Heedless Backbones

CETNet 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)

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

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
CETNet-T	23.4	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	935.0	47.9
CETNet-S	34.0	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	990.0	50.6
CETNet-B	75.0	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1176.0	51.6

model	params (m)	pretrain	finetune	gflops	IN-1k
CETNet-T	23.4	IN-1k : Sup. : 300	— : — : —	4.3	82.7/—
CETNet-T	23.4	IN-1k : Sup. : 300	IN-1k : 30 : 384	12.5	84.2/—
CETNet-S	34.0	IN-1k : Sup. : 300	— : — : —	6.8	83.4/—
CETNet-S	34.0	IN-1k : Sup. : 300	IN-1k : 30 : 384	19.9	84.6/—
CETNet-B	75.0	IN-1k : Sup. : 300	— : — : —	15.1	83.8/—
CETNet-B	75.0	IN-1k : Sup. : 300	IN-1k : 30 : 384	44.5	84.9/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
CETNet-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	261.0	45.5	67.7	50.0	—	—	—
CETNet-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	261.0	46.9	67.9	51.5	—	—	—
CETNet-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	315.0	46.6	68.7	51.4	—	—	—
CETNet-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	315.0	48.6	69.8	53.5	—	—	—
CETNet-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	495.0	47.9	70.3	53.0	—	—	—
CETNet-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	495.0	48.6	69.5	53.7	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
CETNet-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	261.0	40.7	64.4	43.7	—	—	—
CETNet-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	261.0	41.6	65.0	44.7	—	—	—
CETNet-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	315.0	41.6	65.4	44.8	—	—	—
CETNet-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	315.0	43.0	66.9	46.0	—	—	—
CETNet-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	495.0	42.5	67.2	45.6	—	—	—
CETNet-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 36	495.0	43.1	66.9	46.4	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
CETNet-T	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	935.0	47.9	—	—	46.5	—	—
CETNet-S	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	990.0	50.6	—	—	48.9	—	—
CETNet-B	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1176.0	51.6	—	—	50.2	—	—