My implementation of autograd class like Pytorch.autograd using Numpy

I add if node._prev: before node._backward since the original codes have a bug when the node is a leaf node.

Here are the plots of my DVG and Pytorch DVG:

Surprisely, my implementation of sclaer-level backpropagation is faster than Pytorch:

TODO

• scaler-based
• vector-based
• efficiency

• https://medium.com/sfu-cspmp/diy-deep-learning-crafting-your-own-autograd-engine-from-scratch-for-effortless-backpropagation-ddab167faaf5
• https://github.com/karpathy/micrograd