MobileDets: Searching for Object Detection Architectures for Mobile Accelerators

Abstract

Inverted bottleneck layers, which are built upon depthwise convolutions, have been the predominant building blocks in state-of-the-art object detection models on mobile devices. In this work, we investigate the optimality of this design pattern over a broad range of mobile accelerators by revisiting the usefulness of regular convolutions. We discover that regular convolutions are a potent component to boost the latency-accuracy trade-off for object detection on accelerators, provided that they are placed strategically in the network via neural architecture search. By incorporating regular convolutions in the search space and directly optimizing the network architectures for object detection, we obtain a family of object detection models, MobileDets, that achieve state-of-the-art results across mobile accelerators. On the COCO object detection task, MobileDets outperform MobileNetV3+SSDLite by 1.7 mAP at comparable mobile CPU inference latencies. MobileDets also outperform MobileNetV2+SSDLite by 1.9 mAP on mobile CPUs, 3.7 mAP on Google EdgeTPU, 3.4 mAP on Qualcomm Hexagon DSP and 2.7 mAP on Nvidia Jetson GPU without increasing latency. Moreover, MobileDets are comparable with the state-of-the-art MnasFPN on mobile CPUs even without using the feature pyramid, and achieve better mAP scores on both EdgeTPUs and DSPs with up to 2x speedup. Code and models are available in the TensorFlow Object Detection API: https://github.com/tensorflow/models/tree/master/research/object_detection.

Accepted at CVPR 2021; Code and models are available in the TensorFlow Object Detection API: https://github.com/tensorflow/models/tree/master/research/object_detection

References in corpus (12)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
• Neural Architecture Search with Reinforcement Learning
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• AMC: AutoML for Model Compression and Acceleration on Mobile Devices
• Efficient Neural Architecture Search via Parameter Sharing
• SMASH: One-Shot Model Architecture Search through HyperNetworks
• Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decomposition
• Accelerator-aware Neural Network Design using AutoML
• NAS-FCOS: Fast Neural Architecture Search for Object Detection
• ANTNets: Mobile Convolutional Neural Networks for Resource Efficient Image Classification

Cited by in corpus (10)

• EfficientNetV2: Smaller Models and Faster Training
• A Comprehensive Survey on Hardware-Aware Neural Architecture Search
• Weight-Sharing Neural Architecture Search: A Battle to Shrink the Optimization Gap
• HW-NAS-Bench:Hardware-Aware Neural Architecture Search Benchmark
• Characterizing signal propagation to close the performance gap in unnormalized ResNets
• NeBula: Quest for Robotic Autonomy in Challenging Environments; TEAM CoSTAR at the DARPA Subterranean Challenge
• Rethinking Text Line Recognition Models
• Rethinking Co-design of Neural Architectures and Hardware Accelerators
• FBNetV5: Neural Architecture Search for Multiple Tasks in One Run
• NAS-FCOS: Efficient Search for Object Detection Architectures