You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
pytorch-image-models/hfdocs/source/models/efficientnet.mdx

392 lines
13 KiB

# EfficientNet
**EfficientNet** is a convolutional neural network architecture and scaling method that uniformly scales all dimensions of depth/width/resolution using a *compound coefficient*. Unlike conventional practice that arbitrary scales these factors, the EfficientNet scaling method uniformly scales network width, depth, and resolution with a set of fixed scaling coefficients. For example, if we want to use $2^N$ times more computational resources, then we can simply increase the network depth by $\alpha ^ N$, width by $\beta ^ N$, and image size by $\gamma ^ N$, where $\alpha, \beta, \gamma$ are constant coefficients determined by a small grid search on the original small model. EfficientNet uses a compound coefficient $\phi$ to uniformly scales network width, depth, and resolution in a principled way.
The compound scaling method is justified by the intuition that if the input image is bigger, then the network needs more layers to increase the receptive field and more channels to capture more fine-grained patterns on the bigger image.
The base EfficientNet-B0 network is based on the inverted bottleneck residual blocks of [MobileNetV2](https://paperswithcode.com/method/mobilenetv2), in addition to [squeeze-and-excitation blocks](https://paperswithcode.com/method/squeeze-and-excitation-block).
## How do I use this model on an image?
To load a pretrained model:
```py
>>> import timm
>>> model = timm.create_model('efficientnet_b0', pretrained=True)
>>> model.eval()
```
To load and preprocess the image:
```py
>>> import urllib
>>> from PIL import Image
>>> from timm.data import resolve_data_config
>>> from timm.data.transforms_factory import create_transform
>>> config = resolve_data_config({}, model=model)
>>> transform = create_transform(**config)
>>> url, filename = ("https://github.com/pytorch/hub/raw/master/images/dog.jpg", "dog.jpg")
>>> urllib.request.urlretrieve(url, filename)
>>> img = Image.open(filename).convert('RGB')
>>> tensor = transform(img).unsqueeze(0) # transform and add batch dimension
```
To get the model predictions:
```py
>>> import torch
>>> with torch.no_grad():
... out = model(tensor)
>>> probabilities = torch.nn.functional.softmax(out[0], dim=0)
>>> print(probabilities.shape)
>>> # prints: torch.Size([1000])
```
To get the top-5 predictions class names:
```py
>>> # Get imagenet class mappings
>>> url, filename = ("https://raw.githubusercontent.com/pytorch/hub/master/imagenet_classes.txt", "imagenet_classes.txt")
>>> urllib.request.urlretrieve(url, filename)
>>> with open("imagenet_classes.txt", "r") as f:
... categories = [s.strip() for s in f.readlines()]
>>> # Print top categories per image
>>> top5_prob, top5_catid = torch.topk(probabilities, 5)
>>> for i in range(top5_prob.size(0)):
... print(categories[top5_catid[i]], top5_prob[i].item())
>>> # prints class names and probabilities like:
>>> # [('Samoyed', 0.6425196528434753), ('Pomeranian', 0.04062102362513542), ('keeshond', 0.03186424449086189), ('white wolf', 0.01739676296710968), ('Eskimo dog', 0.011717947199940681)]
```
Replace the model name with the variant you want to use, e.g. `efficientnet_b0`. You can find the IDs in the model summaries at the top of this page.
To extract image features with this model, follow the [timm feature extraction examples](https://rwightman.github.io/pytorch-image-models/feature_extraction/), just change the name of the model you want to use.
## How do I finetune this model?
You can finetune any of the pre-trained models just by changing the classifier (the last layer).
```py
>>> model = timm.create_model('efficientnet_b0', pretrained=True, num_classes=NUM_FINETUNE_CLASSES)
```
To finetune on your own dataset, you have to write a training loop or adapt [timm's training
script](https://github.com/rwightman/pytorch-image-models/blob/master/train.py) to use your dataset.
## How do I train this model?
You can follow the [timm recipe scripts](scripts) for training a new model afresh.
## Citation
```BibTeX
@misc{tan2020efficientnet,
title={EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks},
author={Mingxing Tan and Quoc V. Le},
year={2020},
eprint={1905.11946},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
<!--
Type: model-index
Collections:
- Name: EfficientNet
Paper:
Title: 'EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks'
URL: https://paperswithcode.com/paper/efficientnet-rethinking-model-scaling-for
Models:
- Name: efficientnet_b0
In Collection: EfficientNet
Metadata:
FLOPs: 511241564
Parameters: 5290000
File Size: 21376743
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b0
Layers: 18
Crop Pct: '0.875'
Image Size: '224'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1002
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b0_ra-3dd342df.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 77.71%
Top 5 Accuracy: 93.52%
- Name: efficientnet_b1
In Collection: EfficientNet
Metadata:
FLOPs: 909691920
Parameters: 7790000
File Size: 31502706
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b1
Crop Pct: '0.875'
Image Size: '240'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1011
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b1-533bc792.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 78.71%
Top 5 Accuracy: 94.15%
- Name: efficientnet_b2
In Collection: EfficientNet
Metadata:
FLOPs: 1265324514
Parameters: 9110000
File Size: 36788104
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b2
Crop Pct: '0.875'
Image Size: '260'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1020
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b2_ra-bcdf34b7.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 80.38%
Top 5 Accuracy: 95.08%
- Name: efficientnet_b2a
In Collection: EfficientNet
Metadata:
FLOPs: 1452041554
Parameters: 9110000
File Size: 49369973
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b2a
Crop Pct: '1.0'
Image Size: '288'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1029
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b3_ra2-cf984f9c.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 80.61%
Top 5 Accuracy: 95.32%
- Name: efficientnet_b3
In Collection: EfficientNet
Metadata:
FLOPs: 2327905920
Parameters: 12230000
File Size: 49369973
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b3
Crop Pct: '0.904'
Image Size: '300'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1038
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b3_ra2-cf984f9c.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 82.08%
Top 5 Accuracy: 96.03%
- Name: efficientnet_b3a
In Collection: EfficientNet
Metadata:
FLOPs: 2600628304
Parameters: 12230000
File Size: 49369973
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_b3a
Crop Pct: '1.0'
Image Size: '320'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1047
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b3_ra2-cf984f9c.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 82.25%
Top 5 Accuracy: 96.11%
- Name: efficientnet_em
In Collection: EfficientNet
Metadata:
FLOPs: 3935516480
Parameters: 6900000
File Size: 27927309
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_em
Crop Pct: '0.882'
Image Size: '240'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1118
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_em_ra2-66250f76.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 79.26%
Top 5 Accuracy: 94.79%
- Name: efficientnet_es
In Collection: EfficientNet
Metadata:
FLOPs: 2317181824
Parameters: 5440000
File Size: 22003339
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_es
Crop Pct: '0.875'
Image Size: '224'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1110
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_es_ra-f111e99c.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 78.09%
Top 5 Accuracy: 93.93%
- Name: efficientnet_lite0
In Collection: EfficientNet
Metadata:
FLOPs: 510605024
Parameters: 4650000
File Size: 18820005
Architecture:
- 1x1 Convolution
- Average Pooling
- Batch Normalization
- Convolution
- Dense Connections
- Dropout
- Inverted Residual Block
- Squeeze-and-Excitation Block
- Swish
Tasks:
- Image Classification
Training Data:
- ImageNet
ID: efficientnet_lite0
Crop Pct: '0.875'
Image Size: '224'
Interpolation: bicubic
Code: https://github.com/rwightman/pytorch-image-models/blob/a7f95818e44b281137503bcf4b3e3e94d8ffa52f/timm/models/efficientnet.py#L1163
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_lite0_ra-37913777.pth
Results:
- Task: Image Classification
Dataset: ImageNet
Metrics:
Top 1 Accuracy: 75.5%
Top 5 Accuracy: 92.51%
-->