https://github.com/cjf8899/KeepAugment_Pytorch

:low_brightness:Unofficial PyTorch implementation of KeepAugment
https://github.com/cjf8899/KeepAugment_Pytorch

augmentation implementation keepaugment pytorch saliency-map

Last synced: about 1 month ago
JSON representation

:low_brightness:Unofficial PyTorch implementation of KeepAugment

• Host: GitHub
• URL: https://github.com/cjf8899/KeepAugment_Pytorch
• Owner: cjf8899
• Created: 2022-06-02T05:14:45.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-06-06T07:30:07.000Z (over 2 years ago)
• Last Synced: 2024-07-10T07:51:43.535Z (2 months ago)
• Topics: augmentation, implementation, keepaugment, pytorch, saliency-map
• Language: Python
• Homepage:
• Size: 59.6 KB
• Stars: 21
• Watchers: 1
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# KeepAugment_Pytorch

## Unofficial implementation of ["KeepAugment: A Simple Information-Preserving Data Augmentation Approach"](https://openaccess.thecvf.com/content/CVPR2021/papers/Gong_KeepAugment_A_Simple_Information-Preserving_Data_Augmentation_Approach_CVPR_2021_paper.pdf). CVPR2021




## Results 

### CIFAR 10

|              Model            |     ResNet-18     |  Wide ResNet-28-10  |

| :---------------------------: | :---------------: | :-----------------: |

| Cutout                        |  95.6±0.1(paepr)  |  96.9±0.1(paepr)    |

| KeepCutout                    |        96.2       |        97.2         |

| KeepCutout (low resolution)   |        96.1       |        97.1         |

| KeepCutout (early loss)       |        96.2       |        TODO         |

| KeepCutout (low + early)      |        96.2       |        TODO         |

|              Model              |     ResNet-18     |  Wide ResNet-28-10  |

| :-----------------------------: | :---------------: | :-----------------: |

| AutoAugment                     |       96.0        |  97.3±0.1(paepr)    |

| KeepAutoAugment                 |       96.4        |          97.4       |

| KeepAutoAugment (low resolution)|       96.3        |          97.5       |

| KeepAutoAugment (early loss)    |       96.3        |          97.4       |

| KeepAutoAugment (low + early)   |       96.5        |         TODO        |

All results have slight differences(±0.1).

## Run

The type of method is 


'keep_cutout', 'keep_cutout_low', 'keep_cutout_early', 'keep_cutout_low_early',


'keep_autoaugment', 'keep_autoaugment_low', 'keep_autoaugment_early', 'keep_autoaugment_low_early'. 

The type of model is 'resnet', 'wide_resnet'.

```Shell

python train.py --model resnet --method keep_cutout

```

### Any feedback on code simplification and incorrect implementation would be appreciated!

## Referenced. Thank you all:+1:

baseline & cutout code : https://github.com/uoguelph-mlrg/Cutout


autoaugment code : https://github.com/DeepVoltaire/AutoAugment


saliency map code : https://github.com/sunnynevarekar/pytorch-saliency-maps