4b3f8ab69e
* Refactor ModelEstimator: 1. add EvaluatorLoop in engine.runners; 2. add estimator for structures (both subnet & supernet); 3. add layer_counter for each op. * fix lint * update estimator: 1. add ResourceEstimator based on BaseEstimator; 2. add notes & examples for ResourceEstimator & EvaluatorLoop usage; 3. fix a bug of latency test. 4. minor changes according to comments. * add UT & fix a bug caused by UT * add docstrings & remove old estimator * update docstrings for op_spec_counters * rename resource_evaluator_val_loop * support adding resource attrs of each submodule in a measured model * fix lint * refactor estimator file structures * support estimating resources for spec modules * rm old UT * update new estimator UT cases * fix traversal range of the model * cancel unit convert in accumulate_sub_module_flops_params * use estimator_cfg to build ResourceEstimator * fix a broadcast bug * delete fixed input_shape * add assertion and string-format-return when measuring spec_modules * add UT for estimating spec_modules |
||
|---|---|---|
| .dev_scripts | ||
| .github | ||
| configs | ||
| docker | ||
| docs | ||
| mmrazor | ||
| requirements | ||
| resources | ||
| tests | ||
| tools | ||
| .gitignore | ||
| .pre-commit-config.yaml | ||
| .readthedocs.yml | ||
| LICENSE | ||
| MANIFEST.in | ||
| README.md | ||
| README_zh-CN.md | ||
| model-index.yml | ||
| requirements.txt | ||
| setup.cfg | ||
| setup.py | ||
README.md
📘Documentation | 🛠️Installation | 👀Model Zoo | 🤔Reporting Issues
English | 简体中文
Introduction
MMRazor is a model compression toolkit for model slimming and AutoML, which includes 3 mainstream technologies:
- Neural Architecture Search (NAS)
- Pruning
- Knowledge Distillation (KD)
- Quantization (in the next release)
It is a part of the OpenMMLab project.
Major features:
-
Compatibility
MMRazor can be easily applied to various projects in OpenMMLab, due to the similar architecture design of OpenMMLab as well as the decoupling of slimming algorithms and vision tasks.
-
Flexibility
Different algorithms, e.g., NAS, pruning and KD, can be incorporated in a plug-n-play manner to build a more powerful system.
-
Convenience
With better modular design, developers can implement new model compression algorithms with only a few codes, or even by simply modifying config files.
Below is an overview of MMRazor's design and implementation, please refer to tutorials for more details.
What's new
MMRazor v0.3.1 was released in 5/4/2022.
Benchmark and model zoo
Results and models are available in the model zoo.
Supported algorithms:
Neural Architecture Search
Pruning
Knowledge Distillation
Installation
MMRazor depends on PyTorch and MMCV.
Please refer to get_started.md for more detailed instruction.
Getting Started
Please refer to train.md and test.md for the basic usage of MMRazor. There are also tutorials:
- overview
- learn about configs
- customize architectures
- customize nas algorithms
- customize pruning algorithms
- customize kd algorithms
- customize mixed algorithms with our algorithm_components
- apply existing algorithms to other existing tasks
Contributing
We appreciate all contributions to improve MMRazor. Please refer to CONTRUBUTING.md for the contributing guideline.
Acknowledgement
MMRazor is an open source project that is contributed by researchers and engineers from various colleges and companies. We appreciate all the contributors who implement their methods or add new features, as well as users who give valuable feedbacks. We wish that the toolbox and benchmark could serve the growing research community by providing a flexible toolkit to reimplement existing methods and develop their own new model compression methods.
Citation
If you find this project useful in your research, please consider cite:
@misc{2021mmrazor,
title={OpenMMLab Model Compression Toolbox and Benchmark},
author={MMRazor Contributors},
howpublished = {\url{https://github.com/open-mmlab/mmrazor}},
year={2021}
}
License
This project is released under the Apache 2.0 license.
Projects in OpenMMLab
- MMCV: OpenMMLab foundational library for computer vision.
- MIM: MIM installs OpenMMLab packages.
- MMClassification: OpenMMLab image classification toolbox and benchmark.
- MMDetection: OpenMMLab detection toolbox and benchmark.
- MMDetection3D: OpenMMLab's next-generation platform for general 3D object detection.
- MMRotate: OpenMMLab rotated object detection toolbox and benchmark.
- MMSegmentation: OpenMMLab semantic segmentation toolbox and benchmark.
- MMOCR: OpenMMLab text detection, recognition, and understanding toolbox.
- MMPose: OpenMMLab pose estimation toolbox and benchmark.
- MMHuman3D: OpenMMLab 3D human parametric model toolbox and benchmark.
- MMSelfSup: OpenMMLab self-supervised learning toolbox and benchmark.
- MMRazor: OpenMMLab model compression toolbox and benchmark.
- MMFewShot: OpenMMLab fewshot learning toolbox and benchmark.
- MMAction2: OpenMMLab's next-generation action understanding toolbox and benchmark.
- MMTracking: OpenMMLab video perception toolbox and benchmark.
- MMFlow: OpenMMLab optical flow toolbox and benchmark.
- MMEditing: OpenMMLab image and video editing toolbox.
- MMGeneration: OpenMMLab image and video generative models toolbox.
- MMDeploy: OpenMMLab model deployment framework.