[Good First Issue] [NNCF] Make NNCF common accuracy aware training code pass mypy checks about nncf

[siddhant-0707]

.take

[anzr299]

@p-wysocki @vshampor I've created a PR #2637

[github-actions]

Thank you for looking into this issue! Please let us know if you have any questions or require any help.

[tvilight4]

@vshampor missed this!! will wait for the next one!! thanks for informing!!

[vshampor]

@tvilight4 I have created #2491 now, take a look.

[tvilight4]

@vshampor thanks!! got assinged for it.
will start working on the same

[p-wysocki]

Hello @tvilight4, are you still working on that issue or can we reassign it?

[p-wysocki]

@vshampor could you please either link the PR to the development section of the issue or grant me some permissions in NNCF repo to do this myself?

I'd prefer the latter since I'm keeping an eye on all GFIs weekly.

[anzr299]

@p-wysocki Is this issue being actively worked on? If not, I'm down to work on it!

[p-wysocki]

Hi @anzr299, sure! The thing is this issue is in NNCF repository and I do not have correct access rights to assign you. @vshampor could you please do it?

For now let's agree the task is yours, feel free to start working on it and ask questions. We'll figure out the formalities in the background. :)

[anzr299]

Sure, thank you very much! I have already started working on it.

[anzr299]

Hi @p-wysocki, what is the update regarding the assignment? I was off for some time due to my dissertation submission. I am resuming work on this issue now.

[p-wysocki]

Done, please let us know if you have any questions @anzr299. :)

[anzr299]

Sure, I am following the current types that are mentioned in some files. For arguments with kwargs and args, I am planning on using "...", is that fine?

[p-wysocki]

I found some good tips here, I think you can pick one you'd like to use and it's going to be fine - worst case scenario we'll change it.

[anzr299]

Also, I was wondering if the the | method would be preferred over the Optional() method to denote an option between type and None since the latest updates suggest a movement towards the | direction but the current NNCF code mostly uses optional().

[anzr299]

I am facing a problem in the following section of training_loop.py:

Mypy is giving an error saying the CompressionScheduler doesn't contain current_sparsity_level and current_pruning_level. A similar problem exists in the _interpolate_compression_step_update method for runner.maximal_accuracy_drop.