https://github.com/musco-ai/musco-tf

MUSCO: Multi-Stage COmpression of neural networks
https://github.com/musco-ai/musco-tf

cnn-acceleration cnn-compresion cp-decomposition deep-neural-networks low-rank-approximation musco network-compression tensor-decomposition tensorflow truncated-svd tucker vbmf

Last synced: 4 months ago
JSON representation

MUSCO: Multi-Stage COmpression of neural networks

• Host: GitHub
• URL: https://github.com/musco-ai/musco-tf
• Owner: musco-ai
• License: apache-2.0
• Created: 2019-10-01T11:02:44.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2021-01-07T21:34:33.000Z (about 4 years ago)
• Last Synced: 2024-10-03T09:37:04.891Z (4 months ago)
• Topics: cnn-acceleration, cnn-compresion, cp-decomposition, deep-neural-networks, low-rank-approximation, musco, network-compression, tensor-decomposition, tensorflow, truncated-svd, tucker, vbmf
• Language: Python
• Homepage:
• Size: 85.9 KB
• Stars: 8
• Watchers: 3
• Forks: 2
• Open Issues: 5
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# ![](https://user-images.githubusercontent.com/11778655/66068156-bef1a880-e555-11e9-8d26-094071133a11.png) MUSCO: Multi-Stage COmpression of neural networks

This repository contains supplementary code for the paper [MUSCO: Multi-Stage COmpression of neural networks](https://arxiv.org/pdf/1903.09973.pdf). 

It demonstrates how a neural network with convolutional and fully connected layers can be compressed using iterative tensor decomposition of weight tensors.

## Requirements

```

numpy

scipy

scikit-tensor-py3

tensorly-musco

absl-py

tqdm

tensorflow-gpu (TensorRT support)

```

## Installation

```

pip install musco-tf

```

## Quick Start

```python

from musco.tf import CompressorVBMF, Optimizer

model = load_model("model.h5")

compressor = CompressorVBMF(model)

while True:

    model = compressor.compress_iteration(number=5)

    

    # Fine-tune compressed model.

# Compressor decomposes 5 layers on each iteration

# and returns compressed model. You have to fine-tune

# model after each iteration to restore accuracy.

# Compressor automatically selects the best parameters

# for decomposition on each iteration.

# You can freeze and quantize model after compression.

optimizer = Optimizer(precision="FP16", max_batch_size=16)

optimizer.freeze(model)

optimizer.optimize("frozen.pb")

```

## Citing

If you used our research, we kindly ask you to cite the corresponding [paper](https://arxiv.org/abs/1903.09973).

```

@article{gusak2019one,

  title={MUSCO: Multi-Stage Compression of neural networks},

  author={Gusak, Julia and Kholiavchenko, Maksym and Ponomarev, Evgeny and Markeeva, Larisa and Oseledets, Ivan and Cichocki, Andrzej},

  journal={arXiv preprint arXiv:1903.09973},

  year={2019}

}

```

## License

Project is distributed under [Apache License 2.0](https://github.com/musco-ai/musco-tf/blob/master/LICENSE).