Blog
The latest news from Google AI
E"cientNet: Improving Accuracy and E"ciency through AutoML and
Model Scaling
Wednesday, May 29, 2019
Posted by Mingxing Tan, Staff Software Engineer and Quoc V. Le, Principal Scientist, Google AI
Convolutional neural networks (CNNs) are commonly developed at a Pxed resource cost, and
then scaled up in order to achieve better accuracy when more resources are made available.
For example, ResNet can be scaled up from ResNet-18 to ResNet-200 by increasing the
number of layers, and recently, GPipe achieved 84.3% ImageNet top-1 accuracy by scaling up
a baseline CNN by a factor of four. The conventional practice for model scaling is to
arbitrarily increase the CNN depth or width, or to use larger input image resolution for training
and evaluation. While these methods do improve accuracy, they usually require tedious
manual tuning, and still often yield suboptimal performance. What if, instead, we could Pnd a
more principled method to scale up a CNN to obtain better accuracy and eYciency?
In our ICML 2019 paper, “EYcientNet: Rethinking Model Scaling for Convolutional Neural
Networks”, we propose a novel model scaling method that uses a simple yet highly effective
compound coe*cient to scale up CNNs in a more structured manner. Unlike conventional
approaches that arbitrarily scale network dimensions, such as width, depth and resolution,
our method uniformly scales each dimension with a Pxed set of scaling coeYcients.
Powered by this novel scaling method and recent progress on AutoML, we have developed a
family of models, called EYcientNets, which superpass state-of-the-art accuracy with up to
10x better eYciency (smaller and faster).
Compound Model Scaling: A Better Way to Scale Up CNNs
In order to understand the effect of scaling the network, we systematically studied the
impact of scaling different dimensions of the model. While scaling individual dimensions
improves model performance, we observed that balancing all dimensions of the network—
width, depth, and image resolution—against the available resources would best improve
overall performance.
The Prst step in the compound scaling method is to perform a grid search to Pnd the
relationship between different scaling dimensions of the baseline network under a Pxed
resource constraint (e.g., 2x more FLOPS).This determines the appropriate scaling coeYcient
for each of the dimensions mentioned above. We then apply those coeYcients to scale up
the baseline network to the desired target model size or computational budget.
Comparison of different scaling methods. Unlike conventional scaling methods (b)-(d) that arbitrary scale a single
dimension of the network, our compound scaling method uniformly scales up all dimensions in a principled way.
This compound scaling method consistently improves model accuracy and eYciency for
scaling up existing models such as MobileNet (+1.4% imagenet accuracy), and ResNet
(+0.7%), compared to conventional scaling methods.
E"cientNet Architecture
The effectiveness of model scaling also relies heavily on the baseline network. So, to further
improve performance, we have also developed a new baseline network by performing a
neural architecture search using the AutoML MNAS framework, which optimizes both
accuracy and eYciency (FLOPS). The resulting architecture uses mobile inverted bottleneck
convolution (MBConv), similar to MobileNetV2 and MnasNet, but is slightly larger due to an
increased FLOP budget. We then scale up the baseline network to obtain a family of models,
called E*cientNets.
The architecture for our baseline network EYcientNet-B0 is simple and clean, making it easier to scale and generalize.
E"cientNet Performance
We have compared our EYcientNets with other existing CNNs on ImageNet. In general, the
EYcientNet models achieve both higher accuracy and better eYciency over existing CNNs,
reducing parameter size and FLOPS by an order of magnitude. For example, in the high-
accuracy regime, our EYcientNet-B7 reaches state-of-the-art 84.4% top-1 / 97.1% top-5
accuracy on ImageNet, while being 8.4x smaller and 6.1x faster on CPU inference than the
previous Gpipe. Compared with the widely used ResNet-50, our EYcientNet-B4 uses similar
FLOPS, while improving the top-1 accuracy from 76.3% of ResNet-50 to 82.6% (+6.3%).
Model Size vs. Accuracy Comparison. EYcientNet-B0 is the baseline network developed by AutoML MNAS, while EYcient-B1
to B7 are obtained by scaling up the baseline network. In particular, our EYcientNet-B7 achieves new state-of-the-art 84.4%
top-1 / 97.1% top-5 accuracy, while being 8.4x smaller than the best existing CNN.
Though EYcientNets perform well on ImageNet, to be most useful, they should also transfer
to other datasets. To evaluate this, we tested EYcientNets on eight widely used transfer
learning datasets. EYcientNets achieved state-of-the-art accuracy in 5 out of the 8 datasets,
such as CIFAR-100 (91.7%) and Flowers (98.8%), with an order of magnitude fewer
parameters (up to 21x parameter reduction), suggesting that our EYcientNets also transfer
well.
By providing signiPcant improvements to model eYciency, we expect EYcientNets could
potentially serve as a new foundation for future computer vision tasks. Therefore, we have
open-sourced all EYcientNet models, which we hope can benePt the larger machine learning
community. You can Pnd the EYcientNet source code and TPU training scripts here.
Acknowledgements:
Special thanks to Hongkun Yu, Ruoming Pang, Vijay Vasudevan, Alok Aggarwal, Barret Zoph,
Xianzhi Du, Xiaodan Song, Samy Bengio, Jeff Dean, and the Google Brain team.
# ! "
Google · Privacy · Terms