[Feature]: Set up DVC/LFS for PAF23

[MaxJa4]

Description

PAF22-Docs for DVC: Link
Official Docs (about the remote storage): Link
GitHub LFS Docs: Link

DVC was used in PAF22 for storing ML datasets and trained models to iterate/switch faster without commits or PRs and is optimized for large file sizes.

It was used in PAF22 together with Google Drive, where they used a dedicated PAF account (@ll7 has the credentials).

LFS doesn't need Google Drive or similar, it can use GitHub itself (< 2GB per file).

Definition of Done

• Handling ML datasets & trained models is solved

[ll7]

Additional information: The Google Drive account was created as a dedicated PAF-DVC account. I have the credentials, and you can use the account if you like.

[MaxJa4]

Great, thanks. Lets discuss this in the next review/planning.
Updated issue description accordingly.

[MaxJa4]

Hint: Installing DVC with pip install dvc dvc-gdrive like suggested in the PAF22 docs, installs a new DVC version which doesn't work for new repositories (so for us).
Use an older version for now: pip install dvc==3.23.0 dvc-gdrive