Efficient Scale-Permuted Backbone with Learned Resource Distribution
This work addresses efficiency improvements in neural network architectures for computer vision tasks, but it is incremental as it builds on existing scale-permuted models.
The paper tackled the problem of improving scale-permuted backbones like SpineNet by combining efficient operations and compound scaling with learned resource distribution, resulting in models that outperform EfficientNet-based models on object detection and achieve competitive performance on image classification and semantic segmentation.
Recently, SpineNet has demonstrated promising results on object detection and image classification over ResNet model. However, it is unclear if the improvement adds up when combining scale-permuted backbone with advanced efficient operations and compound scaling. Furthermore, SpineNet is built with a uniform resource distribution over operations. While this strategy seems to be prevalent for scale-decreased models, it may not be an optimal design for scale-permuted models. In this work, we propose a simple technique to combine efficient operations and compound scaling with a previously learned scale-permuted architecture. We demonstrate the efficiency of scale-permuted model can be further improved by learning a resource distribution over the entire network. The resulting efficient scale-permuted models outperform state-of-the-art EfficientNet-based models on object detection and achieve competitive performance on image classification and semantic segmentation. Code and models will be open-sourced soon.