- Udemy course: https://www.udemy.com/course/pytorch-for-deep-learning
- repo: https://github.com/mrdbourke/pytorch-deep-learning/
- site: https://www.learnpytorch.io/
- Re-check momentum
- Check torchmetrics: https://torchmetrics.readthedocs.io/en/latest/
- Check out pytorch snippets extension
- Check doc for torchvision
- https://pytorch.org/docs/stable/data.html (especially DataLoader)
- Overhead, memory BW vs compute and others other considerations: https://horace.io/brrr_intro.html
- Setup Ubuntu + docker containers for Pytorch dev (if weems to be worth it)
- Setup bash aliases + Ubuntu settings^
- Check if issues with disk space, where Ubuntu is installed, etc.
- http://numsrvdc001:8090/pages/viewpage.action?pageId=91292102
- https://ubuntu.com/tutorials/enabling-gpu-acceleration-on-ubuntu-on-wsl2-with-the-nvidia-cuda-platform#3-install-nvidia-cuda-on-ubuntu
- Consider docker in Docker 4 windows vs WSL
- Main advantage might be bash aliases^