[NEW!] Check out our latest work involution accepted to CVPR'21 that introduces a new neural operator, other than convolution and self-attention.
Reproduction of EfficientNet V2 architecture as described in EfficientNetV2: Smaller Models and Faster Training by Mingxing Tan, Quoc V. Le with the PyTorch framework.
PyTorch 1.7+ is required to support nn.SiLU
| Architecture | # Parameters | FLOPs | Top-1 Acc. (%) |
|---|---|---|---|
| EfficientNetV2-S | 24M | 8.8B |
More model definitions are pending for architectural details from the authors.
Stay tuned for ImageNet pre-trained weights.
The implementation is heavily borrowed from HBONet or MobileNetV2, please kindly consider citing the following
@InProceedings{Li_2019_ICCV,
author = {Li, Duo and Zhou, Aojun and Yao, Anbang},
title = {HBONet: Harmonious Bottleneck on Two Orthogonal Dimensions},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {Oct},
year = {2019}
}
@InProceedings{Sandler_2018_CVPR,
author = {Sandler, Mark and Howard, Andrew and Zhu, Menglong and Zhmoginov, Andrey and Chen, Liang-Chieh},
title = {MobileNetV2: Inverted Residuals and Linear Bottlenecks},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2018}
}