PyTorch (https://pytorch.org) is an open source machine learning platform for Python. It is specifically designed for deep learning applications, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Generative Adversarial Networks (GANs), and it includes extensive layer definitions for these applications. It has built-in tensor operations that are designed to be used in the same way as NumPy arrays, and they are also optimized to run on GPUs for fast computation. It provides an automatic computational graph scheme so that you won't need to calculate derivatives by hand.
After around 3 years of development and improvements, PyTorch has finally reached its newest milestone, version 1.3! What comes with it is a big package of new features and new functionalities. Don't worry about whether you'll have to...
