FasterViT 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
FasterViT-0	31.4	IN-1k : Sup. : 300	— : — : —	3.3	82.1
FasterViT-1	53.4	IN-1k : Sup. : 300	— : — : —	5.3	83.2
FasterViT-2	75.9	IN-1k : Sup. : 300	— : — : —	8.7	84.2
FasterViT-3	159.5	IN-1k : Sup. : 300	— : — : —	18.2	84.9
FasterViT-4	424.6	IN-1k : Sup. : 300	— : — : —	36.6	85.4
FasterViT-5	957.5	IN-1k : Sup. : 300	— : — : —	113.0	85.6
FasterViT-6	1360.0	IN-1k : Sup. : 300	— : — : —	142.0	85.8
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 224	36.6	86.6
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 384	119.2	87.5
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 512	213.5	87.8
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 768	480.4	87.9

model	params (m)	pretrain	finetune	gflops	IN-1k	IN-A	IN-R	IN-V2
FasterViT-0	31.4	IN-1k : Sup. : 300	— : — : —	3.3	82.1/95.9	23.9/—	45.9/—	70.9/90.0
FasterViT-1	53.4	IN-1k : Sup. : 300	— : — : —	5.3	83.2/96.5	31.2/—	47.5/—	72.6/91.0
FasterViT-2	75.9	IN-1k : Sup. : 300	— : — : —	8.7	84.2/96.8	38.2/—	49.6/—	73.7/91.6
FasterViT-3	159.5	IN-1k : Sup. : 300	— : — : —	18.2	84.9/97.2	44.2/—	51.9/—	75.0/92.2
FasterViT-4	424.6	IN-1k : Sup. : 300	— : — : —	36.6	85.4/97.3	49.0/—	56.0/—	75.7/92.7
FasterViT-5	957.5	IN-1k : Sup. : 300	— : — : —	113.0	85.6/97.4	52.7/—	56.9/—	76.0/93.0
FasterViT-6	1360.0	IN-1k : Sup. : 300	— : — : —	142.0	85.8/97.4	53.7/—	57.1/—	76.1/93.0
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 224	36.6	86.6/97.8	—/—	—/—	—/—
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 384	119.2	87.5/—	—/—	—/—	—/—
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 512	213.5	87.8/98.4	—/—	—/—	—/—
FasterViT-4	424.6	IN-22k : Sup. : 90	IN-1k : 60 : 768	480.4	87.9/98.5	—/—	—/—	—/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
FasterViT-2	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	52.1	71.0	56.6	—	—	—
FasterViT-3	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	52.4	71.1	56.7	—	—	—
FasterViT-4	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	52.9	71.6	57.7	—	—	—
FasterViT-4	IN-22k : Sup. : 90	DINO	COCO (train) : 36	1364.0	58.7	—	—	—	—	—
FasterViT-4	IN-22k : Sup. : 90	DINO	COCO (train) : 12	None	55.16	—	—	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
FasterViT-2	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	45.2	68.4	49.0	—	—	—
FasterViT-3	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	45.4	68.7	49.3	—	—	—
FasterViT-4	IN-1k : Sup. : 300	Cascade Mask R-CNN	COCO (train) : 36	None	45.8	69.1	49.8	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
FasterViT-1	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	None	—	—	—	42.7	—	—
FasterViT-2	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	974.0	48.4	—	—	47.2	—	—
FasterViT-2	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	None	—	—	—	43.5	—	—
FasterViT-3	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1076.0	49.7	—	—	48.7	—	—
FasterViT-4	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1290.0	50.3	—	—	49.1	—	—