# Fixing the train-test resolution discrepancy: FixEfficientNet

arXiv: https://arxiv.org/abs/2003.08237 · HTML: https://mlsanity.com/paper/2003.08237/

## Key facts

- Cited by 31 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Mar 2020.

## Built on by (most cited first, 30 of 31)

- [EfficientNetV2: Smaller Models and Faster Training](https://mlsanity.com/paper/2104.00298/) (2021)
- [Going deeper with Image Transformers](https://mlsanity.com/paper/2103.17239/) (2021)
- [DeiT III: Revenge of the ViT](https://mlsanity.com/paper/2204.07118/) (2022)
- [XCiT: Cross-Covariance Image Transformers](https://mlsanity.com/paper/2106.09681/) (2021)
- [Self-supervised Pretraining of Visual Features in the Wild](https://mlsanity.com/paper/2103.01988/) (2021)
- [Pixel-wise Anomaly Detection in Complex Driving Scenes](https://mlsanity.com/paper/2103.05445/) (2021)
- [Sparse Training via Boosting Pruning Plasticity with Neuroregeneration](https://mlsanity.com/paper/2106.10404/) (2021)
- [Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision](https://mlsanity.com/paper/2202.08360/) (2022)
- [Pufferfish: Communication-efficient Models At No Extra Cost](https://mlsanity.com/paper/2103.03936/) (2021)
- [Compute and Energy Consumption Trends in Deep Learning Inference](https://mlsanity.com/paper/2109.05472/) (2021)
- [Deep Ensembling with No Overhead for either Training or Testing: The All-Round Blessings of Dynamic Sparsity](https://mlsanity.com/paper/2106.14568/) (2021)
- [Momentum Residual Neural Networks](https://mlsanity.com/paper/2102.07870/) (2021)
- [ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition](https://mlsanity.com/paper/2104.03841/) (2021)
- [Stealthy Backdoors as Compression Artifacts](https://mlsanity.com/paper/2104.15129/) (2021)
- [Explaining Bayesian Neural Networks](https://mlsanity.com/paper/2108.10346/) (2021)
- [Backward-Compatible Prediction Updates: A Probabilistic Approach](https://mlsanity.com/paper/2107.01057/) (2021)
- [Revealing the Distributional Vulnerability of Discriminators by Implicit Generators](https://mlsanity.com/paper/2108.09976/) (2021)
- [KonX: Cross-Resolution Image Quality Assessment](https://mlsanity.com/paper/2212.05813/) (2022)
- [Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training](https://mlsanity.com/paper/2106.03640/) (2021)
- [NAF: Zero-Shot Feature Upsampling via Neighborhood Attention Filtering](https://mlsanity.com/paper/2511.18452/) (2025)
- [Tackling the Background Bias in Sparse Object Detection via Cropped Windows](https://mlsanity.com/paper/2106.02288/) (2021)
- [Accelerating Self-Supervised Learning via Efficient Training Strategies](https://mlsanity.com/paper/2212.05611/) (2022)
- [Estimating Galactic Distances From Images Using Self-supervised Representation Learning](https://mlsanity.com/paper/2101.04293/) (2021)
- [Towards an Analytical Definition of Sufficient Data](https://mlsanity.com/paper/2202.03238/) (2022)
- [Exploiting latent representation of sparse semantic layers for improved short-term motion prediction with Capsule Networks](https://mlsanity.com/paper/2103.01644/) (2021)
- [Learnable Adaptive Cosine Estimator (LACE) for Image Classification](https://mlsanity.com/paper/2110.05324/) (2021)
- [Recognition-Aware Learned Image Compression](https://mlsanity.com/paper/2202.00198/) (2022)
- [Multi-modal anticipation of stochastic trajectories in a dynamic environment with Conditional Variational Autoencoders](https://mlsanity.com/paper/2103.03912/) (2021)
- [Content-Augmented Feature Pyramid Network with Light Linear Spatial Transformers for Object Detection](https://mlsanity.com/paper/2105.09464/) (2021)
- [Enhanced Performance of Pre-Trained Networks by Matched Augmentation Distributions](https://mlsanity.com/paper/2201.07894/) (2022)

## Abstract (arXiv)

This paper provides an extensive analysis of the performance of the EfficientNet image classifiers with several recent training procedures, in particular one that corrects the discrepancy between train and test images. The resulting network, called FixEfficientNet, significantly outperforms the initial architecture with the same number of parameters. For instance, our FixEfficientNet-B0 trained without additional training data achieves 79.3% top-1 accuracy on ImageNet with 5.3M parameters. This is a +0.5% absolute improvement over the Noisy student EfficientNet-B0 trained with 300M unlabeled images. An EfficientNet-L2 pre-trained with weak supervision on 300M unlabeled images and further optimized with FixRes achieves 88.5% top-1 accuracy (top-5: 98.7%), which establishes the new state of the art for ImageNet with a single crop. These improvements are thoroughly evaluated with cleaner protocols than the one usually employed for Imagenet, and particular we show that our improvement remains in the experimental setting of ImageNet-v2, that is less prone to overfitting, and with ImageNet Real Labels. In both cases we also establish the new state of the art.