Added active param to augmentations

[klemen1999]

• Added active parameter for each augmentation
• Augmentation gets only currently active augmentations
• This makes it possible for Tuner to turn on/off specific augmentation per trial. Example part of config:

trainer:
  ...
  preprocessing:
    ...
     augmentations:
       - name: Defocus
         params:
           p: 0.1
tuner:
  params:
    trainer.preprocessing.augmentations.0.active_categorical: [True, False]

[klemen1999]

Nice future improvement would be to make _subset keyword for Tuner (e.g. trainer.preprocessing.augmentations: [["Defocus", "Sharpen", "RandomRotate90"], 2] - would make subset of size 2 from the list) but we would have to think how to make it general enough. If we make it for augmentations only this would mean just turning off active parameter for all unselected augs but if for example user tries to use it on parameter that doesn't have this active flag it would fail. Essentially the question is how general do we want to make this "_subset" tunning.

[kozlov721]

LGTM

[github-actions]

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines	Covered	Coverage	Threshold	Status
4717	3654	77%	0%	🟢

New Files

No new covered files...

Modified Files

File	Coverage	Status
luxonis_train/main.py	49%	🟢
luxonis_train/core/core.py	78%	🟢
luxonis_train/utils/config.py	95%	🟢
TOTAL	74%	🟢

updated for commit: 8c06a77 by action🐍

[github-actions]

Test Results

  6 files    6 suites   38m 40s ⏱️
 57 tests  25 ✅  25 💤  7 ❌
342 runs  177 ✅ 150 💤 15 ❌

For more details on these failures, see this check.

Results for commit 8c06a77.