Refactor existing Pytorch code to Pytorch Lightning.

Training on larger images and/or deeper models requires a lot of careful refactoring and debugging. Also, they usually take longer to train.

Some of the advanced deep learning techniques like -

1. Mixed Precision
2. Multi-GPU/Multi-Node
3. Gradient Accumulation.
4. Logging in distributed systems.
5. Running in different environments like CPU/GPU/TPU.

have drastically improved model training time in recent years.

In this video, we will see how Pytorch Lightning makes a lot of these challenges easier. We will demonstrate how you can migrate any of your existing Pytorch code to Pytorch Lightning.