Segmentation Models

InternImage for semantic segmentation: ADE20K, Cityscapes.

ADE20K

UperNet + InternImage

backbone	resolution	mIoU (ss/ms)	train speed	train time	#param	FLOPs	Config	Download
InternImage-T	512x512	47.9 / 48.1	0.23s / iter	10.5h	59M	944G	config	ckpt | log
InternImage-S	512x512	50.1 / 50.9	0.25s / iter	11.5h	80M	1017G	config	ckpt | log
InternImage-B	512x512	50.8 / 51.3	0.26s / iter	12h	128M	1185G	config	ckpt | log
InternImage-L	640x640	53.9 / 54.1	0.42s / iter	19h	256M	2526G	config	ckpt | log
InternImage-XL	640x640	55.0 / 55.3	0.47s / iter	22h	368M	3142G	config	ckpt | log

Cityscapes

UperNet + InternImage

backbone	resolution	mIoU (ss/ms)	train speed	train time	#params	FLOPs	Config	Download
InternImage-T	512x1024	82.58 / 83.40	0.32s / iter	14.5h	59M	1889G	config	ckpt | log
InternImage-S	512x1024	82.74 / 83.45	0.36s / iter	16.5h	80M	2035G	config	ckpt | log
InternImage-B	512x1024	83.18 / 83.97	0.39s / iter	17h	128M	2369G	config	ckpt | log
InternImage-L	512x1024	83.68 / 84.41	0.50s / iter	23h	256M	3234G	config	ckpt | log
InternImage-XL	512x1024	83.62 / 84.28	0.56s / iter	26h	368M	4022G	config	ckpt | log

UperNet + InternImage (with additional data)

Mapillary 80k + Cityscapes (w/ coarse data) 160k

backbone	resolution	mIoU (ss/ms)	train speed	train time	#params	FLOPs	Config	Download
InternImage-L	512x1024	85.94 / 86.22	0.50s / iter	23h	256M	3234G	config	ckpt | log
InternImage-XL	512x1024	86.20 / 86.42	0.56s / iter	26h	368M	4022G	config	ckpt | log

SegFormerHead + InternImage (with additional data)

Mapillary 80k + Cityscapes (w/ coarse data) 160k

backbone	resolution	mIoU (ss/ms)	train speed	train time	#params	FLOPs	Config	Download
InternImage-L	512x1024	85.16 / 85.67	0.37s / iter	17h	220M	1580G	config	ckpt | log
InternImage-XL	512x1024	85.41 / 85.93	0.43s / iter	19.5h	330M	2364G	config	ckpt | log