Meeting notes

[jaivardhankapoor]

• manual seeding function for torch, numpy, determinism
• homogenize everything to torch
• separate run.py (run_{name}.py) for everyone, including configs/conf_{name}.yaml for exp_name, dataset
• Include plots/ in .gitignore
• data.py functions, decorator that changes numpy to torch if needed (@cast_to_torch)
• metrics.py, shape checking to (N,D)
• Add github issue (general with common checklist) and checklist for different issues (writing, metri coding etc.)

Comment which parts you work with

[jaivardhankapoor]

@jaivardhankapoor

• 7. Add github issue (general with common checklist) and checklist for different issues (writing, metri coding etc.)
• 4. Include plots/ in .gitignore

[manuelgloeckler]

@manuelgloeckler

• 1. Seeding
• 5. data decorator
• 6. shape checking