Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366207485> ?p ?o ?g. }
Showing items 1 to 61 of
61
with 100 items per page.
- W4366207485 abstract "We introduce Dynamic Mobile-Former(DMF), maximizes the capabilities of dynamic convolution by harmonizing it with efficient operators.Our Dynamic MobileFormer effectively utilizes the advantages of Dynamic MobileNet (MobileNet equipped with dynamic convolution) using global information from light-weight attention.A Transformer in Dynamic Mobile-Former only requires a few randomly initialized tokens to calculate global features, making it computationally efficient.And a bridge between Dynamic MobileNet and Transformer allows for bidirectional integration of local and global features.We also simplify the optimization process of vanilla dynamic convolution by splitting the convolution kernel into an input-agnostic kernel and an input-dependent kernel.This allows for optimization in a wider kernel space, resulting in enhanced capacity.By integrating lightweight attention and enhanced dynamic convolution, our Dynamic Mobile-Former achieves not only high efficiency, but also strong performance.We benchmark the Dynamic Mobile-Former on a series of vision tasks, and showcase that it achieves impressive performance on image classification, COCO detection, and instanace segmentation.For example, our DMF hits the top-1 accuracy of 79.4% on ImageNet-1K, much higher than PVT-Tiny by 4.3% with only 1/4 FLOPs.Additionally,our proposed DMF-S model performed well on challenging vision datasets such as COCO, achieving a 39.0% mAP,which is 1% higher than that of the Mobile-Former 508M model, despite using 3 GFLOPs less computations.Code and models are available at https://github.com/ysj9909/DMF" @default.
- W4366207485 created "2023-04-19" @default.
- W4366207485 creator A5066092248 @default.
- W4366207485 creator A5078628073 @default.
- W4366207485 date "2023-04-13" @default.
- W4366207485 modified "2023-09-30" @default.
- W4366207485 title "Dynamic Mobile-Former: Strengthening Dynamic Convolution with Attention and Residual Connection in Kernel Space" @default.
- W4366207485 doi "https://doi.org/10.48550/arxiv.2304.07254" @default.
- W4366207485 hasPublicationYear "2023" @default.
- W4366207485 type Work @default.
- W4366207485 citedByCount "0" @default.
- W4366207485 crossrefType "posted-content" @default.
- W4366207485 hasAuthorship W4366207485A5066092248 @default.
- W4366207485 hasAuthorship W4366207485A5078628073 @default.
- W4366207485 hasBestOaLocation W43662074851 @default.
- W4366207485 hasConcept C11413529 @default.
- W4366207485 hasConcept C114614502 @default.
- W4366207485 hasConcept C13280743 @default.
- W4366207485 hasConcept C154945302 @default.
- W4366207485 hasConcept C155512373 @default.
- W4366207485 hasConcept C173608175 @default.
- W4366207485 hasConcept C185798385 @default.
- W4366207485 hasConcept C205649164 @default.
- W4366207485 hasConcept C33923547 @default.
- W4366207485 hasConcept C3826847 @default.
- W4366207485 hasConcept C41008148 @default.
- W4366207485 hasConcept C45347329 @default.
- W4366207485 hasConcept C45374587 @default.
- W4366207485 hasConcept C50644808 @default.
- W4366207485 hasConcept C74193536 @default.
- W4366207485 hasConceptScore W4366207485C11413529 @default.
- W4366207485 hasConceptScore W4366207485C114614502 @default.
- W4366207485 hasConceptScore W4366207485C13280743 @default.
- W4366207485 hasConceptScore W4366207485C154945302 @default.
- W4366207485 hasConceptScore W4366207485C155512373 @default.
- W4366207485 hasConceptScore W4366207485C173608175 @default.
- W4366207485 hasConceptScore W4366207485C185798385 @default.
- W4366207485 hasConceptScore W4366207485C205649164 @default.
- W4366207485 hasConceptScore W4366207485C33923547 @default.
- W4366207485 hasConceptScore W4366207485C3826847 @default.
- W4366207485 hasConceptScore W4366207485C41008148 @default.
- W4366207485 hasConceptScore W4366207485C45347329 @default.
- W4366207485 hasConceptScore W4366207485C45374587 @default.
- W4366207485 hasConceptScore W4366207485C50644808 @default.
- W4366207485 hasConceptScore W4366207485C74193536 @default.
- W4366207485 hasLocation W43662074851 @default.
- W4366207485 hasOpenAccess W4366207485 @default.
- W4366207485 hasPrimaryLocation W43662074851 @default.
- W4366207485 hasRelatedWork W17610788 @default.
- W4366207485 hasRelatedWork W1979781211 @default.
- W4366207485 hasRelatedWork W2093683727 @default.
- W4366207485 hasRelatedWork W2564509292 @default.
- W4366207485 hasRelatedWork W280889461 @default.
- W4366207485 hasRelatedWork W2921438861 @default.
- W4366207485 hasRelatedWork W2992221004 @default.
- W4366207485 hasRelatedWork W401777774 @default.
- W4366207485 hasRelatedWork W4288417960 @default.
- W4366207485 hasRelatedWork W4315882045 @default.
- W4366207485 isParatext "false" @default.
- W4366207485 isRetracted "false" @default.
- W4366207485 workType "article" @default.