Heedless Backbones

MFil-Mamba 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
MFil-Mamba-T	33.46	IN-1k : Sup. : 300	— : — : —	5.56	83.22
MFil-Mamba-S	50.6	IN-1k : Sup. : 300	— : — : —	9.1	83.9
MFil-Mamba-B	93.1	IN-1k : Sup. : 300	— : — : —	16.8	84.2

model	params (m)	pretrain	finetune	gflops	IN-1k
MFil-Mamba-T	33.46	IN-1k : Sup. : 300	— : — : —	5.56	83.22/—
MFil-Mamba-S	50.6	IN-1k : Sup. : 300	— : — : —	9.1	83.9/—
MFil-Mamba-B	93.1	IN-1k : Sup. : 300	— : — : —	16.8	84.2/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
MFil-Mamba-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	261.0	47.3	69.2	51.9	—	—	—
MFil-Mamba-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	320.0	47.9	69.4	52.3	—	—	—
MFil-Mamba-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	467.0	49.0	70.6	53.7	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
MFil-Mamba-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	261.0	42.7	66.2	46.0	—	—	—
MFil-Mamba-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	320.0	43.1	66.8	46.4	—	—	—
MFil-Mamba-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	467.0	43.8	67.7	47.6	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
MFil-Mamba-T	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	961.0	48.9	—	—	48.5	—	—
MFil-Mamba-S	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1028.0	49.9	—	—	49.3	—	—
MFil-Mamba-B	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1198.0	51.3	—	—	50.6	—	—