Vasily Shamporov

Add a workaround for torch<1.11.0 bug with tensor indexing

2

### Changes For torch < 1.11.0, additionally wrap tensor indices for torch.Tensor in a tuple so that indexing a tensor using `TracedTensor` indices has correct behaviour. ### Reason for changes...

Use storage scopes when determining train-only modules

3

### Changes When replacing only non-train-only modules, check the scopes of the modules we iterate on against the scopes corresponding to the storage of the module, not the execution module...

### Changes ### Reason for changes ### Related tickets ### Tests

Synchronize quantizer setups for DistributedDataParallel cases

3

Now that the quantizer setup is being decided during `create_compressed_model`, and for precision init cases the resulting setup is dependent on the data loaders used for initialization, there is a...

Handle __torch_function__ API properly

1

In PyTorch 1.7.0 (at least), the PyTorch framework allows users to subclass `torch.Tensor` and to define a dispatch function called `__torch_function__` in order to perform custom behavior when a public...

### Changes As stated in the title. ### Reason for changes UX improvement due to not failing where we may keep working functionally. ### Related tickets 94348 ### Tests TBA

[Good First Issue] [NNCF] Make NNCF common graph code pass mypy checks

22

## Context [NNCF](https://github.com/openvinotoolkit/nncf) code is mostly typed with conventional type hints, but historically had started out and grown without a CI-enforced `mypy` conformance process in place. Since typing is paramount...