|
gptkbp:instance_of
|
gptkb:Model
|
|
gptkbp:application
|
Image Classification
|
|
gptkbp:architectural_style
|
gptkb:Transformers_character
|
|
gptkbp:author
|
gptkb:Yuan_Li
gptkb:Hao_Zhang
Jingdong Wang
Tong He
Xiaogang Wang
Yunhua Zhang
Ze Liu
|
|
gptkbp:cache_size
|
gptkb:32
|
|
gptkbp:community_support
|
Active
|
|
gptkbp:developed_by
|
gptkb:Microsoft_Research
|
|
gptkbp:features
|
Hierarchical Feature Maps
|
|
gptkbp:focus
|
Shifted Window Attention
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Swin-V
|
|
gptkbp:image
|
RGB
|
|
gptkbp:input_output
|
Class Labels
224x224
|
|
gptkbp:is_a_framework_for
|
gptkb:Py_Torch
|
|
gptkbp:is_a_hub_for
|
over 10,000
over 1,000
|
|
gptkbp:is_evaluated_by
|
F1 Score
Mean Average Precision
|
|
gptkbp:language
|
gptkb:Library
|
|
gptkbp:license
|
gptkb:Apache_2.0
|
|
gptkbp:migration
|
Supported
|
|
gptkbp:orbital_period
|
over 100 million
|
|
gptkbp:performance
|
Top-1 Accuracy
|
|
gptkbp:predecessor
|
gptkb:Swin-T
|
|
gptkbp:provides_information_on
|
gptkb:Image_Net
gptkb:COCO
gptkb:battle
AD E20 K
|
|
gptkbp:reduces
|
0.1
|
|
gptkbp:related_model
|
gptkb:Swin-L
gptkb:Swin-T
gptkb:Dei_T
Large
Vision Transformer (Vi T)
|
|
gptkbp:release_year
|
gptkb:2021
|
|
gptkbp:sister_channel
|
gptkb:television_series
gptkb:3
|
|
gptkbp:successor
|
gptkb:Swin-L
|
|
gptkbp:system_requirements
|
gptkb:CEO
|
|
gptkbp:title
|
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
|
|
gptkbp:training
|
gptkb:Graphics_Processing_Unit
Supervised Learning
0.001
several days
|
|
gptkbp:tuning
|
gptkb:Adam
Supported
|
|
gptkbp:use_case
|
Object Detection
Instance Segmentation
Semantic Segmentation
|
|
gptkbp:bfsParent
|
gptkb:Swin-L
|
|
gptkbp:bfsLayer
|
6
|