From 94d1d7a0336f211070bcb13a4dc9b453d3c473b4 Mon Sep 17 00:00:00 2001 From: Tatiana Savina Date: Mon, 4 Dec 2023 16:36:03 +0100 Subject: [PATCH] [DOCS] Add legacy notes pot (#21449) * add note to pot * change header --- .../post_training_optimization_tool.rst | 8 ++++---- .../post_training_optimization_tool/pot_api_reference.rst | 6 ++++-- .../post_training_optimization_tool/pot_cli.rst | 7 ++++--- .../pot_cli/configuration_file_description.rst | 6 ++++-- .../pot_cli/simplified_mode.rst | 6 ++++-- .../post_training_optimization_tool/pot_examples.rst | 5 +++-- .../pot_examples/pot_api_examples.rst | 6 ++++-- .../pot_api_examples/pot_example_3d_segmentation.rst | 6 ++++-- .../pot_api_examples/pot_example_classification.rst | 6 ++++-- .../pot_api_examples/pot_example_face_detection.rst | 6 ++++-- .../pot_api_examples/pot_example_object_detection.rst | 6 ++++-- .../pot_api_examples/pot_example_segmentation.rst | 6 ++++-- .../pot_examples/pot_api_examples/pot_example_speech.rst | 6 ++++-- .../pot_examples/pot_cli_example.rst | 6 ++++-- .../post_training_optimization_tool/pot_faq.rst | 6 +++--- .../post_training_optimization_tool/protecting_model.rst | 8 ++++++-- .../quantization_best_practices.rst | 5 +++-- .../quantization_best_practices/saturation_issue.rst | 6 ++++-- .../post_training_optimization_tool/quantizing_models.rst | 6 ++++-- .../quantizing_models/default_quantization_algorithm.rst | 6 ++++-- .../quantizing_models_with_accuracy.rst | 5 +++-- .../accuracy_aware_algorithm.rst | 6 ++++-- 22 files changed, 86 insertions(+), 48 deletions(-) diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool.rst index 234f6dca249..840a53b5c23 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool.rst @@ -1,6 +1,6 @@ .. {#pot_introduction} -(Deprecated) Post-training Quantization with POT +[Deprecated] Post-training Quantization with POT ================================================ @@ -14,15 +14,15 @@ API Reference Command-line Interface Examples - pot_docs_FrequentlyAskedQuestions + Post-training Optimization Tool FAQ (Experimental) Protecting Model -.. note:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. -For the needs of post-training optimization, OpenVINO™ provides a **Post-training Optimization Tool (POT)** +For the needs of post-training optimization, OpenVINO provides a **Post-training Optimization Tool (POT)** which supports the **uniform integer quantization** method. This method allows moving from floating-point precision to integer precision (for example, 8-bit) for weights and activations during inference time. It helps to reduce the model size, memory footprint and latency, as well as improve the computational efficiency, using integer arithmetic. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_api_reference.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_api_reference.rst index a837398e652..bbbf724b043 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_api_reference.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_api_reference.rst @@ -1,7 +1,9 @@ .. {#pot_compression_api_README} -API Reference -============= +[Deprecated] API Reference +================================= + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. Post-training Optimization Tool API provides a full set of interfaces and helpers that allow users to implement a custom optimization pipeline for various types of DL models including cascaded or compound models. Below is a full specification of this API: diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli.rst index a0b2887536b..d893a683b2d 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli.rst @@ -1,7 +1,7 @@ .. {#pot_compression_cli_README} -Use Post-Training Optimization Tool Command-Line Interface (Model Zoo flow) -=========================================================================== +[Deprecated] Use Post-Training Optimization Tool Command-Line Interface (Model Zoo flow) +==================================================================================================== .. toctree:: @@ -9,9 +9,10 @@ Use Post-Training Optimization Tool Command-Line Interface (Model Zoo flow) :hidden: Simplified Mode - pot_configs_README + Configuration File Description +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. Introduction #################### diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/configuration_file_description.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/configuration_file_description.rst index 4f0729fb4d6..d4f41bc14e1 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/configuration_file_description.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/configuration_file_description.rst @@ -1,7 +1,9 @@ .. {#pot_configs_README} -Configuration File Description -============================== +[Deprecated] Configuration File Description +============================================== + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. The tool is designed to work with the configuration file where all the parameters required for the optimization are specified. These parameters are organized as a dictionary and stored in diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/simplified_mode.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/simplified_mode.rst index a7b9465aaab..de5fccd0af7 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/simplified_mode.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_cli/simplified_mode.rst @@ -1,7 +1,9 @@ .. {#pot_docs_simplified_mode} -Optimization with Simplified Mode -================================= +[Deprecated] Optimization with Simplified Mode +==================================================== + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. Introduction diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples.rst index 0693a9bdf49..8d0cb6f0444 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples.rst @@ -1,7 +1,7 @@ .. {#pot_examples_description} -Examples -======== +[Deprecated] Examples +======================== .. toctree:: @@ -11,6 +11,7 @@ Examples API Examples Command-line Example +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This section provides a set of examples that demonstrate how to apply the post-training optimization methods to optimize various models from different domains. It contains optimization recipes for concrete models, that unnecessarily cover your case, but which should be sufficient to reuse these recipes to optimize custom models: diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples.rst index 19cb0616e32..f4eea83bed5 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples.rst @@ -1,7 +1,7 @@ .. {#pot_example_README} -Post-training Optimization Tool API Examples -============================================ +[Deprecated] Post-training Optimization Tool API Examples +=============================================================== .. toctree:: @@ -15,6 +15,8 @@ Post-training Optimization Tool API Examples Quantizing 3D Segmentation Model Quantizing for GNA Device +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. + The Post-training Optimization Tool contains multiple examples that demonstrate how to use its :doc:`API ` to optimize DL models. All available examples can be found on `GitHub `__. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_3d_segmentation.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_3d_segmentation.rst index 7217d513142..a80701ad9c5 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_3d_segmentation.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_3d_segmentation.rst @@ -1,7 +1,9 @@ .. {#pot_example_3d_segmentation_README} -Quantizing 3D Segmentation Model -================================ +[Deprecated] Quantizing 3D Segmentation Model +================================================================ + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Tool API ` for the task of quantizing a 3D segmentation model. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_classification.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_classification.rst index fe7f79a8c31..6bf8d2ec310 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_classification.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_classification.rst @@ -1,7 +1,9 @@ .. {#pot_example_classification_README} -Quantizing Image Classification Model -===================================== +[Deprecated] Quantizing Image Classification Model +======================================================== + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Tool API ` for the task of quantizing a classification model. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_face_detection.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_face_detection.rst index 587f7f6e2bc..2f3964bbcd4 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_face_detection.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_face_detection.rst @@ -1,7 +1,9 @@ .. {#pot_example_face_detection_README} -Quantizing Cascaded Face detection Model -======================================== +[Deprecated] Quantizing Cascaded Face detection Model +============================================================ + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Tool API ` for the task of quantizing a face detection model. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_object_detection.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_object_detection.rst index ce158f20195..860d60058d1 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_object_detection.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_object_detection.rst @@ -1,7 +1,9 @@ .. {#pot_example_object_detection_README} -Quantizing Object Detection Model with Accuracy Control -======================================================= +[Deprecated] Quantizing Object Detection Model with Accuracy Control +================================================================================ + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Toolkit API ` to quantize an object detection model in the :doc:`accuracy-aware mode `. The `MobileNetV1 FPN `__ model from TensorFlow for object detection task is used for this purpose. A custom ``DataLoader`` is created to load the `COCO `__ dataset for object detection task and the implementation of mAP COCO is used for the model evaluation. The code of the example is available on `GitHub `__. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_segmentation.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_segmentation.rst index 5b13e99ec09..5c6cf596962 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_segmentation.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_segmentation.rst @@ -1,7 +1,9 @@ .. {#pot_example_segmentation_README} -Quantizing Semantic Segmentation Model -====================================== +[Deprecated] Quantizing Semantic Segmentation Model +============================================================= + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Tool API ` for the task of quantizing a segmentation model. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_speech.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_speech.rst index 6ac3962c283..56f153b4665 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_speech.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_api_examples/pot_example_speech.rst @@ -1,7 +1,9 @@ .. {#pot_example_speech_README} -Quantizing for GNA Device -========================= +[Deprecated] Quantizing for GNA Device +========================================= + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This example demonstrates the use of the :doc:`Post-training Optimization Tool API ` for the task of quantizing a speech model for :doc:`GNA ` device. Quantization for GNA is different from CPU quantization due to device specifics: GNA supports quantized inputs in INT16 and INT32 (for activations) precision and quantized weights in INT8 and INT16 precision. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_cli_example.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_cli_example.rst index 7d0e016647f..c5fd319d988 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_cli_example.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_examples/pot_cli_example.rst @@ -1,7 +1,9 @@ .. {#pot_configs_examples_README} -End-to-end Command-line Interface Example -========================================= +[Deprecated] End-to-end Command-line Interface Example +========================================================= + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. This tutorial describes an example of running post-training quantization for the **MobileNet v2 model from PyTorch** framework, diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_faq.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_faq.rst index bfc2f0276fb..5a495d77c63 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_faq.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/pot_faq.rst @@ -1,10 +1,10 @@ .. {#pot_docs_FrequentlyAskedQuestions} -Post-training Optimization Tool FAQ -=================================== +[Deprecated] Post-training Optimization Tool FAQ +=========================================================== -.. note:: +.. danger:: Post-training Optimization Tool has been deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for post-training quantization instead. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/protecting_model.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/protecting_model.rst index 9ce1b2bf0ce..1eeeddfe0a2 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/protecting_model.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/protecting_model.rst @@ -1,8 +1,12 @@ .. {#pot_ranger_README} -Experimental: Protecting Deep Learning Model through Range Supervision ("RangeSupervision") -=========================================================================================== +[Deprecated] Experimental: Protecting Deep Learning Model through Range Supervision ("RangeSupervision") +================================================================================================================ +.. danger:: + + Post-training Optimization Tool has been deprecated since OpenVINO 2023.0. + :doc:`Neural Network Compression Framework (NNCF) ` is recommended for post-training quantization instead. Introduction #################### diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices.rst index a2b31b35fc8..f5f4e3d08ec 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices.rst @@ -1,7 +1,7 @@ .. {#pot_docs_BestPractices} -Post-Training Quantization Best Practices -========================================= +[Deprecated] Post-Training Quantization Best Practices +======================================================== .. toctree:: @@ -10,6 +10,7 @@ Post-Training Quantization Best Practices Saturation Issue +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. The :doc:`Default Quantization ` of the Post-training Optimization Tool (POT) is the fastest and easiest way to get a quantized model. It requires only some unannotated representative dataset to be provided in most cases. Therefore, it is recommended to use it as a starting point when it comes to model optimization. However, it can lead to significant accuracy deviation in some cases. The purpose of this article is to provide tips to address this issue. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices/saturation_issue.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices/saturation_issue.rst index d78a29ed03e..369adc5b981 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices/saturation_issue.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantization_best_practices/saturation_issue.rst @@ -1,7 +1,9 @@ .. {#pot_saturation_issue} -Saturation (overflow) Issue Workaround -====================================== +[Deprecated] Saturation (overflow) Issue Workaround +======================================================= + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. Introduction diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models.rst index 9005f296d58..48bfb24fe9c 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models.rst @@ -1,7 +1,7 @@ .. {#pot_default_quantization_usage} -Quantizing Models -================= +[Deprecated] Quantizing Models +====================================== @@ -12,6 +12,8 @@ Quantizing Models DefaultQuantization Method +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. + This guide describes how to apply model quantization with the Default Quantization method without accuracy control, using an unannotated dataset. To use this method, create a Python script using an API of Post-Training Optimization Tool (POT) and implement data preparation logic and quantization pipeline. If you are not familiar with Python, try :doc:`command-line interface ` of POT which is designed to quantize models from diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models/default_quantization_algorithm.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models/default_quantization_algorithm.rst index 20793e7e8f8..78c5f34df09 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models/default_quantization_algorithm.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models/default_quantization_algorithm.rst @@ -1,7 +1,9 @@ .. {#pot_compression_algorithms_quantization_default_README} -DefaultQuantization Parameters -============================== +[Deprecated] DefaultQuantization Parameters +======================================================== + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. The DefaultQuantization Algorithm is designed to perform fast and accurate quantization. It does not offer direct control over the accuracy metric itself but provides many options that can be used to improve it. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy.rst index 679394d6a53..0cac04e0e61 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy.rst @@ -1,7 +1,7 @@ .. {#pot_accuracyaware_usage} -Quantizing Models with Accuracy Control -======================================= +[Deprecated] Quantizing Models with Accuracy Control +================================================================ .. toctree:: @@ -10,6 +10,7 @@ Quantizing Models with Accuracy Control AccuracyAwareQuantization Method +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. The Accuracy-aware Quantization algorithm allows performing quantization while maintaining accuracy within a pre-defined range. Note that it should be used only if the :doc:`Default Quantization ` introduces a significant accuracy degradation. The reason for it not being the primary choice is its potential for performance degradation, due to some layers getting reverted to the original precision. diff --git a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy/accuracy_aware_algorithm.rst b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy/accuracy_aware_algorithm.rst index f47bf5f8743..42ba337e79b 100644 --- a/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy/accuracy_aware_algorithm.rst +++ b/docs/articles_en/documentation/openvino_legacy_features/post_training_optimization_tool/quantizing_models_with_accuracy/accuracy_aware_algorithm.rst @@ -1,7 +1,9 @@ .. {#accuracy_aware_README} -AccuracyAwareQuantization Parameters -==================================== +[Deprecated] AccuracyAwareQuantization Parameters +======================================================== + +.. danger:: Post-training Optimization Tool is deprecated since OpenVINO 2023.0. :doc:`Neural Network Compression Framework (NNCF) ` is recommended for the post-training quantization instead. Introduction