基于pytorch的CV入门教程(图像分类)
Author: Ruilong Ren
Date: 2022-1-12
GPU: 4 × GTX 1080Ti
- LetNet
- AlexNet
- ResNet18/34/50/101/152
- GoogLeNet
- VGG16/19
- DenseNet
- ResNeXt
- SENet
- MobileNetv2
- EfficientNetB0
- Darknet53
- CSPDarknet53
- MNIST
- CIFAR10
- PASCAL VOC2007/2012
- EPOCH: 100
- LR: CosineAnnealing
METHODS | BEST-ACC | TIME(s) |
---|---|---|
LetNet | 71.8% | 1754.7 |
AlexNet | 70.09% | 6292.1 |
ResNet18 | 95.23% | 3987.9 |
ResNet50 | 95.2% | 8468.4 |
GoogLeNet | 95.02% | 13144.3 |
VGG16 | 93.8% | 9618 |
DenseNet121 | 94.55% | 7735.3 |
ResNeXt29_32×4d | 95.18% | 9573.5 |
SENet18 | 95.22 % | 6988.3 |
MobileNetv2 | 93.37% | 5995.6 |
EfficientNetB0 | 90.49% | 7210.8 |
Darknet53 | 91.85% | 8440.4 |
CSPDarknet53 | 88.75% | 5898.9 |
MobileNetv3 | ||
ShuffleNetv2_x0.5 |