1 Find the best model on the ImageNet dataset-paperswithcode
We look for the best model on the ImageNet dataset on paperswithcode
2 The best model on ImageNet-FixEfficientNet-L2 (by Sep, 13th, 2020)
The FixEfficientNet-L2 model comes from the paper "Fixing the train-test resolution discrepancy: FixEfficientNet"
link: https://arxiv.org/abs/2003.08237
The main idea of this article is the data augmentation strategy .
This article is the author's previous paper " "Fixing the train-test reso-lution discrepancy" expansion,
we can see the main idea of the article from here , we can see that the main idea
here is the strategy of data augmentation;
because we are to learn the idea of the backbone network model, So we will not delve into this article for the time being, but learn the backbone network model he used;
you can see that the model used in this article is EfficientNet;
so we continue to learn EfficientNet, "EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks"
. Please refer to my blog post "Image Classification-EfficientNet Study Notes";