| Name | mct-nightly JSON |
| Version |
1.4.0.31052022.post408
JSON |
| download |
| home_page | |
| Summary | A Model Compression Toolkit for neural networks |
| upload_time | 2022-05-31 00:04:11 |
| maintainer | |
| docs_url | None |
| author | |
| requires_python | >=3.6 |
| license | |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# Model Compression Toolkit (MCT)
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 tools for optimizing and deploying state-of-the-art neural networks on efficient hardware. Specifically, this project aims to apply quantization and pruning schemes to compress neural networks.
<img src="MCT_Block_Diagram.svg" width="800">
Currently, this project supports hardware-friendly post-training quantization (HPTQ) with Tensorflow 2 and Pytorch [1].
The MCT project is developed by researchers and engineers working at Sony Semiconductors Israel.
For more information, please visit our [project website](https://sony.github.io/model_optimization/).
## Table of Contents
- [Getting Started](#getting-started)
- [Supported features](#supported-features)
- [Results](#results)
- [Contributions](#contributions)
- [License](#license)
## Getting Started
This section provides a quick starting guide. We begin with installation via source code or pip server. Then, we provide a short usage example.
### Installation
See the MCT install guide for the pip package, and build from the source.
#### From Source
```
git clone https://github.com/sony/model_optimization.git
python setup.py install
```
#### From PyPi - latest stable release
```
pip install model-compression-toolkit
```
A nightly package is also available (unstable):
```
pip install mct-nightly
```
To run MCT, one of the supported frameworks, Tenosflow/Pytorch, needs to be installed.
For using with Tensorflow please install the packages:
[tensorflow](https://www.tensorflow.org/install),
[tensorflow-model-optimization](https://www.tensorflow.org/model_optimization/guide/install)
For using with Pytorch (experimental) please install the packages:
[torch](https://pytorch.org/)
MCT is tested with:
* Tensorflow version 2.7
* Pytorch version 1.10.0
### Usage Example
For an example of how to use the post-training quantization, using Keras,
please use this [link](tutorials/example_keras_mobilenet.py).
For an example using Pytorch (experimental), please use this [link](tutorials/example_pytorch_mobilenet_v2.py).
For more examples please see the [tutorials' directory](tutorials).
## Supported Features
Quantization:
* Post Training Quantization for Keras models.
* Post Training Quantization for Pytorch models (experimental).
* Gradient-based post-training (Experimental, Keras only).
* Mixed-precision post-training quantization (Experimental).
Tensorboard Visualization (Experimental):
* CS Analyzer: compare a model compressed with the original model to analyze large accuracy drops.
* Activation statistics and errors
## Results
### Keras
As part of the MCT library, we have a set of example networks on image classification. These networks can be used as examples when using the package.
* Image Classification Example with MobileNet V1 on ImageNet dataset
| Network Name | Float Accuracy | 8Bit Accuracy | Comments |
| -------------------------| ---------------:| ---------------:| ------------------------------------:|
| MobileNetV1 [2] | 70.558 | 70.418 | |
For more results please see [1]
### Pytorch
We quantized classification networks from the torchvision library.
In the following table we present the ImageNet validation results for these models:
| Network Name | Float Accuracy | 8Bit Accuracy |
| --------------------------| ---------------:| ---------------:|
| MobileNet V2 [3] | 71.886 | 71.444 |
| ResNet-18 [3] | 69.86 | 69.63 |
| SqueezeNet 1.1 [3] | 58.128 | 57.678 |
## Contributions
MCT aims at keeping a more up-to-date fork and welcomes contributions from anyone.
*You will find more information about contributions in the [Contribution guide](CONTRIBUTING.md).
## License
[Apache License 2.0](LICENSE).
## References
[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).
[2] [MobilNet](https://keras.io/api/applications/mobilenet/#mobilenet-function) from Keras applications.
[3] [TORCHVISION.MODELS](https://pytorch.org/vision/stable/models.html)
Raw data
{
"_id": null,
"home_page": "",
"name": "mct-nightly",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": "",
"keywords": "",
"author": "",
"author_email": "",
"download_url": "https://files.pythonhosted.org/packages/af/f8/c282ae419e6185466ce337aadaae5685a6abd051c962ac99603ae8a13d84/mct-nightly-1.4.0.31052022.post408.tar.gz",
"platform": null,
"description": "# Model Compression Toolkit (MCT)\n\n\nModel Compression Toolkit (MCT) is an open-source project for neural network model optimization under efficient, constrained hardware. This project provides researchers, developers, and engineers tools for optimizing and deploying state-of-the-art neural networks on efficient hardware. Specifically, this project aims to apply quantization and pruning schemes to compress neural networks. \n<img src=\"MCT_Block_Diagram.svg\" width=\"800\">\n\nCurrently, this project supports hardware-friendly post-training quantization (HPTQ) with Tensorflow 2 and Pytorch [1]. \n\nThe MCT project is developed by researchers and engineers working at Sony Semiconductors Israel.\n\nFor more information, please visit our [project website](https://sony.github.io/model_optimization/).\n\n## Table of Contents\n\n- [Getting Started](#getting-started)\n- [Supported features](#supported-features)\n- [Results](#results)\n- [Contributions](#contributions)\n- [License](#license)\n\n## Getting Started\n\nThis section provides a quick starting guide. We begin with installation via source code or pip server. Then, we provide a short usage example.\n\n### Installation\nSee the MCT install guide for the pip package, and build from the source.\n\n\n#### From Source\n```\ngit clone https://github.com/sony/model_optimization.git\npython setup.py install\n```\n#### From PyPi - latest stable release\n```\npip install model-compression-toolkit\n```\n\nA nightly package is also available (unstable):\n```\npip install mct-nightly\n```\n\nTo run MCT, one of the supported frameworks, Tenosflow/Pytorch, needs to be installed.\n\nFor using with Tensorflow please install the packages: \n[tensorflow](https://www.tensorflow.org/install), \n[tensorflow-model-optimization](https://www.tensorflow.org/model_optimization/guide/install)\n\nFor using with Pytorch (experimental) please install the packages: \n[torch](https://pytorch.org/)\n\nMCT is tested with:\n* Tensorflow version 2.7 \n* Pytorch version 1.10.0 \n\n### Usage Example \nFor an example of how to use the post-training quantization, using Keras,\nplease use this [link](tutorials/example_keras_mobilenet.py).\n\nFor an example using Pytorch (experimental), please use this [link](tutorials/example_pytorch_mobilenet_v2.py).\n\nFor more examples please see the [tutorials' directory](tutorials).\n\n\n## Supported Features\n\nQuantization:\n\n * Post Training Quantization for Keras models.\n * Post Training Quantization for Pytorch models (experimental).\n * Gradient-based post-training (Experimental, Keras only).\n * Mixed-precision post-training quantization (Experimental).\n\nTensorboard Visualization (Experimental):\n\n * CS Analyzer: compare a model compressed with the original model to analyze large accuracy drops.\n * Activation statistics and errors\n\n\n## Results\n### Keras\nAs part of the MCT library, we have a set of example networks on image classification. These networks can be used as examples when using the package.\n\n* Image Classification Example with MobileNet V1 on ImageNet dataset\n\n| Network Name | Float Accuracy | 8Bit Accuracy | Comments |\n| -------------------------| ---------------:| ---------------:| ------------------------------------:|\n| MobileNetV1 [2] | 70.558 | 70.418 | |\n\n\nFor more results please see [1]\n\n### Pytorch\nWe quantized classification networks from the torchvision library. \nIn the following table we present the ImageNet validation results for these models:\n\n| Network Name | Float Accuracy | 8Bit Accuracy | \n| --------------------------| ---------------:| ---------------:| \n| MobileNet V2 [3] | 71.886 | 71.444 | \n| ResNet-18 [3] | 69.86 | 69.63 | \n| SqueezeNet 1.1 [3] | 58.128 | 57.678 | \n\n\n\n## Contributions\nMCT aims at keeping a more up-to-date fork and welcomes contributions from anyone.\n\n*You will find more information about contributions in the [Contribution guide](CONTRIBUTING.md).\n\n\n## License\n[Apache License 2.0](LICENSE).\n\n## References \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] [MobilNet](https://keras.io/api/applications/mobilenet/#mobilenet-function) from Keras applications.\n\n[3] [TORCHVISION.MODELS](https://pytorch.org/vision/stable/models.html) \n\n\n",
"bugtrack_url": null,
"license": "",
"summary": "A Model Compression Toolkit for neural networks",
"version": "1.4.0.31052022.post408",
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"md5": "37643e29f8abf78513426712c41c8ec1",
"sha256": "99f3d99f2c2fac0e74d685c058813e840c0f3f84d6b698e2faa4d3023e72dbbc"
},
"downloads": -1,
"filename": "mct_nightly-1.4.0.31052022.post408-py3-none-any.whl",
"has_sig": false,
"md5_digest": "37643e29f8abf78513426712c41c8ec1",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.6",
"size": 399836,
"upload_time": "2022-05-31T00:04:09",
"upload_time_iso_8601": "2022-05-31T00:04:09.934319Z",
"url": "https://files.pythonhosted.org/packages/f2/16/cdf070add884a1c1da4c02d7d3f392f61c293040ed3c5d42b05e85911a0b/mct_nightly-1.4.0.31052022.post408-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"md5": "1a260e42ed974abea9376e42ade7aa81",
"sha256": "3852b791627ea5a01ccdbd348aae99ca4564306e6871dc4e4de244fa7fe9e21b"
},
"downloads": -1,
"filename": "mct-nightly-1.4.0.31052022.post408.tar.gz",
"has_sig": false,
"md5_digest": "1a260e42ed974abea9376e42ade7aa81",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 218560,
"upload_time": "2022-05-31T00:04:11",
"upload_time_iso_8601": "2022-05-31T00:04:11.496472Z",
"url": "https://files.pythonhosted.org/packages/af/f8/c282ae419e6185466ce337aadaae5685a6abd051c962ac99603ae8a13d84/mct-nightly-1.4.0.31052022.post408.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2022-05-31 00:04:11",
"github": false,
"gitlab": false,
"bitbucket": false,
"lcname": "mct-nightly"
}
