Statements (70)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:Transformers
gptkb:neural_networks |
| gptkbp:applies_to |
image classification
object detection image segmentation semantic segmentation |
| gptkbp:architectural_style |
gptkb:social_hierarchy
|
| gptkbp:based_on |
gptkb:Transformers
shifted windows |
| gptkbp:competes_with |
gptkb:Vision_Transformers
|
| gptkbp:developed_by |
gptkb:Microsoft_Research
|
| gptkbp:has |
high accuracy
multiple variants |
| gptkbp:has_achieved |
state-of-the-art results
state-of-the-art performance |
| gptkbp:has_variants |
gptkb:Swin-B
gptkb:Swin-L gptkb:Swin-T Swin-S |
| https://www.w3.org/2000/01/rdf-schema#label |
Swin Transformer
|
| gptkbp:improves |
computational efficiency
image classification performance |
| gptkbp:influences |
subsequent models
|
| gptkbp:introduced_in |
gptkb:2021
|
| gptkbp:is |
open-source
scalable based on a hierarchical structure designed for dense prediction tasks flexible for different tasks part of the family of vision transformers widely adopted in research |
| gptkbp:is_adopted_by |
gptkb:academic_research
industry applications |
| gptkbp:is_cited_in |
numerous research papers
|
| gptkbp:is_considered_as |
a leading model in vision tasks
|
| gptkbp:is_documented_in |
ar Xiv preprint
|
| gptkbp:is_evaluated_by |
gptkb:COCO_dataset
gptkb:Image_Net_dataset gptkb:ADE20_K_dataset real-world scenarios benchmark competitions |
| gptkbp:is_implemented_in |
gptkb:Tensor_Flow
gptkb:Py_Torch |
| gptkbp:is_known_for |
scalability
efficient computation flexible windowing mechanism |
| gptkbp:is_optimized_for |
large-scale datasets
|
| gptkbp:is_part_of |
CVPR 2021 conference
Vision Transformer family |
| gptkbp:is_recognized_by |
AI research community
|
| gptkbp:is_related_to |
gptkb:neural_networks
gptkb:Vision_Transformers Attention Mechanisms |
| gptkbp:is_supported_by |
community contributions
open-source implementations |
| gptkbp:is_trained_in |
gptkb:COCO_2017
gptkb:Image_Net-1_K gptkb:ADE20_K_dataset |
| gptkbp:is_used_in |
transfer learning
self-supervised learning |
| gptkbp:performance |
image classification tasks
object detection tasks image segmentation tasks |
| gptkbp:supports |
multi-scale feature extraction
hierarchical feature representation |
| gptkbp:used_for |
computer vision tasks
|
| gptkbp:utilizes |
shifted windows
local and global attention mechanisms |
| gptkbp:bfsParent |
gptkb:Transformers
|
| gptkbp:bfsLayer |
4
|