Heedless Backbones

UniNeXt 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

pretrain dataset

Select an option

---------

ImageNet-1k

ImageNet-22k

JFT-300M

JFT-3B

MegData73M

pretrain method

Select an option

----------

Supervised

Sup. + TL

FCMAE

MAE

CL

MAP

Legend

Legend attribute

Select an option

Family

Pretrain Dataset

Pretrain Method

Classification Resolution

Legend attribute (second)

Select an option

----------

Pretrain Dataset

Pretrain Method

Classification Resolution

model	params (m)	pretrain	finetune	GFLOPs	Top-1
UniNeXt-T	24.0	IN-1k : Sup. : 300	— : — : —	4.3	83.6
UniNeXt-S	51.0	IN-1k : Sup. : 300	— : — : —	9.5	84.1
UniNeXt-B	91.0	IN-1k : Sup. : 300	— : — : —	17.1	84.4

model	params (m)	pretrain	finetune	gflops	IN-1k
UniNeXt-T	24.0	IN-1k : Sup. : 300	— : — : —	4.3	83.6/—
UniNeXt-S	51.0	IN-1k : Sup. : 300	— : — : —	9.5	84.1/—
UniNeXt-B	91.0	IN-1k : Sup. : 300	— : — : —	17.1	84.4/—

COCO (val)

model	pretrain	head	train	gflops	mAP^b	AP^b₅₀	AP^b₇₅	mAP^b_s	mAP^b_m	mAP^b_l
UniNeXt-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	266.0	48.6	70.6	53.4	—	—	—
UniNeXt-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	333.0	49.0	71.3	54.1	—	—	—
UniNeXt-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	460.0	49.3	71.4	54.1	—	—	—

COCO (val)

model	pretrain	head	train	gflops	mAP^m	AP^m₅₀	AP^m₇₅	mAP^m_s	mAP^m_m	mAP^m_l
UniNeXt-T	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	266.0	43.4	67.6	46.7	—	—	—
UniNeXt-S	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	333.0	43.7	68.0	46.9	—	—	—
UniNeXt-B	IN-1k : Sup. : 300	Mask R-CNN	COCO (train) : 12	460.0	43.9	68.3	47.3	—	—	—

ADE20K (val)

model	pretrain	head	train	gflops	mIoU^ms	pAcc^ms	mAcc^ms	mIoU^ss	pAcc^ss	mAcc^ss
UniNeXt-T	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	942.0	50.6	—	—	49.7	—	—
UniNeXt-S	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1010.0	51.8	—	—	51.0	—	—
UniNeXt-B	IN-1k : Sup. : 300	UPerNet	ADE20K (train) : 128 : 512	1142.0	52.2	—	—	51.4	—	—