Alexander Suslov

Results 24 issues of Alexander Suslov

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

NNCF PT
NNCF Common

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

NNCF Common

### Changes Fixes of examples for Windows ### Reason for changes Support for running examples on Windows ### Related tickets N/A ### Tests test_examples 297

NNCF PT
NNCF OpenVINO
NNCF ONNX

### Discussed in https://github.com/openvinotoolkit/nncf/discussions/2547 Originally posted by **MinGiSa** March 5, 2024 I've been working on converting Torch models into OpenVINO models recently. I was able to easily do this for...

bug

### Context This issue proposes adding a test to the [post-training compression conformance suite](https://github.com/openvinotoolkit/nncf/blob/develop/tests/post_training/README.md) to verify that the weights of Tinyllama-1.1b PyTorch model can be compressed to INT8 in a...

good first issue

### Changes Bump up to `tensorflow_datasets==4.8.2` ### Reason for changes Regular update ### Related tickets 100587 ### Tests Sota eval build: 299

NNCF TF

Create the QAT example based on ultralytics yolov8. The example should be added in `examples/quantization_aware_training/torch/yolov8/` Motivation: - Demonstrating NNCF QAT API - Demonstrating the benefits of QAT - https://github.com/openvinotoolkit/nncf/issues/2486

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

NNCF PT
experimental
NNCF OpenVINO
NNCF ONNX
NNCF PTQ

### Changes Fix implementation of `argsort` function ### Reason for changes Align PyTorch and Numpy implementation ### Related tickets ref: 2588 ### Tests tests/common/test_tensor.py

experimental

### Context Using `compress_to_fp16=False` when saving a **compressed/quantized** OpenVINO model is no longer required following changes to OpenVINO https://github.com/openvinotoolkit/openvino/pull/23609 that are included in the OpenVINO 2024.1.0 release. This issue suggests...

good first issue