{"id":18965327,"url":"https://github.com/sony/model_optimization","last_synced_at":"2025-05-14T16:07:19.476Z","repository":{"id":36968465,"uuid":"378795568","full_name":"sony/model_optimization","owner":"sony","description":"Model Compression Toolkit (MCT) is an open source project for neural network model optimization under efficient, constrained hardware. This project provides researchers, developers, and engineers advanced quantization and compression tools for deploying state-of-the-art neural networks.","archived":false,"fork":false,"pushed_at":"2025-04-10T07:31:35.000Z","size":23092,"stargazers_count":382,"open_issues_count":8,"forks_count":64,"subscribers_count":22,"default_branch":"main","last_synced_at":"2025-04-10T08:33:06.650Z","etag":null,"topics":["deep-learning","deep-neural-networks","edge-ai","machine-learning","network-compression","network-quantization","neural-network","optimizer","ptq","pytorch","qat","quantization","tensorflow"],"latest_commit_sha":null,"homepage":"https://sony.github.io/model_optimization/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/sony.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE.md","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2021-06-21T03:28:09.000Z","updated_at":"2025-04-10T07:31:39.000Z","dependencies_parsed_at":"2023-10-16T06:17:25.393Z","dependency_job_id":"b6177337-4c5c-4e3a-99fd-1df6bc963fb7","html_url":"https://github.com/sony/model_optimization","commit_stats":{"total_commits":1168,"total_committers":32,"mean_commits":36.5,"dds":0.7611301369863014,"last_synced_commit":"3f04be79372f81149fe555f1371f27f0b18d7e82"},"previous_names":[],"tags_count":25,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sony%2Fmodel_optimization","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sony%2Fmodel_optimization/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sony%2Fmodel_optimization/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sony%2Fmodel_optimization/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/sony","download_url":"https://codeload.github.com/sony/model_optimization/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248688544,"owners_count":21145764,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deep-learning","deep-neural-networks","edge-ai","machine-learning","network-compression","network-quantization","neural-network","optimizer","ptq","pytorch","qat","quantization","tensorflow"],"created_at":"2024-11-08T14:28:52.728Z","updated_at":"2025-05-14T16:07:19.464Z","avatar_url":"https://github.com/sony.png","language":"Python","readme":"\u003cdiv align=\"center\" markdown=\"1\"\u003e\n\u003cp\u003e\n      \u003ca href=\"https://sony.github.io/model_optimization/\" target=\"_blank\"\u003e\n        \u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/mctHeader1-cropped.svg\" width=\"1000\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n  \n______________________________________________________________________\n\n\u003c/div\u003e  \n\u003cdiv align=\"center\"\u003e\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"#getting-started\"\u003eGetting Started\u003c/a\u003e •\n  \u003ca href=\"#tutorials-and-examples\"\u003eTutorials\u003c/a\u003e •\n  \u003ca href=\"#high-level-features-and-techniques\"\u003eHigh level features and techniques\u003c/a\u003e •\n  \u003ca href=\"#resources\"\u003eResources\u003c/a\u003e • \n  \u003ca href=\"#contributions\"\u003eCommunity\u003c/a\u003e •\n  \u003ca href=\"#license\"\u003eLicense\u003c/a\u003e\n\u003c/p\u003e\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://sony.github.io/model_optimization#prerequisites\"\u003e\u003cimg src=\"https://img.shields.io/badge/pytorch-2.3%20%7C%202.4%20%7C%202.5%20%7C%202.6-blue\" /\u003e\u003c/a\u003e\n  \u003ca href=\"https://sony.github.io/model_optimization#prerequisites\"\u003e\u003cimg src=\"https://img.shields.io/badge/tensorflow-2.14%20%7C%202.15-blue\" /\u003e\u003c/a\u003e\n  \u003ca href=\"https://sony.github.io/model_optimization#prerequisites\"\u003e\u003cimg src=\"https://img.shields.io/badge/python-3.9%20%7C%203.10%20%7C%203.11%20%7C%203.12-blue\" /\u003e\u003c/a\u003e\n  \u003ca href=\"https://github.com/sony/model_optimization/releases\"\u003e\u003cimg src=\"https://img.shields.io/github/v/release/sony/model_optimization\" /\u003e\u003c/a\u003e\n  \u003ca href=\"https://github.com/sony/model_optimization/blob/main/LICENSE.md\"\u003e\u003cimg src=\"https://img.shields.io/badge/license-Apache%202.0-blue\" /\u003e\u003c/a\u003e\n  \n \u003c/p\u003e    \n\u003c/div\u003e\n\n__________________________________________________________________________________________________________\n\n## \u003cdiv align=\"center\"\u003eGetting Started\u003c/div\u003e\n### Quick Installation\nPip install the model compression toolkit package in a Python\u003e=3.9 environment with PyTorch\u003e=2.3 or Tensorflow\u003e=2.14.\n```\npip install model-compression-toolkit\n```\nFor installing the nightly version or installing from source, refer to the [installation guide](https://github.com/sony/model_optimization/blob/main/INSTALLATION.md).\n\n**Important note**: In order to use MCT, you’ll need to provide a pre-trained floating point model (PyTorch/Keras) as an input.\n\n### Tutorials and Examples \n\nOur [tutorials](https://github.com/sony/model_optimization/blob/main/tutorials/README.md) section will walk you through the basics of the MCT tool, covering various compression techniques for both Keras and PyTorch models. \nAccess interactive notebooks for hands-on learning with popular models/tasks or move on to [Resources](#resources) section.\n\n### Supported Quantization Methods\u003c/div\u003e  \nMCT supports various quantization methods as appears below. \n\u003cdiv align=\"center\"\u003e\n\u003cp align=\"center\"\u003e\n\n  Quantization Method  | Complexity | Computational Cost | API | Tutorial \n-------------------- | -----------|--------------------|---------|--------\nPTQ (Post Training Quantization)  | Low | Low (~1-10 CPU minutes) | [PyTorch API](https://sony.github.io/model_optimization/api/api_docs/methods/pytorch_post_training_quantization.html) / [Keras API](https://sony.github.io/model_optimization/api/api_docs/methods/keras_post_training_quantization.html) | \u003ca href=\"https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_post_training_quantization.ipynb\"\u003e\u003cimg src=\"https://img.shields.io/badge/Pytorch-green\"/\u003e\u003c/a\u003e \u003ca href=\"https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_post-training_quantization.ipynb\"\u003e\u003cimg src=\"https://img.shields.io/badge/Keras-green\"/\u003e\u003c/a\u003e\nGPTQ (parameters fine-tuning using gradients)  | Moderate | Moderate (~1-3 GPU hours) | [PyTorch API](https://sony.github.io/model_optimization/api/api_docs/methods/pytorch_gradient_post_training_quantization.html) / [Keras API](https://sony.github.io/model_optimization/api/api_docs/methods/keras_gradient_post_training_quantization.html) | \u003ca href=\"https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_mobilenet_gptq.ipynb\"\u003e\u003cimg src=\"https://img.shields.io/badge/PyTorch-green\"/\u003e\u003c/a\u003e \u003ca href=\"https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_mobilenet_gptq.ipynb\"\u003e\u003cimg src=\"https://img.shields.io/badge/Keras-green\"/\u003e\u003c/a\u003e \nQAT (Quantization Aware Training)  | High | High (~12-36 GPU hours) | [QAT API](https://sony.github.io/model_optimization/api/api_docs/index.html#qat) | \u003ca href=\"https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_qat.ipynb\"\u003e\u003cimg src=\"https://img.shields.io/badge/Keras-green\"/\u003e\u003c/a\u003e\n\n\u003c/p\u003e    \n\u003c/div\u003e\n\nFor each flow, **Quantization core** utilizes various algorithms and hyper-parameters for optimal [hardware-aware](https://github.com/sony/model_optimization/blob/main/model_compression_toolkit/target_platform_capabilities/README.md) quantization results. \nFor further details, please see [Supported features and algorithms](#high-level-features-and-techniques). \n\n **Required input**: Pre-trained floating point model (PyTorch/Keras)\n \n **Optional input**: Representative dataset - can be either provided by the user, or generated utilizing the [Data Generation](#data-generation-) capability\n\n\u003cdiv align=\"center\"\u003e\n\u003cp align=\"center\"\u003e\n\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/mctDiagram_clean.svg\" width=\"800\"\u003e\n\u003c/p\u003e    \n\u003c/div\u003e\n\n## \u003cdiv align=\"center\"\u003eHigh level features and techniques\u003c/div\u003e\n\nMCT offers a range of powerful features to optimize models for efficient edge deployment. These supported features include:\n\n### Quantization Core Features\n\n🏆 **Mixed-precision search** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_mixed_precision_ptq.ipynb). Assigning optimal quantization bit-width per layer (for weights/activations) \n\n📈 **Graph optimizations**. \nTransforming the model to be best fitted for quantization process.\n\n🔎 **Quantization parameter search** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_activation_threshold_search.ipynb). Minimizing expected quantization-noise during thresholds search using methods such as MSE, No-Clipping and MAE.\n\n🧮 **Advanced quantization algorithms** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_activation_z_score_threshold.ipynb). Enhancing quantization performance for advanced cases is available with some algorithms that can be applied, such as Shift negative correction, Outliers filtering and clustering.\n__________________________________________________________________________________________________________\n### Hardware-aware optimization \n\n🎯 **TPC (Target Platform Capabilities)**. Describes the target hardware’s constrains, for which the model optimization is targeted. See [TPC Readme](./model_compression_toolkit/target_platform_capabilities/README.md) for more information.\n__________________________________________________________________________________________________________\n### Data-free quantization (Data Generation) [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_data_generation.ipynb)\nGenerates synthetic images based on the statistics stored in the model's batch normalization layers, according to your specific needs, for when image data isn’t available. See [Data Generation Library](https://github.com/sony/model_optimization/blob/main/model_compression_toolkit/data_generation/README.md) for more.\nThe specifications of the method are detailed in the paper: _\"**Data Generation for Hardware-Friendly Post-Training Quantization**\"_ [5].\n__________________________________________________________________________________________________________\n### Structured Pruning [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_pruning_mnist.ipynb)\nReduces model size/complexity and ensures better channels utilization by removing redundant input channels from layers and reconstruction of layer weights. Read more ([Pytorch API](https://sony.github.io/model_optimization/api/api_docs/methods/pytorch_pruning_experimental.html) / [Keras API](https://sony.github.io/model_optimization/api/api_docs/methods/keras_pruning_experimental.html)).\n__________________________________________________________________________________________________________\n### **Debugging and Visualization**\n**🎛️ Network Editor (Modify Quantization Configurations)** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/keras/example_keras_network_editor.ipynb). \nModify your model's quantization configuration for specific layers or apply a custom edit rule (e.g adjust layer's bit-width) using MCT’s network editor.\n\n**🖥️ Visualization**. Observe useful information for troubleshooting the quantized model's performance using TensorBoard. [Read more](https://sony.github.io/model_optimization/guidelines/visualization.html).\n\n**🔑 XQuant (Explainable Quantization)** [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sony/model_optimization/blob/main/tutorials/notebooks/mct_features_notebooks/pytorch/example_pytorch_xquant.ipynb). Get valuable insights regarding the quality and success of the quantization process of your model. The report includes histograms and similarity metrics between the original float model and the quantized model in key points of the model. The report can be visualized using TensorBoard.\n__________________________________________________________________________________________________________\n### Enhanced Post-Training Quantization (EPTQ)\nAs part of the GPTQ capability, we provide an advanced optimization algorithm called EPTQ.\nThe specifications of the algorithm are detailed in the paper: _\"**EPTQ: Enhanced Post-Training Quantization via Hessian-guided Network-wise Optimization**\"_ [4].\nMore details on how to use EPTQ via MCT can be found in the [GPTQ guidelines](https://github.com/sony/model_optimization/blob/main/model_compression_toolkit/gptq/README.md).\n\n## \u003cdiv align=\"center\"\u003eResources\u003c/div\u003e\n* [User Guide](https://sony.github.io/model_optimization/index.html)  contains detailed information about MCT and guides you from installation through optimizing models for your edge AI applications.\n\n* MCT's [API Docs](https://sony.github.io/model_optimization/api/api_docs/) is separated per quantization methods:  \n\n  * [Post-training quantization](https://sony.github.io/model_optimization/api/api_docs/index.html#ptq) | PTQ API docs\n  * [Gradient-based post-training quantization](https://sony.github.io/model_optimization/api/api_docs/index.html#gptq) | GPTQ API docs\n  * [Quantization-aware training](https://sony.github.io/model_optimization/api/api_docs/index.html#qat) | QAT API docs\n    \n* [Debug](https://sony.github.io/model_optimization/guidelines/visualization.html) – modify optimization process or generate an explainable report\n  \n* [Release notes](https://github.com/sony/model_optimization/releases)\n\n\n## \u003cdiv align=\"center\"\u003eSupported Versions\u003c/div\u003e\n\nCurrently, MCT is being tested on various Python, Pytorch and TensorFlow versions:\n\u003cdetails id=\"supported-versions\"\u003e\n  \u003csummary\u003eSupported Versions Table\u003c/summary\u003e\n\n|                                                                                                                                                                                                                   | PyTorch 2.3                                                                                                                                                                                                              | PyTorch 2.4                                                                                                                                                                                                              | PyTorch 2.5                                                                                                                                                                                                              | PyTorch 2.6                                                                                                                                                                                                              |\n|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| Python 3.9  | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch23.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch23.yml)   | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch24.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch24.yml)   | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch25.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch25.yml)   | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch26.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_pytorch26.yml)   |\n| Python 3.10 | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch23.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch23.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch24.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch24.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch25.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch25.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch26.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_pytorch26.yml) |\n| Python 3.11 | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch23.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch23.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch24.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch24.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch25.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch25.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch26.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_pytorch26.yml) |\n| Python 3.12 | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch23.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch23.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch24.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch24.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch25.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch25.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch26.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python312_pytorch26.yml) |\n\n|             | TensorFlow 2.14                                                                                                                                                                                                        | TensorFlow 2.15                                                                                                                                                                                                        |\n|-------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| Python 3.9  | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_keras214.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_keras214.yml)   | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_keras215.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python39_keras215.yml)   |\n| Python 3.10 | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_keras214.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_keras214.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_keras215.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python310_keras215.yml) |\n| Python 3.11 | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_keras214.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_keras214.yml) | [![Run Tests](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_keras215.yml/badge.svg)](https://github.com/sony/model_optimization/actions/workflows/run_tests_python311_keras215.yml) |\n\n\u003c/details\u003e\n\n## \u003cdiv align=\"center\"\u003eResults\u003c/div\u003e\n\n\u003cp align=\"center\"\u003e\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/Classification.png\" width=\"200\"\u003e\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/SemSeg.png\" width=\"200\"\u003e\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/PoseEst.png\" width=\"200\"\u003e\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/ObjDet.png\" width=\"200\"\u003e\n\nMCT can quantize an existing 32-bit floating-point model to an 8-bit fixed-point (or less) model without compromising accuracy. \nBelow is a graph of [MobileNetV2](https://pytorch.org/vision/main/models/generated/torchvision.models.mobilenet_v2.html) accuracy on ImageNet vs average bit-width of weights (X-axis), using **single-precision** quantization, **mixed-precision** quantization, and mixed-precision quantization with GPTQ. \n\n\u003cp align=\"center\"\u003e\n\u003cimg src=\"https://raw.githubusercontent.com/sony/model_optimization/refs/heads/main/docsrc/images/torch_mobilenetv2.png\" width=\"800\"\u003e\n\nFor more results, please see [1]\n\n\n### Pruning Results\n\nResults for applying pruning to reduce the parameters of the following models by 50%:\n\n| Model           | Dense Model Accuracy | Pruned Model Accuracy |\n|-----------------|----------------------|-----------------------|\n| ResNet50 [2]    | 75.1                 | 72.4                  |\n| DenseNet121 [3] | 74.44                | 71.71                 |\n\n## \u003cdiv align=\"center\"\u003eTroubleshooting and Community\u003c/div\u003e\n\nIf you encountered a large accuracy degradation with MCT, check out the [Quantization Troubleshooting](https://github.com/sony/model_optimization/tree/main/quantization_troubleshooting.md)\nfor common pitfalls and some tools to improve the quantized model's accuracy.\n\nCheck out the [FAQ](https://github.com/sony/model_optimization/tree/main/FAQ.md) for common issues. \n\nYou are welcome to ask questions and get support on our [issues section](https://github.com/sony/model_optimization/issues) and manage community discussions under the [discussions section](https://github.com/sony/model_optimization/discussions).\n\n\n## \u003cdiv align=\"center\"\u003eContributions\u003c/div\u003e\nWe'd love your input! MCT would not be possible without help from our community, and welcomes contributions from anyone! \n\n*Checkout our [Contribution guide](https://github.com/sony/model_optimization/blob/main/CONTRIBUTING.md) for more details.\n\nThank you 🙏 to all our contributors!\n\n## \u003cdiv align=\"center\"\u003eLicense\u003c/div\u003e\nMCT is licensed under Apache License Version 2.0. By contributing to the project, you agree to the license and copyright terms therein and release your contribution under these terms.\n\n\u003ca href=\"https://github.com/sony/model_optimization/blob/main/LICENSE.md\"\u003e\u003cimg src=\"https://img.shields.io/badge/license-Apache%202.0-blue\" /\u003e\u003c/a\u003e\n\n## \u003cdiv align=\"center\"\u003eReferences\u003c/div\u003e\n\n[1] Habi, H.V., Peretz, R., Cohen, E., Dikstein, L., Dror, O., Diamant, I., Jennings, R.H. and Netzer, A., 2021. [HPTQ: Hardware-Friendly Post Training Quantization. arXiv preprint](https://arxiv.org/abs/2109.09113).\n\n[2] [Keras Applications](https://keras.io/api/applications/)\n\n[3] [TORCHVISION.MODELS](https://pytorch.org/vision/stable/models.html) \n\n[4] Gordon, O., Cohen, E., Habi, H. V., \u0026 Netzer, A., 2024. [EPTQ: Enhanced Post-Training Quantization via Hessian-guided Network-wise Optimization, European Conference on Computer Vision Workshop 2024, Computational Aspects of Deep Learning (CADL)](https://arxiv.org/abs/2309.11531)\n\n[5] Dikstein, L., Lapid, A., Netzer, A., \u0026 Habi, H. V., 2024. [Data Generation for Hardware-Friendly Post-Training Quantization, Accepted to IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025](https://openaccess.thecvf.com/content/WACV2025/papers/Dikstein_Data_Generation_for_Hardware-Friendly_Post-Training_Quantization_WACV_2025_paper.pdf)\n","funding_links":[],"categories":["Networks"],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsony%2Fmodel_optimization","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsony%2Fmodel_optimization","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsony%2Fmodel_optimization/lists"}