style-based-recalibration-module

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

style-based-recalibration-module

(★ 61)

PyTorch code for our paper : "SRM : A Style-based Recalibration Module for Convolutional Neural Networks" (https://arxiv.org/abs/1903.10829)

style-based-recalibration-module 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Style-Based Recalibration Module
The official PyTorch implementation of "SRM : A Style-based Recalibration Module for Convolutional Neural Networks" for ImageNet.
SRM is a lightweight architectural unit that dynamically recalibrates feature responses based on style importance.

![](https://raw.githubusercontent.com/hyunjaelee410/style-based-recalibration-module/master/figures/overview.png)

## Overview of Results

### Training and validation curves on ImageNet with ResNet-50

![](https://raw.githubusercontent.com/hyunjaelee410/style-based-recalibration-module/master/figures/imagenet_curve.png)

### Top-1 and top-5  accuracy (%) on the ImageNet-1K validation set

![](https://raw.githubusercontent.com/hyunjaelee410/style-based-recalibration-module/master/figures/imagenet.png)

### Example results of style transfer

![](https://raw.githubusercontent.com/hyunjaelee410/style-based-recalibration-module/master/figures/style_transfer.png)

## Prerequisites
- PyTorch 0.4.0+
- Python 3.6
- CUDA 8.0+

## Training Examples
* Train **ResNet-50**
```
python imagenet.py --depth 50 --data /data/imagenet/ILSVRC2012 --gpu-id 0,1,2,3,4,5,6,7 --checkpoint resnet50/baseline
```

* Train **SRM-ResNet-50**
```
python imagenet.py --depth 50 --data /data/imagenet/ILSVRC2012 --gpu-id 0,1,2,3,4,5,6,7 --checkpoint resnet50/srm --recalibration-type srm
```

* Train **SE-ResNet-50**
```
python imagenet.py --depth 50 --data /data/imagenet/ILSVRC2012 --gpu-id 0,1,2,3,4,5,6,7 --checkpoint resnet50/se --recalibration-type se
```

* Train **GE-ResNet-50**
```
python imagenet.py --depth 50 --data /data/imagenet/ILSVRC2012 --gpu-id 0,1,2,3,4,5,6,7 --checkpoint resnet50/ge --recalibration-type ge
```

## Acknowledgment
This code is heavily borrowed from [pytorch-classification](https://github.com/bearpaw/pytorch-classification).

## Note
* 28/05/2019: initial code for ImageNet is released

style-based-recalibration-module

style-based-recalibration-module 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

style-based-recalibration-module

style-based-recalibration-module 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README