Stable diffusion pretraining improvements

[marwan116]

Summary of changes in ascending order of commit history (first commit up top):

• pep8 file renaming
• update filenaming references inside notebooks and readme
• use a README notebook instead of markdown for better rendering
• remove links to notebooks inside the table to avoid broken pages
• add a reference to our stable diffusion blog post in the README
• make a note in README that the template runs on a small subset of the…
• add a link back to the preprocessing guide inside preprocessing files
• merge the Scale to 2 billion and Train with your custom data sections
• use improved filenames (more concise and imply ordering)
• generate a README.md file for compatibility with template requirements
• keep code ordering and comments consistent across python scripts

Note given code changes are involved, I confirmed the template runs end-to-end successfully by cloning this branch onto this workspace and executing the notebooks

[marwan116]

@woshiyyya all your comments have been addressed - this should be good to merge from what I can tell, please let me know if there is anything else you'd like me to address