Heedless Backbones

EViT 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
EViT-Tiny	11.8	IN-1k : Sup. : 300	— : — : —	2.1	79.9
EViT-Small	24.9	IN-1k : Sup. : 300	— : — : —	4.1	82.6
EViT-Base	42.8	IN-1k : Sup. : 300	— : — : —	7.2	83.9
EViT-Large	61.9	IN-1k : Sup. : 300	— : — : —	10.3	84.4
EViT-Large	61.9	IN-1k : Sup. : 300	— : — : —	13.7	84.9

model	params (m)	pretrain	finetune	gflops	IN-1k	IN-C↓
EViT-Tiny	11.8	IN-1k : Sup. : 300	— : — : —	2.1	79.9/—	—/—
EViT-Small	24.9	IN-1k : Sup. : 300	— : — : —	4.1	82.6/—	52.4/—
EViT-Base	42.8	IN-1k : Sup. : 300	— : — : —	7.2	83.9/—	—/—
EViT-Large	61.9	IN-1k : Sup. : 300	— : — : —	10.3	84.4/—	—/—
EViT-Large	61.9	IN-1k : Sup. : 300	— : — : —	13.7	84.9/—	—/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
EViT-Small	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	45.1	66.0	48.4	28.3	49.2	59.7
EViT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	46.0	67.6	50.3	—	—	—
EViT-Base	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	46.5	67.5	49.8	29.4	51.3	62.1
EViT-Base	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	47.5	68.8	52.3	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
EViT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	41.7	64.8	44.9	—	—	—
EViT-Base	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	43.1	66.3	46.3	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
EViT-Small	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	44.8	—	—	—	46.1	—	—
EViT-Base	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	60.1	—	—	—	48.5	—	—