Proposal to add SOTA major computing platforms such as PyTorch, Tensorflow, MXNet, PaddlePaddle and Skitcit-Learn, etc.

[zhimin-z]

Extending #348, I propose to add these tools due to their prevalence in today's production-level machine-learning pipeline.

For this pr, I would add limited tools (here FIVE only) since the computing platform market is maturing now.
Also, this pr would NOT include general backend computing libraries such as numpy, scipy, pandas, etc.