myself basing on code samples of the book, to help me understand the book.🤣
Network.py is based on the book. According to the author's question, I improved it to train all data as a matrix in a mini_batch in one time, rather than train data one after another, to speed the training.
Network2.py is based on the book and uses the matrix improvement in network.py.(Because I haven't use git in command line for a long time, I use git checkout -- network2.py carelessly. I don't know what changes have been discard, so……I choose to give up it and wouldn't try to restore those changes.😭)
Network3.py……emmm, I don't want to write it because I am lazy to install Theano.