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.
262 lines
9.1 KiB
262 lines
9.1 KiB
# (Tensorflow) EfficientNet Lite
|
|
|
|
**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).
|
|
|
|
EfficientNet-Lite makes EfficientNet more suitable for mobile devices by introducing [ReLU6](https://paperswithcode.com/method/relu6) activation functions and removing [squeeze-and-excitation blocks](https://paperswithcode.com/method/squeeze-and-excitation).
|
|
|
|
The weights from this model were ported from [Tensorflow/TPU](https://github.com/tensorflow/tpu).
|
|
|
|
## How do I use this model on an image?
|
|
|
|
To load a pretrained model:
|
|
|
|
```py
|
|
>>> import timm
|
|
>>> model = timm.create_model('tf_efficientnet_lite0', 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. `tf_efficientnet_lite0`. 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](../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('tf_efficientnet_lite0', 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: TF EfficientNet Lite
|
|
Paper:
|
|
Title: 'EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks'
|
|
URL: https://paperswithcode.com/paper/efficientnet-rethinking-model-scaling-for
|
|
Models:
|
|
- Name: tf_efficientnet_lite0
|
|
In Collection: TF EfficientNet Lite
|
|
Metadata:
|
|
FLOPs: 488052032
|
|
Parameters: 4650000
|
|
File Size: 18820223
|
|
Architecture:
|
|
- 1x1 Convolution
|
|
- Average Pooling
|
|
- Batch Normalization
|
|
- Convolution
|
|
- Dense Connections
|
|
- Dropout
|
|
- Inverted Residual Block
|
|
- RELU6
|
|
Tasks:
|
|
- Image Classification
|
|
Training Data:
|
|
- ImageNet
|
|
ID: tf_efficientnet_lite0
|
|
Crop Pct: '0.875'
|
|
Image Size: '224'
|
|
Interpolation: bicubic
|
|
Code: https://github.com/rwightman/pytorch-image-models/blob/9a25fdf3ad0414b4d66da443fe60ae0aa14edc84/timm/models/efficientnet.py#L1596
|
|
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite0-0aa007d2.pth
|
|
Results:
|
|
- Task: Image Classification
|
|
Dataset: ImageNet
|
|
Metrics:
|
|
Top 1 Accuracy: 74.83%
|
|
Top 5 Accuracy: 92.17%
|
|
- Name: tf_efficientnet_lite1
|
|
In Collection: TF EfficientNet Lite
|
|
Metadata:
|
|
FLOPs: 773639520
|
|
Parameters: 5420000
|
|
File Size: 21939331
|
|
Architecture:
|
|
- 1x1 Convolution
|
|
- Average Pooling
|
|
- Batch Normalization
|
|
- Convolution
|
|
- Dense Connections
|
|
- Dropout
|
|
- Inverted Residual Block
|
|
- RELU6
|
|
Tasks:
|
|
- Image Classification
|
|
Training Data:
|
|
- ImageNet
|
|
ID: tf_efficientnet_lite1
|
|
Crop Pct: '0.882'
|
|
Image Size: '240'
|
|
Interpolation: bicubic
|
|
Code: https://github.com/rwightman/pytorch-image-models/blob/9a25fdf3ad0414b4d66da443fe60ae0aa14edc84/timm/models/efficientnet.py#L1607
|
|
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite1-bde8b488.pth
|
|
Results:
|
|
- Task: Image Classification
|
|
Dataset: ImageNet
|
|
Metrics:
|
|
Top 1 Accuracy: 76.67%
|
|
Top 5 Accuracy: 93.24%
|
|
- Name: tf_efficientnet_lite2
|
|
In Collection: TF EfficientNet Lite
|
|
Metadata:
|
|
FLOPs: 1068494432
|
|
Parameters: 6090000
|
|
File Size: 24658687
|
|
Architecture:
|
|
- 1x1 Convolution
|
|
- Average Pooling
|
|
- Batch Normalization
|
|
- Convolution
|
|
- Dense Connections
|
|
- Dropout
|
|
- Inverted Residual Block
|
|
- RELU6
|
|
Tasks:
|
|
- Image Classification
|
|
Training Data:
|
|
- ImageNet
|
|
ID: tf_efficientnet_lite2
|
|
Crop Pct: '0.89'
|
|
Image Size: '260'
|
|
Interpolation: bicubic
|
|
Code: https://github.com/rwightman/pytorch-image-models/blob/9a25fdf3ad0414b4d66da443fe60ae0aa14edc84/timm/models/efficientnet.py#L1618
|
|
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite2-dcccb7df.pth
|
|
Results:
|
|
- Task: Image Classification
|
|
Dataset: ImageNet
|
|
Metrics:
|
|
Top 1 Accuracy: 77.48%
|
|
Top 5 Accuracy: 93.75%
|
|
- Name: tf_efficientnet_lite3
|
|
In Collection: TF EfficientNet Lite
|
|
Metadata:
|
|
FLOPs: 2011534304
|
|
Parameters: 8199999
|
|
File Size: 33161413
|
|
Architecture:
|
|
- 1x1 Convolution
|
|
- Average Pooling
|
|
- Batch Normalization
|
|
- Convolution
|
|
- Dense Connections
|
|
- Dropout
|
|
- Inverted Residual Block
|
|
- RELU6
|
|
Tasks:
|
|
- Image Classification
|
|
Training Data:
|
|
- ImageNet
|
|
ID: tf_efficientnet_lite3
|
|
Crop Pct: '0.904'
|
|
Image Size: '300'
|
|
Interpolation: bilinear
|
|
Code: https://github.com/rwightman/pytorch-image-models/blob/9a25fdf3ad0414b4d66da443fe60ae0aa14edc84/timm/models/efficientnet.py#L1629
|
|
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite3-b733e338.pth
|
|
Results:
|
|
- Task: Image Classification
|
|
Dataset: ImageNet
|
|
Metrics:
|
|
Top 1 Accuracy: 79.83%
|
|
Top 5 Accuracy: 94.91%
|
|
- Name: tf_efficientnet_lite4
|
|
In Collection: TF EfficientNet Lite
|
|
Metadata:
|
|
FLOPs: 5164802912
|
|
Parameters: 13010000
|
|
File Size: 52558819
|
|
Architecture:
|
|
- 1x1 Convolution
|
|
- Average Pooling
|
|
- Batch Normalization
|
|
- Convolution
|
|
- Dense Connections
|
|
- Dropout
|
|
- Inverted Residual Block
|
|
- RELU6
|
|
Tasks:
|
|
- Image Classification
|
|
Training Data:
|
|
- ImageNet
|
|
ID: tf_efficientnet_lite4
|
|
Crop Pct: '0.92'
|
|
Image Size: '380'
|
|
Interpolation: bilinear
|
|
Code: https://github.com/rwightman/pytorch-image-models/blob/9a25fdf3ad0414b4d66da443fe60ae0aa14edc84/timm/models/efficientnet.py#L1640
|
|
Weights: https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite4-741542c3.pth
|
|
Results:
|
|
- Task: Image Classification
|
|
Dataset: ImageNet
|
|
Metrics:
|
|
Top 1 Accuracy: 81.54%
|
|
Top 5 Accuracy: 95.66%
|
|
--> |