Heedless Backbones

FaViT 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
FaViT-B0	3.0	IN-1k : Sup. : 300	— : — : —	0.6	71.5
FaViT-B1	13.0	IN-1k : Sup. : 300	— : — : —	2.4	79.4
FaViT-B2	24.0	IN-1k : Sup. : 300	— : — : —	4.5	82.3
FaViT-B3	48.0	IN-1k : Sup. : 300	— : — : —	8.5	83.4

model	params (m)	pretrain	finetune	gflops	IN-1k
FaViT-B0	3.0	IN-1k : Sup. : 300	— : — : —	0.6	71.5/—
FaViT-B1	13.0	IN-1k : Sup. : 300	— : — : —	2.4	79.4/—
FaViT-B2	24.0	IN-1k : Sup. : 300	— : — : —	4.5	82.3/—
FaViT-B3	48.0	IN-1k : Sup. : 300	— : — : —	8.5	83.4/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
FaViT-B0	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	37.4	57.2	39.8	22.9	40.6	49.5
FaViT-B0	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	37.9	59.6	41.0	—	—	—
FaViT-B1	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	41.4	61.8	44.0	25.7	44.9	55.0
FaViT-B1	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	42.4	64.4	46.3	—	—	—
FaViT-B2	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	44.4	65.0	47.7	27.7	48.2	58.8
FaViT-B2	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	45.4	67.1	49.4	—	—	—
FaViT-B3	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	47.1	68.0	51.4	—	—	—
FaViT-B3	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	46.0	66.7	49.1	28.4	50.3	62.0

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
FaViT-B0	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	35.4	56.6	37.8	—	—	—
FaViT-B1	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	38.9	61.2	41.8	—	—	—
FaViT-B2	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	41.0	64.0	44.1	—	—	—
FaViT-B3	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	42.7	65.9	46.1	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
FaViT-B0	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 128 : 512	None	—	—	—	37.2	—	—
FaViT-B1	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 128 : 512	None	—	—	—	42.0	—	—
FaViT-B2	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 128 : 512	None	—	—	—	45.0	—	—
FaViT-B3	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 128 : 512	None	—	—	—	47.2	—	—