Kipling is a simple deep learning framework implemented from the scratch using Python. API style imitates Keras, the bottom layer uses the Tensor to carry on the gradient backpropagation automatically. the computational model implements the sequence model and the computational graph model, in which the computational graph model can realize any form of graph calculation (including residual connection, shared parameters, etc), also it has all modern layers like RNN, LSTM, CNN. This framework is mainly used to explore the internal implementation principle of deep learning framework.

• Websitehttps://github.com/AlbertBagdos256/MACHINE-LEARNING/tree/master/KIpling
• StackTensor, SGD Optimizer, RNN CNN and LSTM layers