Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networks

By Jierun Chen, Shiu-hong Kao, Hao He, Weipeng Zhuo, Song Wen, Chul-Ho Lee, Shueng-Han Gary Chan. Computer Vision and Pattern Recognition, 2023

1,893 citations recorded by the source corpus.

To design fast neural networks, many works have been focusing on reducing the number of floating-point operations (FLOPs). We observe that such reduction in FLOPs, however, does not necessarily lead to a similar level of re-duction in latency. This mainly stems from inefficiently low floating-point operations per second (FLOPS). To achieve faster networks, we revisit popular operators and demonstrate that such low FLOPS is mainly due to frequent memory access of the operators, especially the depthwise con-volution. We hence propose a novel partial convolution (PConv) that extracts spatial features more efficiently, by cutting down redundant computation and memory access simultaneously. Building upon our PConv, we further propose FasterNet, a new family of neural networks, which attains substantially higher running speed than others on a wide range of devices, without compromising on accuracy for various vision tasks. For example, on ImageNet-lk, our tiny FasterNet-TO is 2.8×, 3.3×, and 2.4× faster than MobileViT-XXS on GPU, CPU, and ARM processors, respectively, while being 2.9% more accurate. Our large FasterNet-L achieves impressive 83.5% top-1 accuracy, on par with the emerging Swin-B, while having 36% higher inference throughput on GPU, as well as saving 37% compute time on CPU. Code is available at https://github.com/JierunChen/FasterNet.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/a3aa1323a7f08c40207eaa359041e5bd72b25b27

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. G. Wang
2. Xinchao Wang

In Computer Vision and Pattern Recognition

1. Xinchao Wang
2. Weihao Yu
3. Pengyu Guo
4. Yue Bai
5. Hanyu Hong
6. Pei An
7. G. Wang
8. Junhua Yan

Citations in the corpus

• Rewrite the Stars (2024, positive): “Capitalizing on the efficacy of the star operation, our StarNet can even surpasses various meticulously designed efficient models, like MobileNetv3 [27], EdgeViT [42], FasterNet [6], etc .”
• MambaOut: Do We Really Need Mamba for Vision?* (2024, positive): “Besides, to improve the practical speed, we only conduct depthwise convolution on partial channels [9, 60, 101], following InceptionNeXt [101].”
• SCSA: Exploring the Synergistic Effects Between Spatial and Channel Attention (2024, neutral): “2.”
• UAV-YOLOv8: A Small-Object-Detection Model Based on Improved YOLOv8 for UAV Aerial Photography Scenarios (2023, methodology): “[29] proposed a simple nd efficie t convoluti : partial convolution (PConv).”
• FFCA-YOLO for Small Object Detection in Remote Sensing Images (2024, methodology): “To alleviate this phenomenon, FasterNet uses PConv, which considers the redundancy in feature maps [51], and applies standard convolution on only a portion of input channels.”