EfficientNetV2-Pytorch Image Classification

EfficientNetV2: Smaller Models and Faster Training

Training efficiency is important to deep learning as model size and training data size are increasingly larger.

PyTorch Implementation of EfficientNetV2 Family PyTorch implementation of EfficientNet V2 Reproduction of EfficientNet V2 architecture as described in EfficientNetV2: Smaller Models and Faster Training by Mingxing Tan, Quoc V. Le with the PyTorch framework.

Models Architecture # Parameters FLOPs Top-1 Acc. (%)

EfficientNetV2-S 22.10M 8.42G @ 384

EfficientNetV2-M 55.30M 24.74G @ 480

EfficientNetV2-L 119.36M 56.13G @ 480

EfficientNetV2-XL 208.96M 93.41G @ 512

References

https://paperswithcode.com/paper/efficientnetv2-smaller-models-and-faster

https://wandb.ai/wandb_fc/pytorch-image-models/reports/EfficientNetV2--Vmlldzo2NTkwNTQ

http://pytorch.org/vision/main/generated/torchvision.models.efficientnet_v2_s.html

https://pytorch.org/vision/main/_modules/torchvision/models/efficientnet.html