CvT: Introducing Convolutions to Vision Transformers
Author:
Affiliation:
1. McGill University
2. Microsoft Cloud + AI
Publisher
IEEE
Link
http://xplorestaging.ieee.org/ielx7/9709627/9709628/09710031.pdf?arnumber=9710031
Reference44 articles.
1. Learning Texture Transformer Network for Image Super-Resolution
2. Lite transformer with long-short range attention;wu,2020
3. Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions
4. Max-deeplab: End-to-end panoptic segmentation with mask transformers;wang,2020
5. GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Cited by 885 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. SPA: Self-Peripheral-Attention for central–peripheral interactions in endoscopic image classification and segmentation;Expert Systems with Applications;2024-07
2. PPFGED: Federated learning for graphic element detection with privacy preservation in multi-source substation drawings;Expert Systems with Applications;2024-06
3. HumanPoseNet: An all-transformer architecture for pose estimation with efficient patch expansion and attentional feature refinement;Expert Systems with Applications;2024-06
4. A survey of multimodal hybrid deep learning for computer vision: Architectures, applications, trends, and challenges;Information Fusion;2024-05
5. Two-pathway spatiotemporal representation learning for extreme water temperature prediction;Engineering Applications of Artificial Intelligence;2024-05
1.学者识别学者识别
2.学术分析学术分析
3.人才评估人才评估
"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370
www.globalauthorid.com
TOP
Copyright © 2019-2024 北京同舟云网络信息技术有限公司 京公网安备11010802033243号 京ICP备18003416号-3