- Yes this is another attempt at making a pytorch wrapper. Hopefully this will be the last
- A lot of loss functions
- Callbacks for optimizer
- Gradient Clipping
- Seed everything
- Check if certain layer
- Tabulate/count parameters
- Plot loss every n epochs
- I would like to be clear here. This is a fork of Tez. But they will accept no PRs and.. I plan to add a lot of things which may or may not be accepted. I am very grateful to Abhishek for being such a huge inspiration.
- Many changes that I will add will be based on fastai (Another huge thank you to Jeremy Howard, Sebastian Gugger and Rachel Thomas) and pytorch lightning (More thanks to William Falcon)
Firstly a huge thank you to all of these repositories
- Tez
- fastai
- pytorch lightning And also my own previous efforts
- sprintdl
- lightningAddons
- tinyDL Yes I have been trying this for a solid while :)