{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/mobilevitv3-mobile-friendly-vision","title":"MobileViTv3: Mobile-Friendly Vision Transformer with Simple and Effective Fusion of Local, Global and Input Features","arxiv_id":"2209.15159","date":"2022-09-30","proceeding":null,"authors":["Shakti N. Wadekar","Abhishek Chaurasia"],"abstract":"MobileViT (MobileViTv1) combines convolutional neural networks (CNNs) and vision transformers (ViTs) to create light-weight models for mobile vision tasks. Though the main MobileViTv1-block helps to achieve competitive state-of-the-art results, the fusion block inside MobileViTv1-block, creates scaling challenges and has a complex learning task. We propose changes to the fusion block that are simple and effective to create MobileViTv3-block, which addresses the scaling and simplifies the learning task. Our proposed MobileViTv3-block used to create MobileViTv3-XXS, XS and S models outperform MobileViTv1 on ImageNet-1k, ADE20K, COCO and PascalVOC2012 datasets. On ImageNet-1K, MobileViTv3-XXS and MobileViTv3-XS surpasses MobileViTv1-XXS and MobileViTv1-XS by 2% and 1.9% respectively. Recently published MobileViTv2 architecture removes fusion block and uses linear complexity transformers to perform better than MobileViTv1. We add our proposed fusion block to MobileViTv2 to create MobileViTv3-0.5, 0.75 and 1.0 models. These new models give better accuracy numbers on ImageNet-1k, ADE20K, COCO and PascalVOC2012 datasets as compared to MobileViTv2. MobileViTv3-0.5 and MobileViTv3-0.75 outperforms MobileViTv2-0.5 and MobileViTv2-0.75 by 2.1% and 1.0% respectively on ImageNet-1K dataset. For segmentation task, MobileViTv3-1.0 achieves 2.07% and 1.1% better mIOU compared to MobileViTv2-1.0 on ADE20K dataset and PascalVOC2012 dataset respectively. Our code and the trained models are available at: https://github.com/micronDLA/MobileViTv3","url_abs":"https://arxiv.org/abs/2209.15159v2","url_pdf":"https://arxiv.org/pdf/2209.15159v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mobilevitv3-mobile-friendly-vision","repo_url":"https://github.com/microndla/mobilevitv3","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"mobilevitv3-mobile-friendly-vision","repo_url":"https://github.com/jaiwei98/mobile-vit-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"depthwise-convolution","method_name":"Depthwise Convolution"},{"method_slug":"depthwise-separable-convolution","method_name":"Depthwise Separable Convolution"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"mobilevitv2","method_name":"MobileViTv2"},{"method_slug":"mobilevitv3","method_name":"MobileViTv3"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"pointwise-convolution","method_name":"Pointwise Convolution"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"vision-transformer","method_name":"Vision Transformer"}],"datasets_introduced":[],"methods_introduced":[{"slug":"mobilevitv3","name":"MobileViTv3","full_name":"MobileViTv3"}],"results":[{"leaderboard":"/sota/on-1","task":"","dataset":"","model":"","rank_in_archive_order":6,"of":6,"metrics":{},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-S","rank_in_archive_order":765,"of":1060,"metrics":{"GFLOPs":"1.841","Number of params":"5.8M","Top 1 Accuracy":"79.3%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-1.0","rank_in_archive_order":815,"of":1060,"metrics":{"GFLOPs":"1.876","Top 1 Accuracy":"78.64%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-XS","rank_in_archive_order":902,"of":1060,"metrics":{"GFLOPs":"0.927","Number of params":"2.5M","Top 1 Accuracy":"76.7%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-0.75","rank_in_archive_order":911,"of":1060,"metrics":{"GFLOPs":"1.064","Number of params":"3M","Top 1 Accuracy":"76.55%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-0.5","rank_in_archive_order":998,"of":1060,"metrics":{"GFLOPs":"0.481","Number of params":"1.4M","Top 1 Accuracy":"72.33%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"MobileViTv3-XXS","rank_in_archive_order":1015,"of":1060,"metrics":{"GFLOPs":"0.289","Number of params":"1.2M","Top 1 Accuracy":"70.98%"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2209.15159","atlas_url":"https://app.syntology.ai/?focus=2209.15159","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}