A Keras implementation of EfficientNet
 Mingxing Tan and Quoc V. Le. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. ICML 2019. Arxiv link: https://arxiv.org/abs/1905.11946.
1. About EfficientNet Models
EfficientNets are a family of image classification models, which achieve state-of-the-art accuracy, yet being an order-of-magnitude smaller and faster than previous models.
We develop EfficientNets based on AutoML and Compound Scaling. In particular, we first use AutoML Mobile framework to develop a mobile-size baseline network, named as EfficientNet-B0; Then, we use the compound scaling method to scale up this baseline to obtain EfficientNet-B1 to B7.
EfficientNets achieve state-of-the-art accuracy on ImageNet with an order of magnitude better efficiency:
In high-accuracy regime, our EfficientNet-B7 achieves state-of-the-art 84.4% top-1 / 97.1% top-5 accuracy on ImageNet with 66M parameters and 37B FLOPS, being 8.4x smaller and 6.1x faster on CPU inference than previous best Gpipe.
In middle-accuracy regime, our EfficientNet-B1 is 7.6x smaller and 5.7x faster on CPU inference than ResNet-152, with similar ImageNet accuracy.
Compared with the widely used ResNet-50, our EfficientNet-B4 improves the top-1 accuracy from 76.3% of ResNet-50 to 82.6% (+6.3%), under similar FLOPS constraint.
2. Get the weights
In order to get the TF official weights.
- Create a Colab file in your Google Drive.
- Mount the drive with
from google.colab import drive drive.mount('/content/drive/')
- Run this command in your created Colab
!gsutil cp -r gs://cloud-tpu-checkpoints/efficientnet "/content/drive/My Drive/effnets/"
Download the effnets folder from your drive and extract somewhere.
Extract the weights from each layer to a directory by.
from extract_weights import extract_tensors_from_checkpoint_file extract_tensors_from_checkpoint_file('efficientnet-b0/model.ckpt-109400') # change this line to your extracted directory
- Each TF weights directory should be like.
best_eval.txt checkpoint model.ckpt-12345.data-00000-of-00001 model.ckpt-12345.index model.ckpt-12345.meta
model.ckpt-12345 in this case.
Create the hdf5 weights by run the
load_weights.py, make sure to change
The example of testing a panda image is in