Heedless Backbones

QuadMamba 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
QuadMamba-Li	5.47	IN-1k : Sup. : 300	— : — : —	0.82	74.2
QuadMamba-T	10.32	IN-1k : Sup. : 300	— : — : —	2.07	78.2
QuadMamba-S	31.25	IN-1k : Sup. : 300	— : — : —	5.51	82.4
QuadMamba-B	50.6	IN-1k : Sup. : 300	— : — : —	9.3	83.8

model	params (m)	pretrain	finetune	gflops	IN-1k
QuadMamba-Li	5.47	IN-1k : Sup. : 300	— : — : —	0.82	74.2/92.1
QuadMamba-T	10.32	IN-1k : Sup. : 300	— : — : —	2.07	78.2/94.3
QuadMamba-S	31.25	IN-1k : Sup. : 300	— : — : —	5.51	82.4/95.6
QuadMamba-B	50.6	IN-1k : Sup. : 300	— : — : —	9.3	83.8/96.7

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
QuadMamba-Li	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	186.0	39.3	61.7	42.4	—	—	—
QuadMamba-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	213.0	42.3	64.6	46.2	—	—	—
QuadMamba-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	301.0	46.7	69.0	51.3	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
QuadMamba-Li	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	186.0	36.9	58.8	39.4	—	—	—
QuadMamba-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	213.0	38.8	61.6	41.4	—	—	—
QuadMamba-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	301.0	42.4	65.9	45.6	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
QuadMamba-T	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	886.0	45.1	—	—	44.3	—	—
QuadMamba-S	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	961.0	48.1	—	—	47.2	—	—
QuadMamba-B	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1042.0	50.8	—	—	49.7	—	—