Matches in SemOpenAlex for { <https://semopenalex.org/work/W3146097248> ?p ?o ?g. }
- W3146097248 abstract "Transformers have been recently adapted for large scale image classification, achieving high scores shaking up the long supremacy of convolutional neural networks. However the optimization of image transformers has been little studied so far. In this work, we build and optimize deeper transformer networks for image classification. In particular, we investigate the interplay of architecture and optimization of such dedicated transformers. We make two transformers architecture changes that significantly improve the accuracy of deep transformers. This leads us to produce models whose performance does not saturate early with more depth, for instance we obtain 86.5% top-1 accuracy on Imagenet when training with no external data, we thus attain the current SOTA with less FLOPs and parameters. Moreover, our best model establishes the new state of the art on Imagenet with Reassessed labels and Imagenet-V2 / match frequency, in the setting with no additional training data. We share our code and models." @default.
- W3146097248 created "2021-04-13" @default.
- W3146097248 creator A5022871131 @default.
- W3146097248 creator A5032644449 @default.
- W3146097248 creator A5033867133 @default.
- W3146097248 creator A5041907084 @default.
- W3146097248 creator A5067991583 @default.
- W3146097248 date "2021-03-31" @default.
- W3146097248 modified "2023-10-01" @default.
- W3146097248 title "Going deeper with Image Transformers" @default.
- W3146097248 cites W1026270304 @default.
- W3146097248 cites W1533861849 @default.
- W3146097248 cites W1690739335 @default.
- W3146097248 cites W1836465849 @default.
- W3146097248 cites W2097073572 @default.
- W3146097248 cites W2097117768 @default.
- W3146097248 cites W2108598243 @default.
- W3146097248 cites W2117539524 @default.
- W3146097248 cites W2138011018 @default.
- W3146097248 cites W2163605009 @default.
- W3146097248 cites W2194775991 @default.
- W3146097248 cites W2331143823 @default.
- W3146097248 cites W2533598788 @default.
- W3146097248 cites W2549139847 @default.
- W3146097248 cites W2601564443 @default.
- W3146097248 cites W2626778328 @default.
- W3146097248 cites W2752782242 @default.
- W3146097248 cites W2767206889 @default.
- W3146097248 cites W2768282280 @default.
- W3146097248 cites W2805003733 @default.
- W3146097248 cites W2907121943 @default.
- W3146097248 cites W2911925209 @default.
- W3146097248 cites W2922509574 @default.
- W3146097248 cites W2937843571 @default.
- W3146097248 cites W2940744433 @default.
- W3146097248 cites W2946948417 @default.
- W3146097248 cites W2949427019 @default.
- W3146097248 cites W2950181225 @default.
- W3146097248 cites W2950541952 @default.
- W3146097248 cites W2950739196 @default.
- W3146097248 cites W2953937638 @default.
- W3146097248 cites W2962835968 @default.
- W3146097248 cites W2963091558 @default.
- W3146097248 cites W2963139417 @default.
- W3146097248 cites W2963341956 @default.
- W3146097248 cites W2964065616 @default.
- W3146097248 cites W2965373594 @default.
- W3146097248 cites W2970389371 @default.
- W3146097248 cites W2971315489 @default.
- W3146097248 cites W2972818416 @default.
- W3146097248 cites W2975381464 @default.
- W3146097248 cites W2981413347 @default.
- W3146097248 cites W2991213871 @default.
- W3146097248 cites W2993466051 @default.
- W3146097248 cites W3010768098 @default.
- W3146097248 cites W3012324658 @default.
- W3146097248 cites W3016719260 @default.
- W3146097248 cites W3017022649 @default.
- W3146097248 cites W3030163527 @default.
- W3146097248 cites W3033210410 @default.
- W3146097248 cites W3034363135 @default.
- W3146097248 cites W3034421924 @default.
- W3146097248 cites W3034429256 @default.
- W3146097248 cites W3034885317 @default.
- W3146097248 cites W3035452548 @default.
- W3146097248 cites W3035618017 @default.
- W3146097248 cites W3035743198 @default.
- W3146097248 cites W3035794324 @default.
- W3146097248 cites W3093533780 @default.
- W3146097248 cites W3093579165 @default.
- W3146097248 cites W3096609285 @default.
- W3146097248 cites W3103005696 @default.
- W3146097248 cites W3116489684 @default.
- W3146097248 cites W3118608800 @default.
- W3146097248 cites W3119786062 @default.
- W3146097248 cites W3120885796 @default.
- W3146097248 cites W3121334524 @default.
- W3146097248 cites W3121523901 @default.
- W3146097248 cites W3128099838 @default.
- W3146097248 cites W3128633047 @default.
- W3146097248 cites W3133696297 @default.
- W3146097248 cites W3172509117 @default.
- W3146097248 cites W3180562345 @default.
- W3146097248 hasPublicationYear "2021" @default.
- W3146097248 type Work @default.
- W3146097248 sameAs 3146097248 @default.
- W3146097248 citedByCount "67" @default.
- W3146097248 countsByYear W31460972482020 @default.
- W3146097248 countsByYear W31460972482021 @default.
- W3146097248 countsByYear W31460972482022 @default.
- W3146097248 crossrefType "posted-content" @default.
- W3146097248 hasAuthorship W3146097248A5022871131 @default.
- W3146097248 hasAuthorship W3146097248A5032644449 @default.
- W3146097248 hasAuthorship W3146097248A5033867133 @default.
- W3146097248 hasAuthorship W3146097248A5041907084 @default.
- W3146097248 hasAuthorship W3146097248A5067991583 @default.
- W3146097248 hasConcept C119599485 @default.
- W3146097248 hasConcept C119857082 @default.
- W3146097248 hasConcept C123657996 @default.
- W3146097248 hasConcept C127413603 @default.