Shortcuts

EfficientNetV2

摘要

This paper introduces EfficientNetV2, a new family of convolutional networks that have faster training speed and better parameter efficiency than previous models. To develop this family of models, we use a combination of training-aware neural architecture search and scaling, to jointly optimize training speed and parameter efficiency. The models were searched from the search space enriched with new ops such as Fused-MBConv. Our experiments show that EfficientNetV2 models train much faster than state-of-the-art models while being up to 6.8x smaller. Our training can be further sped up by progressively increasing the image size during training, but it often causes a drop in accuracy. To compensate for this accuracy drop, we propose to adaptively adjust regularization (e.g., dropout and data augmentation) as well, such that we can achieve both fast training and good accuracy. With progressive learning, our EfficientNetV2 significantly outperforms previous models on ImageNet and CIFAR/Cars/Flowers datasets. By pretraining on the same ImageNet21k, our EfficientNetV2 achieves 87.3% top-1 accuracy on ImageNet ILSVRC2012, outperforming the recent ViT by 2.0% accuracy while training 5x-11x faster using the same computing resources. Code will be available at https://github.com/google/automl/tree/master/efficientnetv2.

使用方式

from mmpretrain import inference_model

predict = inference_model('efficientnetv2-b0_3rdparty_in1k', 'demo/bird.JPEG')
print(predict['pred_class'])
print(predict['pred_score'])

Models and results

Pretrained models

模型

Params (M)

Flops (G)

配置文件

下载

efficientnetv2-s_3rdparty_in21k*

48.16

3.31

config

model

efficientnetv2-m_3rdparty_in21k*

80.84

5.86

config

model

efficientnetv2-l_3rdparty_in21k*

145.22

13.11

config

model

efficientnetv2-xl_3rdparty_in21k*

234.82

18.86

config

model

Models with * are converted from the timm. The config files of these models are only for inference. We haven’t reproduce the training results.

Image Classification on ImageNet-1k

模型

预训练

Params (M)

Flops (G)

Top-1 (%)

Top-5 (%)

配置文件

下载

efficientnetv2-b0_3rdparty_in1k*

从头训练

7.14

0.92

78.52

94.44

config

model

efficientnetv2-b1_3rdparty_in1k*

从头训练

8.14

1.44

79.80

94.89

config

model

efficientnetv2-b2_3rdparty_in1k*

从头训练

10.10

1.99

80.63

95.30

config

model

efficientnetv2-b3_3rdparty_in1k*

从头训练

14.36

3.50

82.03

95.88

config

model

efficientnetv2-s_3rdparty_in1k*

从头训练

21.46

9.72

83.82

96.67

config

model

efficientnetv2-m_3rdparty_in1k*

从头训练

54.14

26.88

85.01

97.26

config

model

efficientnetv2-l_3rdparty_in1k*

从头训练

118.52

60.14

85.43

97.31

config

model

efficientnetv2-s_in21k-pre_3rdparty_in1k*

ImageNet-21k

21.46

9.72

84.29

97.26

config

model

efficientnetv2-m_in21k-pre_3rdparty_in1k*

ImageNet-21k

54.14

26.88

85.47

97.76

config

model

efficientnetv2-l_in21k-pre_3rdparty_in1k*

ImageNet-21k

118.52

60.14

86.31

97.99

config

model

efficientnetv2-xl_in21k-pre_3rdparty_in1k*

ImageNet-21k

208.12

98.34

86.39

97.83

config

model

Models with * are converted from the timm. The config files of these models are only for inference. We haven’t reproduce the training results.

引用

@inproceedings{tan2021efficientnetv2,
  title={Efficientnetv2: Smaller models and faster training},
  author={Tan, Mingxing and Le, Quoc},
  booktitle={International Conference on Machine Learning},
  pages={10096--10106},
  year={2021},
  organization={PMLR}
}
Read the Docs v: latest
Versions
latest
stable
mmcls-1.x
mmcls-0.x
dev
Downloads
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.