# Mix & Match: training convnets with mixed image sizes for improved accuracy, speed and scale resiliency

arXiv: https://arxiv.org/abs/1908.08986 · HTML: https://mlsanity.com/paper/1908.08986/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Aug 2019.

## Built on by (most cited first, 12 of 12)

- [Restormer: Efficient Transformer for High-Resolution Image Restoration](https://mlsanity.com/paper/2111.09881/) (2021)
- [EfficientNetV2: Smaller Models and Faster Training](https://mlsanity.com/paper/2104.00298/) (2021)
- [Learning Enriched Features for Fast Image Restoration and Enhancement](https://mlsanity.com/paper/2205.01649/) (2022)
- [S4ND: Modeling Images and Videos as Multidimensional Signals Using State Spaces](https://mlsanity.com/paper/2210.06583/) (2022)
- [Compute-Efficient Deep Learning: Algorithmic Trends and Opportunities](https://mlsanity.com/paper/2210.06640/) (2022)
- [Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training](https://mlsanity.com/paper/2106.03640/) (2021)
- [Accelerating Self-Supervised Learning via Efficient Training Strategies](https://mlsanity.com/paper/2212.05611/) (2022)
- [On the Efficacy of Multi-scale Data Samplers for Vision Applications](https://mlsanity.com/paper/2309.04502/) (2023)
- [IrrNet: Advancing Irrigation Mapping with Incremental Patch Size Training on Remote Sensing Imagery](https://mlsanity.com/paper/2404.11762/) (2024)
- [Characterizing and Taming Resolution in Convolutional Neural Networks](https://mlsanity.com/paper/2110.14819/) (2021)
- [Temporally Resolution Decrement: Utilizing the Shape Consistency for Higher Computational Efficiency](https://mlsanity.com/paper/2112.00954/) (2021)
- [Exploring Deep-to-Shallow Transformable Neural Networks for Intelligent Embedded Systems](https://mlsanity.com/paper/2512.19731/) (2025)

## Abstract (arXiv)

Convolutional neural networks (CNNs) are commonly trained using a fixed spatial image size predetermined for a given model. Although trained on images of aspecific size, it is well established that CNNs can be used to evaluate a wide range of image sizes at test time, by adjusting the size of intermediate feature maps. In this work, we describe and evaluate a novel mixed-size training regime that mixes several image sizes at training time. We demonstrate that models trained using our method are more resilient to image size changes and generalize well even on small images. This allows faster inference by using smaller images attest time. For instance, we receive a 76.43% top-1 accuracy using ResNet50 with an image size of 160, which matches the accuracy of the baseline model with 2x fewer computations. Furthermore, for a given image size used at test time, we show this method can be exploited either to accelerate training or the final test accuracy. For example, we are able to reach a 79.27% accuracy with a model evaluated at a 288 spatial size for a relative improvement of 14% over the baseline.