Heedless Backbones

ViT-5 Family

Paper | GitHub

axis

Select an option

Results

Parameters (M)

Images / Second

Publication Date

task

Select an option

---------

Object Detection

Instance Segmentation

Classification

Semantic Segmentation

Panoptic Segmentation

dataset

Select an option

---------

ImageNet-1k

ImageNet-A

ImageNet-R

ImageNet-Sketch

ImageNet-C

ImageNet-C-bar

ImageNet-V2

ImageNet-ReaL

PASCAL VOC 2007 (val)

PASCAL VOC 2007 (test)

metric

Select an option

----------

Top-1

Top-5

GFLOPs

resolution

Select an option

----------

224x224

384x384

512x512

axis

Select an option

Results

Parameters (M)

Images / Second

GFLOPs

Publication Date

Filters

Legend

Legend attribute

Select an option

Family

Pretrain Dataset

Classification Resolution

Legend attribute (second)

Select an option

----------

Family

Pretrain Method

Classification Resolution

model	params (m)	pretrain	finetune	GFLOPs	Top-1
ViT-5-S	22.0	IN-1k : Sup. : 800	IN-1k : 20 : 224	4.7	82.16
ViT-5-B	87.0	IN-1k : Sup. : 800	IN-1k : 20 : 224	17.9	84.16
ViT-5-B	87.0	IN-1k : Sup. : 800	IN-1k : 20 : 384	55.9	85.4
ViT-5-L	304.0	IN-1k : Sup. : 400	IN-1k : 20 : 224	62.8	84.86
ViT-5-L	304.0	IN-1k : Sup. : 400	IN-1k : 20 : 384	192.5	86.0

model	params (m)	pretrain	finetune	gflops	IN-1k
ViT-5-S	22.0	IN-1k : Sup. : 800	IN-1k : 20 : 224	4.7	82.16/—
ViT-5-B	87.0	IN-1k : Sup. : 800	IN-1k : 20 : 224	17.9	84.16/—
ViT-5-B	87.0	IN-1k : Sup. : 800	IN-1k : 20 : 384	55.9	85.4/—
ViT-5-L	304.0	IN-1k : Sup. : 400	IN-1k : 20 : 224	62.8	84.86/—
ViT-5-L	304.0	IN-1k : Sup. : 400	IN-1k : 20 : 384	192.5	86.0/—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
ViT-5-S	IN-1k : Sup. : 800	UPerNet	ADE20K (train) : 128 : 512	None	—	—	—	47.5	—	—
ViT-5-B	IN-1k : Sup. : 800	UPerNet	ADE20K (train) : 128 : 512	None	—	—	—	49.1	—	—
ViT-5-L	IN-1k : Sup. : 400	UPerNet	ADE20K (train) : 128 : 512	None	—	—	—	52.0	—	—