PVT 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

---------

Cityscapes (val)

Cityscapes (test)

ADE20K (val)

ADE20K (test)

PASCAL VOC 2007 (val)

PASCAL VOC 2007 (test)

PASCAL VOC 2012 (val)

metric

Select an option

mIoU^ms

pAcc^ms

mAcc^ms

mIoU^ss

pAcc^ss

mAcc^ss

GFLOPs

head

Select an option

---------

UPerNet

Mask2Former

Panoptic FPN

SETR

Trans2Seg

DeepLabv3

PSPNet

Linear

resolution

Select an option

----------

512x2048

640x2560

axis

Select an option

Results

Parameters (M)

Images / Second

GFLOPs

Publication Date

Filters

Legend

Legend attribute

Select an option

Family

Pretrain Dataset

Semantic Segmentation Head

Semantic Segmentation Resolution

Semantic Segmentation Training Epochs

Legend attribute (second)

Select an option

----------

Family

Pretrain Method

Semantic Segmentation Head

Semantic Segmentation Resolution

Semantic Segmentation Training Epochs

model	params (m)	pretrain	head	train	GFLOPs	mIoU^ms
PVT-Large	61.4	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 256 : 512	79.6	44.8

model	params (m)	pretrain	finetune	gflops	IN-1k
PVT-Tiny	13.2	IN-1k : Sup. : 300	— : — : —	1.9	75.1/—
PVT-Small	24.5	IN-1k : Sup. : 300	— : — : —	3.8	79.8/—
PVT-Medium	44.2	IN-1k : Sup. : 300	— : — : —	6.7	81.2/—
PVT-Large	61.4	IN-1k : Sup. : 300	— : — : —	9.8	81.7/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
PVT-Tiny	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	36.7	56.9	38.9	22.6	38.8	50.0
PVT-Tiny	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 36	None	39.4	59.8	42.0	25.5	42.0	52.1
PVT-Tiny	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	36.7	59.2	39.3	—	—	—
PVT-Tiny	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	39.8	62.2	43.0	—	—	—
PVT-Small	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	285.8	40.4	61.3	43.0	25.0	42.9	55.7
PVT-Small	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 36	None	42.2	62.7	45.0	26.2	45.2	57.2
PVT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	304.4	40.4	62.9	43.8	—	—	—
PVT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	43.0	65.3	46.9	—	—	—
PVT-Small	IN-1k : Sup. : 300	DETR	COCO (train) : 50	None	34.7	55.7	35.4	12.0	36.4	56.7
PVT-Medium	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	41.9	63.1	44.3	25.0	44.9	57.6
PVT-Medium	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 36	None	43.2	63.8	46.1	27.3	46.3	58.9
PVT-Medium	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	42.0	64.4	45.6	—	—	—
PVT-Medium	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	44.2	66.0	48.2	—	—	—
PVT-Large	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 12	None	42.6	63.7	45.4	25.8	46.0	58.4
PVT-Large	IN-1k : Sup. : 300	RetinaNet	COCO (train) : 36	None	43.4	63.6	46.1	26.1	46.0	59.5
PVT-Large	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	42.9	65.0	46.6	—	—	—
PVT-Large	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	44.5	66.0	48.3	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
PVT-Tiny	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	35.1	56.7	37.3	—	—	—
PVT-Tiny	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	37.4	59.3	39.9	—	—	—
PVT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	304.4	37.8	60.1	40.3	—	—	—
PVT-Small	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	39.9	62.5	42.8	—	—	—
PVT-Medium	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	39.0	61.6	42.1	—	—	—
PVT-Medium	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	40.5	63.1	43.5	—	—	—
PVT-Large	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	None	39.5	61.9	42.5	—	—	—
PVT-Large	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	None	40.7	63.4	43.7	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
PVT-Tiny	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	33.2	—	—	—	35.7	—	—
PVT-Small	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	44.5	—	—	—	39.8	—	—
PVT-Small	IN-1k : Sup. : 300	Trans2Seg	ADE20K (train) : 32 : 512	31.6	—	—	—	42.6	—	—
PVT-Medium	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	61.0	—	—	—	41.6	—	—
PVT-Large	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 64 : 512	79.6	—	—	—	42.1	—	—
PVT-Large	IN-1k : Sup. : 300	Panoptic FPN	ADE20K (train) : 256 : 512	79.6	44.8	—	—	—	—	—