[Common] Help gc to collect unused model outputs (#3643)

### Changes

Variable which is containing model outputs during statistic collection
is explicitly freed after each iteration

### Reason for changes

The quantization cell was failing on my machine with 125GB of RAM in the
[SD v3 TorchFX
notebook](https://github.com/openvinotoolkit/openvino_notebooks/blob/latest/notebooks/stable-diffusion-v3-torch-fx/stable-diffusion-v3-torch-fx.ipynb).
After checking that the pytorch code is not leaking by the [pytorch
profiler tool](https://docs.pytorch.org/docs/stable/profiler)

<img width="1870" height="1145" alt="image"
src="https://github.com/user-attachments/assets/dba21040-8455-49b3-84fa-74e9aa3f9fd8"
/>

I confirmed that the problem is not related to statistic collection
directly. After that I forced the garbage collection by the `del`
statement - and my memory didn't exceed healthy 50GB during the runtime,
and SQ was successfully applied to the model (in contrast with previous
runs)

### Related tickets


### Tests

I'm not sure which test should I do with that

Author: daniil-lyakhov

src/nncf/common/tensor_statistics/aggregator.py

@@ -86,6 +86,11 @@ def collect_statistics(self, model: TModel, graph: NNCFGraph) -> None:
             outputs = engine.infer(input_data)
             processed_outputs = self._process_outputs(outputs)
             self._register_statistics(processed_outputs, merged_statistics)
+            # Manually dereference output tensors to hint gc to remove them. Without it,
+            # the processed_outputs and outputs remain during model inference,
+            # increasing the peak memory consumption.
+            del processed_outputs
+            del outputs
             processed_samples += 1
         if processed_samples == 0:
             raise nncf.ValidationError(EMPTY_DATASET_ERROR)