Aleksei Kashapov

### Changes Extend graph tests with new detection and classification modesl; Add ONNX metatypes ### Reason for changes To cover all metatypes in the models ### Related tickets 88882 ###...

What is the reason for logging the message - "Please, provide execution parameters for optimal model initialization"? When I run the code below I see this message. It could confuse...

convert_version for MaskRCNN and FasterRCNN causes segmentation fault

1

# Bug Report ### Is the issue related to model conversion? No ### Describe the bug When I try to `convert_version()` function for Mask RCNN or Faster RCNN models from...

Introduce automatic adding post-processing nodes to ignored scopes for quantization

1

## Reopend #1255 ### Changes Add pre-quantization step in ONNXMinMaxQuantization to add nodes in ignored scope. Fix incorrect default parameters creation for PostTrainingQuantization and ONNXMinMaxQuantization. Now all parameters needed for...

### Changes Boost up requirements; Remove skl2onnx from requirements; Remove add_input_from_initializer ### Reason for changes ### Related tickets 91534 ### Tests No tests

[PTQ] Add support of arbitrary batch size for PTQ

7

### Changes Add a new advanced bool option for quantization - `batchwise_statistics`. When set to _True_ then statistics collection for supported algorithms (see below) are calculated with the assumption that...

[Good First Issue][NNCF]: Optimize memory footprint by removing redundant collected statistics

7

### Context One of the most frequently used algorithm inside NNCF is [PostTrainingQuantization](https://github.com/openvinotoolkit/nncf/blob/develop/nncf/quantization/algorithms/post_training/algorithm.py#L32). This algorithm operates within a pipeline structure, comprising multiple steps. Each step is independent and may involve...

[ONNX] Fix adding output nodes in NNCFGraph

2

### Changes Do not add edges for the output node if it has no parents. E.g. output is put on the initializer. ### Reason for changes 125248 ### Related tickets...

[OpenVINO] Update models for tests

1

### Changes Replace model at tests from OMZ models to analogs at `tests/torch/test_models`. All graph, scales references were updated accordingly. New dependency for OV tests were added: `timm`, `addict`, `efficientnet_pytorch`...

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