Merge branch 'openvinotoolkit:develop' into support-transpose

Author: Shehrozkashif

…raining_time_compression/Quantization.md docs/usage/Quantization.mddocs/usage/training_time_compression/Quantization.md renamed to docs/usage/Quantization.md docs/usage/training_time_compression/Quantization.md renamed to docs/usage/Quantization.md

@@ -2,7 +2,7 @@
 
 Post-Training Quantization is a quantization algorithm that doesn't demand retraining of a quantized model.
 It utilizes a small subset of the initial dataset to calibrate quantization constants.
-Please refer to this [document](/docs/usage/training_time_compression/Quantization.md) for details of the implementation.
+Please refer to this [document](/docs/usage/Quantization.md) for details of the implementation.
 
 NNCF provides an advanced Post-Training Quantization algorithm, which consists of the following techniques:
 

docs/usage/post_training_compression/post_training_quantization/Usage.md

@@ -27,8 +27,8 @@ By default, the algorithm applies asymmetric 8-bit integer quantization (INT8_AS
 
 | Compression Mode | Element type | Scale type | Granularity              | Description                |
 |------------------|--------------|------------|--------------------------|----------------------------|
-| INT8_ASYM        | INT8         | FP16       | Per-channel              | [Asymmetric quantization](/docs/usage/training_time_compression/Quantization.md#asymmetric-quantization) |
-| INT8_SYM         | INT8         | FP16       | Per-channel              | [Symmetric quantization](/docs/usage/training_time_compression/Quantization.md#symmetric-quantization) |
+| INT8_ASYM        | INT8         | FP16       | Per-channel              | [Asymmetric quantization](/docs/usage/Quantization.md#asymmetric-quantization) |
+| INT8_SYM         | INT8         | FP16       | Per-channel              | [Symmetric quantization](/docs/usage/Quantization.md#symmetric-quantization) |
 
 #### Mixed precision modes
 
@@ -40,8 +40,8 @@ NNCF can automatically distribute precision assignments based on quantization se
 
 | Compression Mode | Element type | Scale type | Granularity              | Description |
 |------------------|--------------|------------|--------------------------|-------------|
-| INT4_SYM         | INT4         | FP16       | Per-channel / Group-wise | [Symmetric quantization](/docs/usage/training_time_compression/Quantization.md#symmetric-quantization) |
-| INT4_ASYM        | INT4         | FP16       | Per-channel / Group-wise | [Asymmetric quantization](/docs/usage/training_time_compression/Quantization.md#asymmetric-quantization) |
+| INT4_SYM         | INT4         | FP16       | Per-channel / Group-wise | [Symmetric quantization](/docs/usage/Quantization.md#symmetric-quantization) |
+| INT4_ASYM        | INT4         | FP16       | Per-channel / Group-wise | [Asymmetric quantization](/docs/usage/Quantization.md#asymmetric-quantization) |
 | NF4              | FP32         | FP16       | Per-channel / Group-wise | [NormalFloat-4](https://arxiv.org/pdf/2305.14314v1.pdf) lookup table with 16 FP32 values |
 | CODEBOOK         | Any          | FP16       | Per-channel / Group-wise | Arbitrary lookup table (codebook) |
 | CB4              | E4M3         | FP16       | Per-channel / Group-wise | A fixed lookup table with 16 E4M3 values based on NF4 values |

docs/usage/post_training_compression/weights_compression/Usage.md

@@ -3,7 +3,7 @@
 This is a step-by-step tutorial on how to integrate the NNCF package into the existing PyTorch projects.
 The use case implies that the user already has a training pipeline that reproduces training of the model in the floating point precision and pretrained model.
 The task is to prepare this model for accelerated inference by simulating the compression at train time.
-Please refer to this [document](/docs/usage/training_time_compression/Quantization.md) for details of the implementation.
+Please refer to this [document](/docs/usage/Quantization.md) for details of the implementation.
 
 ## Basic usage
 
@@ -70,27 +70,6 @@ quantized_model.load_state_dict(state_dict)
 
 You can save the `compressed_model` object `torch.save` as usual: via `state_dict` and `load_state_dict` methods.
 
-## Advanced usage
-
-### Compression of custom modules
-
-With no target model code modifications, NNCF only supports native PyTorch modules with respect to trainable parameter (weight) compressed, such as `torch.nn.Conv2d`.
-If your model contains a custom, non-PyTorch standard module with trainable weights that should be compressed, you can register it using the `@nncf.register_module` decorator:
-
-```python
-import nncf
-
-@nncf.register_module(ignored_algorithms=[...])
-class MyModule(torch.nn.Module):
-    def __init__(self, ...):
-        self.weight = torch.nn.Parameter(...)
-    # ...
-```
-
-If registered module should be ignored by specific algorithms use `ignored_algorithms` parameter of decorator.
-
-In the example above, the NNCF-compressed models that contain instances of `MyModule` will have the corresponding modules extended with functionality that will allow NNCF to quantize the `weight` parameter of `MyModule` before it takes part in `MyModule`'s `forward` calculation.
-
 ## Examples
 
 - See a PyTorch [example](/examples/quantization_aware_training/torch/resnet18/README.md) for **Quantization** Compression scenario on Tiny ImageNet-200 dataset.