MobileNetV2: Inverted Residuals and Linear Bottlenecks
By Mark Sandler, Andrew G. Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen. 2018
25,197 citations recorded by the source corpus.
In this paper we describe a new mobile architecture, MobileNetV2, that improves the state of the art performance of mobile models on multiple tasks and benchmarks as well as across a spectrum of different model sizes. We also describe efficient ways of applying these mobile models to object detection in a novel framework we call SSDLite. Additionally, we demonstrate how to build mobile semantic segmentation models through a reduced form of DeepLabv3 which we call Mobile DeepLabv3. is based on an inverted residual structure where the shortcut connections are between the thin bottleneck layers. The intermediate expansion layer uses lightweight depthwise convolutions to filter features as a source of non-linearity. Additionally, we find that it is important to remove non-linearities in the narrow layers in order to maintain representational power. We demonstrate that this improves performance and provide an intuition that led to this design. Finally, our approach allows decoupling of the input/output domains from the expressiveness of the transformation, which provides a convenient framework for further analysis. We measure our performance on ImageNet [1] classification, COCO object detection [2], VOC image segmentation [3]. We evaluate the trade-offs between accuracy, and number of operations measured by multiply-adds (MAdd), as well as actual latency, and the number of parameters.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Efficiency & Compression
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Quoc V. Le
- Mohit Bansal
- Jianfeng Gao
- Dacheng Tao
- Yoshua Bengio
- Trevor Darrell
- Sergey Levine
- Percy Liang
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Ming-Hsuan Yang
- Xiaogang Wang
- Trevor Darrell
- Yi Yang
- Alan L. Yuille
- Jitendra Malik
In Efficiency & Compression
Citations in the corpus
- Do Better ImageNet Models Transfer Better? (2018, methodology): “These networks encompassed widely used Inception architectures [77, 39, 78, 76]; ResNets [33, 30, 29]; DenseNets [36]; MobileNets [35, 68]; and NASNets [92].”
- EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks (2019, methodology): “Its main building block is mobile inverted bottleneck MBConv (Sandler et al., 2018; Tan et al., 2019), to which we also add squeeze-and-excitation optimization (Hu et al.”
- Visual Transformers: Token-based Image Representation and Processing for Computer Vision (2020, neutral): “Early work in this direction includes [23, 31, 13, 17, 32, 16, 48, 27, 43].”
- Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation (2020, methodology): “Additionally, we multiply all the channels [34,69,33] by 0.5, 0.75, and 2, resulting in Axial- ResNet-{ S, M, XL } , respectively.”
- Stand-Alone Self-Attention in Vision Models (2019, background): “Concurrently, [33] also see significant gains in object detection and image classification through augmenting convolutional features with global self-attention features.”
- Attention Augmented Convolutional Networks (2019, methodology): “In particular, the MnasNet was found by neural architecture search [54], using only the highly optimized mobile inverted bottleneck block [36] and the Squeeze-andExcitation operation [17] as the primitives in its search space.”
- Measuring Robustness to Natural Distribution Shifts in Image Classification (2020, neutral): “mobilenet_v2 [73].”
- Making Convolutional Networks Shift-Invariant Again (2019, methodology): “, 2016) and MobileNet (Sandler et al., 2018).”
- Measuring the Algorithmic Efficiency of Neural Networks (2020, methodology): “…2014, He et al., 2015, Xie et al., 2016, Huang et al., 2016, Iandola et al., 2016, Zagoruyko & Komodakis, 2016,Zhang et al., 2017,Howard et al., 2017,Sandler et al., 2018,Ma et al., 2018,Tan & Le, 2019] we used PyTorch’s example models [Paszke et al., 2017] with Pytorch’s suggested hyperparameters.”
- ContextNet: Improving Convolutional Neural Networks for Automatic Speech Recognition with Global Context (2020, methodology): “In this paper, we choose depthwise separable convolution as conv(·), because such a design has been previously shown in various applications [6, 4, 28] to achieve better parameter efficiency without impacting accuracy.”
- MERLOT: Multimodal Neural Script Knowledge Models (2021, methodology): “6 We trained a lightweight MobileNet-V2 CNN [93] to score whether a COCO object class is present in an image or not, using a sigmoid cross entropy loss.”
- High-Performance Large-Scale Image Recognition Without Normalization (2021, methodology): “The current state of the art on image classification is generally held by the EfficientNet family of models (Tan & Le, 2019), which are based on a variant of inverted bottleneck blocks (Sandler et al., 2018) with a backbone and model scaling strategy derived from neural architecture search.”
- Reinforce Data, Multiply Impact: Improved Model Accuracy and Robustness with Dataset Reinforcement (2023, positive): “We study light-weight CNN-based (MobileNetV1 [26]/ V2 [50]/ V3[25]), heavy-weight CNN-based (ResNet [21] and EfficientNet [58]), and transformer-based (ViT [16] and SwinTransformer [38]) models.”
- Designing Network Design Spaces (2020, methodology): “We observe that REGNETS are surprisingly effective in this regime considering the substantial body of work on finding better mobile networks via both manual design [7, 21, 16] and NAS [30, 19, 14, 15].”
- Green AI (2019, background): “For example, a significant amount of work in the computer vision community has addressed efficient inference, which is necessary for real-time processing of images for applications like self-driving cars [23, 30, 22], or for placing models on devices such as mobile phones [16, 33].”
- MnasNet: Platform-Aware Neural Architecture Search for Mobile (2018, methodology): “Such increases in computational demands make it difficult to deploy state-of-the-art CNN models on resource-constrained platforms [29, 36, 26] on ImageNet.”