Heedless Backbones

HTC++

Paper | GitHub

axis

Select an option

Results

Parameters (M)

Images / Second

Publication Date

task

Select an option

---------

Object Detection

Instance Segmentation

dataset

Select an option

---------

COCO (val)

COCO (test)

Cityscapes (val)

Cityscapes (test)

ADE20K (val)

ADE20K (test)

metric

Select an option

----------

mAP

AP₅₀

AP₇₅

mAP_s

mAP_m

mAP_l

GFLOPs

axis

Select an option

Results

Parameters (M)

Images / Second

GFLOPs

Publication Date

Filters

pretrain dataset

Select an option

---------

JFT-3B

ImageNet-1k

ImageNet-22k

JFT-300M

pretrain method

Select an option

----------

Supervised

FCMAE

MAE

CL

Legend

Legend attribute

Select an option

Family

Pretrain Dataset

Pretrain Method

Instance Head

Instance Training Epochs

Legend attribute (second)

Select an option

----------

Pretrain Dataset

Pretrain Method

Instance Head

Instance Training Epochs

family	model	params (m)	pretrain	head	train	GFLOPs	mAP
Swin	Swin-B	88.0	IN-22k : Sup. : 90	HTC++	COCO (train) : 72	1043.0	49.1
Swin	Swin-L	197.0	IN-22k : Sup. : 90	HTC++	COCO (train) : 72	1470.0	49.5

COCO (test)

family	model	pretrain	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
Swin	Swin-L	IN-22k : Sup. : 90	COCO (train) : 72	1470.0	57.7	—	—	—	—	—

COCO (val)

family	model	pretrain	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
Swin	Swin-B	IN-22k : Sup. : 90	COCO (train) : 72	1043.0	56.4	—	—	—	—	—
Swin	Swin-L	IN-22k : Sup. : 90	COCO (train) : 72	1470.0	57.1	—	—	—	—	—

COCO (test)

family	model	pretrain	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
Swin	Swin-L	IN-22k : Sup. : 90	COCO (train) : 72	1470.0	50.2	—	—	—	—	—

COCO (val)

family	model	pretrain	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
Swin	Swin-B	IN-22k : Sup. : 90	COCO (train) : 72	1043.0	49.1	—	—	—	—	—
Swin	Swin-L	IN-22k : Sup. : 90	COCO (train) : 72	1470.0	49.5	—	—	—	—	—