Mha files impl

[krisssnv]

No description provided.

[fennievdg]

Script looks clean, and it runs as expected nice job!
I see basically no differences in the results between what I ran and yours (see attached), but to try to reduce randomness as it is best practice, could you add a torch.seed(any number) in the appropriate files to try to control for randomness from pytorch operations? You can just add this at the top, one such file is serie.py. You can just check a bit with chatgpt which scripts it would make sense (based on their operations) to add a torch seed; always use the same number and make note of this number. No need to be very thorough about this; could be that serie.py is the only appropriate file.
comparison_results.csv

Can you remove the adding of .mha to the seriesinstanceuid in the SeriesInstanceUID collumn of the output?

Once these changes are made I can approve!

*General reminder, with you experiments on performance, please run each experiment 5x with that same seed just to demonstrate reliability of your results