PyTorch implementation for "Sharpness-aware Quantization for Deep Neural Networks".

Related tags

Deep LearningSAQ
Overview

Sharpness-aware Quantization for Deep Neural Networks

License

Recent Update

2021.11.23: We release the source code of SAQ.

Setup the environments

  1. Clone the repository locally:
git clone https://github.com/zhuang-group/SAQ
  1. Install pytorch 1.8+, tensorboard and prettytable
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch
pip install tensorboard
pip install prettytable

Data preparation

ImageNet

  1. Download the ImageNet 2012 dataset from here, and prepare the dataset based on this script.

  2. Change the dataset path in link_imagenet.py and link the ImageNet-100 by

python link_imagenet.py

CIFAR-100

Download the CIFAR-100 dataset from here.

After downloading ImageNet and CIFAR-100, the file structure should look like:

dataset
├── imagenet
    ├── train
    │   ├── class1
    │   │   ├── img1.jpeg
    │   │   ├── img2.jpeg
    │   │   └── ...
    │   ├── class2
    │   │   ├── img3.jpeg
    │   │   └── ...
    │   └── ...
    └── val
        ├── class1
        │   ├── img4.jpeg
        │   ├── img5.jpeg
        │   └── ...
        ├── class2
        │   ├── img6.jpeg
        │   └── ...
        └── ...
├── cifar100
    ├── cifar-100-python
    │   ├── meta
    │   ├── test
    │   ├── train
    │   └── ...
    └── ...

Training

Fixed-precision quantization

  1. Download the pre-trained full-precision models from the model zoo.

  2. Train low-precision models.

To train low-precision ResNet-20 on CIFAR-100, run:

sh script/train_qsam_cifar_r20.sh

To train low-precision ResNet-18 on ImageNet, run:

sh script/train_qsam_imagenet_r18.sh

Mixed-precision quantization

  1. Download the pre-trained full-precision models from the model zoo.

  2. Train the configuration generator.

To train the configuration generator of ResNet-20 on CIFAR-100, run:

sh script/train_generator_cifar_r20.sh

To train the configuration generator on ImageNet, run:

sh script/train_generator_imagenet_r18.sh
  1. After training the configuration generator, run following commands to fine-tune the resulting models with the obtained bitwidth configurations on CIFAR-100 and ImageNet.
sh script/finetune_cifar_r20.sh
sh script/finetune_imagenet_r18.sh

Results on CIFAR-100

Network Method Bitwidth BOPs (M) Top-1 Acc. (%) Top-5 Acc. (%)
ResNet-20 SAQ 4 674.6 68.7 91.2
ResNet-20 SAMQ MP 659.3 68.7 91.2
ResNet-20 SAQ 3 392.1 67.7 90.8
ResNet-20 SAMQ MP 374.4 68.6 91.2
MobileNetV2 SAQ 4 1508.9 75.6 93.7
MobileNetV2 SAMQ MP 1482.1 75.5 93.6
MobileNetV2 SAQ 3 877.1 74.4 93.2
MobileNetV2 SAMQ MP 869.5 75.5 93.7

Results on ImageNet

Network Method Bitwidth BOPs (G) Top-1 Acc. (%) Top-5 Acc. (%)
ResNet-18 SAQ 4 34.7 71.3 90.0
ResNet-18 SAMQ MP 33.7 71.4 89.9
ResNet-18 SAQ 2 14.4 67.1 87.3
MobileNetV2 SAQ 4 5.3 70.2 89.4
MobileNetV2 SAMQ MP 5.3 70.3 89.4

License

This repository is released under the Apache 2.0 license as found in the LICENSE file.

Acknowledgement

This repository has adopted codes from SAM, ASAM and ESAM, we thank the authors for their open-sourced code.

You might also like...
Objective of the repository is to learn and build machine learning models using Pytorch. 30DaysofML Using Pytorch
Objective of the repository is to learn and build machine learning models using Pytorch. 30DaysofML Using Pytorch

30 Days Of Machine Learning Using Pytorch Objective of the repository is to learn and build machine learning models using Pytorch. List of Algorithms

Pretrained SOTA Deep Learning models, callbacks and more for research and production with PyTorch Lightning and PyTorch
Pretrained SOTA Deep Learning models, callbacks and more for research and production with PyTorch Lightning and PyTorch

Pretrained SOTA Deep Learning models, callbacks and more for research and production with PyTorch Lightning and PyTorch

Amazon Forest Computer Vision: Satellite Image tagging code using PyTorch / Keras with lots of PyTorch tricks
Amazon Forest Computer Vision: Satellite Image tagging code using PyTorch / Keras with lots of PyTorch tricks

Amazon Forest Computer Vision Satellite Image tagging code using PyTorch / Keras Here is a sample of images we had to work with Source: https://www.ka

The Incredible PyTorch: a curated list of tutorials, papers, projects, communities and more relating to PyTorch.
The Incredible PyTorch: a curated list of tutorials, papers, projects, communities and more relating to PyTorch.

This is a curated list of tutorials, projects, libraries, videos, papers, books and anything related to the incredible PyTorch. Feel free to make a pu

Amazon Forest Computer Vision: Satellite Image tagging code using PyTorch / Keras with lots of PyTorch tricks
Amazon Forest Computer Vision: Satellite Image tagging code using PyTorch / Keras with lots of PyTorch tricks

Amazon Forest Computer Vision Satellite Image tagging code using PyTorch / Keras Here is a sample of images we had to work with Source: https://www.ka

A bunch of random PyTorch models using PyTorch's C++ frontend
A bunch of random PyTorch models using PyTorch's C++ frontend

PyTorch Deep Learning Models using the C++ frontend Gettting started Clone the repo 1. https://github.com/mrdvince/pytorchcpp 2. cd fashionmnist or

PyTorch Autoencoders - Implementing a Variational Autoencoder (VAE) Series in Pytorch.

PyTorch Autoencoders Implementing a Variational Autoencoder (VAE) Series in Pytorch. Inspired by this repository Model List check model paper conferen

PyTorch-LIT is the Lite Inference Toolkit (LIT) for PyTorch which focuses on easy and fast inference of large models on end-devices.

PyTorch-LIT PyTorch-LIT is the Lite Inference Toolkit (LIT) for PyTorch which focuses on easy and fast inference of large models on end-devices. With

A general framework for deep learning experiments under PyTorch based on pytorch-lightning

torchx Torchx is a general framework for deep learning experiments under PyTorch based on pytorch-lightning. TODO list gan-like training wrapper text

Comments
  • Quantize_first_last_layer

    Quantize_first_last_layer

    Hi! I noticed that in your code, you set bits_weights=8 and bits_activations=32 for first layer as default, it's not what is claimed in your paper " For the first and last layers of all quantized models, we quantize both weights and activations to 8-bit. " And I see an accuracy drop if I adjust the bits_activations to 8 for the first layer, could u please explain what is the reason? Thanks!

    opened by mmmiiinnnggg 0
  • 代码问题请求帮助

    代码问题请求帮助

    你好,带佬的代码写的很好,有部分代码不太懂,想请教一下, parser.add_argument( "--arch_bits", type=lambda s: [float(item) for item in s.split(",")] if len(s) != 0 else "", default=" ", help="bits configuration of each layer",

    if len(args.arch_bits) != 0: if args.wa_same_bit: set_wae_bits(model, args.arch_bits) elif args.search_w_bit: set_w_bits(model, args.arch_bits) else: set_bits(model, args.arch_bits) show_bits(model) logger.info("Set arch bits to: {}".format(args.arch_bits)) logger.info(model) 这个arch_bits主要是做什么的呢,卡在这里有段时间了

    opened by LKAMING97 0
Releases(v0.1.1)
Owner
Zhuang AI Group
Zhuang AI Group
Implementation of the paper "Self-Promoted Prototype Refinement for Few-Shot Class-Incremental Learning"

Self-Promoted Prototype Refinement for Few-Shot Class-Incremental Learning This is the implementation of the paper "Self-Promoted Prototype Refinement

Kai Zhu 78 Dec 02, 2022
Joint learning of images and text via maximization of mutual information

mutual_info_img_txt Joint learning of images and text via maximization of mutual information. This repository incorporates the algorithms presented in

Ruizhi Liao 10 Dec 22, 2022
Example scripts for the detection of lanes using the ultra fast lane detection model in Tensorflow Lite.

TFlite Ultra Fast Lane Detection Inference Example scripts for the detection of lanes using the ultra fast lane detection model in Tensorflow Lite. So

Ibai Gorordo 12 Aug 27, 2022
Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20. model in ONNX

ONNX msg_chn_wacv20 depth completion Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20 model in

Ibai Gorordo 19 Oct 22, 2022
git《Beta R-CNN: Looking into Pedestrian Detection from Another Perspective》(NeurIPS 2020) GitHub:[fig3]

Beta R-CNN: Looking into Pedestrian Detection from Another Perspective This is the pytorch implementation of our paper "[Beta R-CNN: Looking into Pede

35 Sep 08, 2021
Group project for MFIN7036. Our goal is to predict firm profitability with text-based competition measures.

NLP_0-project Group project for MFIN7036. Our goal is to predict firm profitability with text-based competition measures1. We are a "democratic" and c

3 Mar 16, 2022
Group Fisher Pruning for Practical Network Compression(ICML2021)

Group Fisher Pruning for Practical Network Compression (ICML2021) By Liyang Liu*, Shilong Zhang*, Zhanghui Kuang, Jing-Hao Xue, Aojun Zhou, Xinjiang W

Shilong Zhang 129 Dec 13, 2022
[BMVC2021] The official implementation of "DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations"

DomainMix [BMVC2021] The official implementation of "DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations" [paper] [de

Wenhao Wang 17 Dec 20, 2022
Compressed Video Action Recognition

Compressed Video Action Recognition Chao-Yuan Wu, Manzil Zaheer, Hexiang Hu, R. Manmatha, Alexander J. Smola, Philipp Krähenbühl. In CVPR, 2018. [Proj

Chao-Yuan Wu 479 Dec 26, 2022
[ICCV 2021] Deep Hough Voting for Robust Global Registration

Deep Hough Voting for Robust Global Registration, ICCV, 2021 Project Page | Paper | Video Deep Hough Voting for Robust Global Registration Junha Lee1,

57 Nov 28, 2022
Resources complimenting the Machine Learning Course led in the Faculty of mathematics and informatics part of Sofia University.

Machine Learning and Data Mining, Summer 2021-2022 How to learn data science and machine learning? Programming. Learn Python. Basic Statistics. Take a

Simeon Hristov 8 Oct 04, 2022
The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Training

[ICLR 2022] The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Training The Unreasonable Effectiveness of

VITA 44 Dec 23, 2022
Intel® Nervana™ reference deep learning framework committed to best performance on all hardware

DISCONTINUATION OF PROJECT. This project will no longer be maintained by Intel. Intel will not provide or guarantee development of or support for this

Nervana 3.9k Dec 20, 2022
[CVPR 2021] Anycost GANs for Interactive Image Synthesis and Editing

Anycost GAN video | paper | website Anycost GANs for Interactive Image Synthesis and Editing Ji Lin, Richard Zhang, Frieder Ganz, Song Han, Jun-Yan Zh

MIT HAN Lab 726 Dec 28, 2022
Reinforcement learning algorithms in RLlib

raylab Reinforcement learning algorithms in RLlib and PyTorch. Installation pip install raylab Quickstart Raylab provides agents and environments to b

Ângelo 50 Sep 08, 2022
TalkingHead-1KH is a talking-head dataset consisting of YouTube videos

TalkingHead-1KH Dataset TalkingHead-1KH is a talking-head dataset consisting of YouTube videos, originally created as a benchmark for face-vid2vid: On

173 Dec 29, 2022
✨风纪委员会自动投票脚本,利用Github Action帮你进行裁决操作(为了让其他风纪委员有案件可判,本程序从中午12点才开始运行,有需要请自己修改运行时间)

风纪委员会自动投票 本脚本通过使用Github Action来实现B站风纪委员的自动投票功能,喜欢请给我点个STAR吧! 如果你不是风纪委员,在符合风纪委员申请条件的情况下,本脚本会自动帮你申请 投票时间是早上八点,如果有需要请自行修改.github/workflows/Judge.yml中的时间,

Pesy Wu 25 Feb 17, 2021
TinyML Cookbook, published by Packt

TinyML Cookbook This is the code repository for TinyML Cookbook, published by Packt. Author: Gian Marco Iodice Publisher: Packt About the book This bo

Packt 93 Dec 29, 2022
Entity-Based Knowledge Conflicts in Question Answering.

Entity-Based Knowledge Conflicts in Question Answering Run Instructions | Paper | Citation | License This repository provides the Substitution Framewo

Apple 35 Oct 19, 2022
Automatically measure the facial Width-To-Height ratio and get facial analysis results provided by Microsoft Azure

fwhr-calc-website This project is to automatically measure the facial Width-To-Height ratio and get facial analysis results provided by Microsoft Azur

SoohyunPark 1 Feb 07, 2022