Benchmarks · fastforks

kernels (our CUDA kernels) against portable (plain PyTorch, each model's tested precision) and the original authors' code. One RTX 5060 Ti 16 GB, torch 2.14.1+cu130, CUDA 12.8 kernels (13.0 for four stereo families), cuDNN 9.24. Stereo and depth rows use real images (KITTI 2015 000000_10, Middlebury Motorcycle) as uint8 CUDA tensors with default options; other tasks use a synthetic input at the task's default size. Speeds are medians of warm runs (5 for most rows; each row's protocol is in docs/data/speed.jsonl); peak memory is torch's peak allocated / reserved GPU MiB (CUDA, weights included) over the first and the timed calls. fastforks bench MODEL --size HxW measures your GPU (it uses 10 runs).

Sizes are written height x width everywhere (375x1242, as in fastforks bench --size).

To choose a model, use the overview: one standard size per task, generated from the same data. The tables below give one table per task and scene (each model at the standard size, or its own input where it has no other; the model pages list every measured size), the startup times and, at the end, the same models at fixed sizes up to 12 MP.

Accuracy matches the original code. On ETH3D and Middlebury (the "Measured" tables on the model pages) both modes stay within 0.15 bad-% / 0.025 EPE of the original with the same checkpoint for AANet, AANet+, CoEx, IGEV, Selective-IGEV, LiteAnyStereo V2 M/H, DEFOM, FoundationStereo-L and CREStereo. S2M2 is better (up to 1 bad-% lower) because our port uses the cost-volume lookup convention the released weights were trained with.

The accuracy columns on the model pages check parity (original code vs portable vs kernels) on public images; they are not a ranking. Variants trained or fine-tuned on ETH3D / Middlebury are marked † in those columns (? where the training data is not stated); to rank models use the official Middlebury and ETH3D test-set leaderboards.

Modes, states and devices

These terms apply to every device:

Modes: original = the authors' code with the same checkpoint (fp32 and the authors' own fp16 option, else torch.autocast), portable = fastforks in plain PyTorch, kernels = fastforks with its own kernels (CUDA, or Metal on Apple silicon). Every time column names its dtype or its original precision.
Timing: median of warm, synchronised calls after warm-up, one model at a time. Startup (load, first call) is listed separately.
Memory is recorded as one of two kinds, never compared with each other: peak torch memory (allocated / reserved, CUDA and ROCm) or the peak process footprint (Apple silicon, where weights and activations share the system memory).
Speed-ups name their baseline (vs original, vs portable, vs fastest portable). When a baseline is missing the cell stays empty; another baseline is not used in its place.
States: unsupported = the card does not offer that mode or dtype on this backend, failed = it ran and raised, OOM = out of memory, – = not tested.
The reference device is the RTX 5060 Ti below: its rows are the regression floor (tests/test_speed_floor.py). Every other device is a survey with its own page; no floor is kept for it. Adding a device: "Devices" in docs/data/README.md.

Other devices

Apple M1 Pro 16 GB (mps): 169 models, 422 rows, measured 2026-11-06 to 2026-11-07.
NVIDIA L4 (cuda): 119 models, 238 rows, measured 2026-11-07 to 2026-11-08.
NVIDIA L40S (cuda): 119 models, 238 rows, measured 2026-11-07 to 2026-11-08.
NVIDIA A10 (cuda): 119 models, 238 rows, measured 2026-11-07 to 2026-11-08.
NVIDIA A100-SXM4-40GB (cuda): 117 models, 234 rows, measured 2026-11-07 to 2026-11-08.
NVIDIA RTX PRO 6000 (cuda): 174 models, 348 rows, measured 2026-11-08 to 2026-11-08.

Across devices

Each task at its standard input (the overview's). A cell is the fastest validated fastforks mode on that device (portable or kernels, time and dtype), or why there is none; a row needs a survey device measured at that input. Cells compare only rows of the same input and protocol (warm-median: median of warm, synchronised calls after warm-up (at least 5; 2 for calls over 2 s on surveys)); no speed-up is derived across devices, as the modes and dtypes differ. – = not measured at this input.

Stereo, 994x1482

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`aanet-plus-sceneflow`	88.9 ms · kernels fp16	818 ms · kernels fp16	111 ms · kernels fp16	46.0 ms · kernels fp16	97.6 ms · kernels fp16	50.8 ms · kernels fp16	22.1 ms · kernels fp16
`aanet-sceneflow`	101 ms · kernels fp16	631 ms · kernels fp16	140 ms · kernels fp16	44.7 ms · kernels fp16	105 ms · kernels fp16	49.6 ms · kernels fp16	20.9 ms · kernels fp16
`bridgedepth-rvc-pretrain`	354 ms · kernels fp16	3756 ms · kernels fp16	–	–	–	143 ms · kernels fp16	68.8 ms · kernels fp16
`bridgedepth-sf`	348 ms · kernels fp16	3693 ms · kernels fp16	430 ms · kernels fp16	122 ms · kernels fp16	333 ms · kernels fp16	141 ms · kernels fp16	68.8 ms · kernels fp16
`c-fast-foundationstereo`	135 ms · kernels fp16	4881 ms · portable fp32	–	–	–	85.1 ms · kernels fp16	37.3 ms · kernels fp16
`c-foundationstereo-s`	1362 ms · kernels fp16	OOM	1908 ms · kernels fp16	551 ms · kernels fp16	1537 ms · kernels fp16	668 ms · kernels fp16	319 ms · kernels fp16
`cas-gwcnet-c-sceneflow`	225 ms · kernels fp16	2489 ms · portable fp16	–	–	–	87.9 ms · kernels fp16	46.7 ms · kernels fp16
`cas-gwcnet-gc-sceneflow`	358 ms · kernels fp16	2995 ms · portable fp16	500 ms · kernels fp16	159 ms · kernels fp16	345 ms · kernels fp16	136 ms · kernels fp16	78.8 ms · kernels fp16
`cfnet-sceneflow`	401 ms · kernels fp16	2206 ms · kernels fp16	484 ms · kernels fp16	152 ms · kernels fp16	432 ms · kernels fp16	145 ms · kernels fp16	76.4 ms · kernels fp16
`coex-sceneflow`	23.5 ms · kernels fp16	411 ms · portable fp16	32.4 ms · kernels fp16	11.9 ms · kernels fp16	27.3 ms · kernels fp16	16.3 ms · kernels fp16	6.43 ms · kernels fp16
`crestereo-eth3d`	229 ms · kernels fp16	3342 ms · kernels fp16	318 ms · kernels fp16	102 ms · kernels fp16	255 ms · kernels fp16	108 ms · kernels fp16	60.9 ms · kernels fp16
`defom-stereo-vitl-sceneflow`	659 ms · kernels fp16	8224 ms · kernels fp16	921 ms · kernels fp16	271 ms · kernels fp16	689 ms · kernels fp16	248 ms · kernels fp16	146 ms · kernels fp16
`dlnr-middlebury`	214 ms · kernels fp16	2143 ms · portable fp16	–	–	–	–	48.6 ms · kernels fp16
`dlnr-sceneflow`	214 ms · kernels fp16	2143 ms · portable fp16	293 ms · kernels fp16	94.6 ms · kernels fp16	224 ms · kernels fp16	95.4 ms · kernels fp16	48.6 ms · kernels fp16
`fast-acvnet-generalization`	38.2 ms · kernels fp16	944 ms · portable fp32	51.7 ms · kernels fp16	19.2 ms · kernels fp16	43.6 ms · kernels fp16	21.5 ms · kernels fp16	9.33 ms · kernels fp16
`fast-acvnet-kitti12`	38.2 ms · kernels fp16	942 ms · portable fp32	–	–	–	–	9.58 ms · kernels fp16
`fast-acvnet-kitti15`	38.2 ms · kernels fp16	943 ms · portable fp32	–	–	–	–	9.34 ms · kernels fp16
`fast-acvnet-plus-generalization`	46.4 ms · kernels fp16	1322 ms · portable fp32	–	–	–	–	11.7 ms · kernels fp16
`fast-acvnet-plus-kitti12`	46.4 ms · kernels fp16	1321 ms · portable fp32	–	–	–	–	11.5 ms · kernels fp16
`fast-acvnet-plus-kitti15`	46.4 ms · kernels fp16	1322 ms · portable fp32	–	–	–	–	11.6 ms · kernels fp16
`fast-acvnet-plus-sceneflow`	46.4 ms · kernels fp16	1322 ms · portable fp32	62.8 ms · kernels fp16	20.3 ms · kernels fp16	51.2 ms · kernels fp16	28.7 ms · kernels fp16	11.8 ms · kernels fp16
`fast-acvnet-sceneflow`	38.2 ms · kernels fp16	942 ms · portable fp32	–	–	–	–	9.55 ms · kernels fp16
`fast-foundationstereo`	135 ms · kernels fp16	5195 ms · portable fp32	201 ms · kernels fp16	70.2 ms · kernels fp16	170 ms · kernels fp16	85.3 ms · kernels fp16	36.8 ms · kernels fp16
`foundationstereo-vit-l`	1513 ms · kernels fp16	24480 ms · kernels fp16	2035 ms · kernels fp16	605 ms · kernels fp16	1674 ms · kernels fp16	680 ms · kernels fp16	337 ms · kernels fp16
`foundationstereo-vit-s`	1120 ms · kernels fp16	19496 ms · kernels fp16	1713 ms · kernels fp16	487 ms · kernels fp16	1369 ms · kernels fp16	561 ms · kernels fp16	270 ms · kernels fp16
`great-igev-eth3d`	393 ms · kernels fp16	9636 ms · kernels fp16	–	–	–	–	97.2 ms · kernels fp16
`great-igev-kitti`	393 ms · kernels fp16	9634 ms · kernels fp16	–	–	–	–	97.4 ms · kernels fp16
`great-igev-middlebury`	579 ms · kernels fp16	25368 ms · kernels fp16	–	–	–	–	143 ms · kernels fp16
`great-igev-sceneflow`	393 ms · kernels fp16	9861 ms · kernels fp16	579 ms · kernels fp16	176 ms · kernels fp16	477 ms · kernels fp16	196 ms · kernels fp16	98.3 ms · kernels fp16
`gwcnet-g-sceneflow`	238 ms · kernels fp16	4340 ms · portable fp32	–	–	–	–	50.5 ms · kernels fp16
`gwcnet-gc-sceneflow`	257 ms · kernels fp16	4461 ms · portable fp32	342 ms · kernels fp16	109 ms · kernels fp16	244 ms · kernels fp16	97.7 ms · kernels fp16	56.7 ms · kernels fp16
`hitnet-eth3d`	23.2 ms · kernels fp16	136 ms · kernels fp16	31.4 ms · kernels fp16	11.0 ms · kernels fp16	28.3 ms · kernels fp16	16.9 ms · kernels fp16	6.22 ms · kernels fp16
`hitnet-middlebury-d400`	52.7 ms · kernels fp16	440 ms · kernels fp16	–	–	–	–	14.3 ms · kernels fp16
`igev-plusplus-eth3d`	393 ms · kernels fp16	6139 ms · kernels fp16	–	–	–	–	104 ms · kernels fp16
`igev-plusplus-kitti12`	393 ms · kernels fp16	6124 ms · kernels fp16	–	–	–	–	104 ms · kernels fp16
`igev-plusplus-kitti15`	393 ms · kernels fp16	6125 ms · kernels fp16	–	–	–	–	104 ms · kernels fp16
`igev-plusplus-middlebury`	393 ms · kernels fp16	6130 ms · kernels fp16	–	–	–	–	104 ms · kernels fp16
`igev-plusplus-sceneflow`	393 ms · kernels fp16	6135 ms · kernels fp16	547 ms · kernels fp16	177 ms · kernels fp16	463 ms · kernels fp16	186 ms · kernels fp16	104 ms · kernels fp16
`igev-stereo-eth3d`	288 ms · kernels fp16	3998 ms · kernels fp16	–	–	–	–	72.6 ms · kernels fp16
`igev-stereo-kitti12`	288 ms · kernels fp16	4006 ms · kernels fp16	–	–	–	–	72.2 ms · kernels fp16
`igev-stereo-kitti15`	288 ms · kernels fp16	4008 ms · kernels fp16	–	–	–	–	72.3 ms · kernels fp16
`igev-stereo-middlebury`	288 ms · kernels fp16	4007 ms · kernels fp16	–	–	–	–	72.6 ms · kernels fp16
`igev-stereo-sceneflow`	288 ms · kernels fp16	4001 ms · kernels fp16	430 ms · kernels fp16	134 ms · kernels fp16	349 ms · kernels fp16	137 ms · kernels fp16	72.6 ms · kernels fp16
`las2-h`	53.3 ms · kernels fp16	734 ms · kernels fp16	77.0 ms · kernels fp16	36.3 ms · kernels fp16	65.1 ms · kernels fp16	38.3 ms · kernels fp16	15.5 ms · kernels fp16
`las2-l`	29.9 ms · kernels fp16	519 ms · kernels fp16	–	–	–	–	8.96 ms · kernels fp16
`las2-m`	18.3 ms · kernels fp16	281 ms · kernels fp16	–	–	–	–	6.91 ms · kernels fp16
`las2-s`	14.3 ms · kernels fp16	198 ms · kernels fp16	23.0 ms · kernels fp16	15.3 ms · kernels fp16	16.9 ms · kernels fp16	15.7 ms · kernels fp16	4.99 ms · kernels fp16
`mocha-v1-sceneflow`	455 ms · kernels fp16	6693 ms · kernels fp16	631 ms · kernels fp16	224 ms · kernels fp16	527 ms · kernels fp16	227 ms · kernels fp16	116 ms · kernels fp16
`mocha-v2-sceneflow`	448 ms · kernels fp16	6501 ms · kernels fp16	–	–	–	–	115 ms · kernels fp16
`monster-mix-all`	668 ms · kernels fp16	9646 ms · kernels fp16	–	–	–	–	164 ms · kernels fp16
`monster-plusplus-mix-all`	698 ms · kernels fp16	10202 ms · kernels fp16	949 ms · kernels fp16	288 ms · kernels fp16	792 ms · kernels fp16	289 ms · kernels fp16	167 ms · kernels fp16
`monster-sceneflow`	668 ms · kernels fp16	9642 ms · kernels fp16	951 ms · kernels fp16	287 ms · kernels fp16	783 ms · kernels fp16	286 ms · kernels fp16	164 ms · kernels fp16
`nmrf-sceneflow`	196 ms · kernels fp16	1204 ms · kernels fp16	239 ms · kernels fp16	64.6 ms · kernels fp16	173 ms · kernels fp16	84.4 ms · kernels fp16	37.7 ms · kernels fp16
`psmnet-sceneflow`	232 ms · kernels fp16	4615 ms · portable fp32	290 ms · kernels fp16	90.6 ms · kernels fp16	218 ms · kernels fp16	91.3 ms · kernels fp16	48.9 ms · kernels fp16
`raft-stereo-eth3d`	273 ms · kernels fp16	3638 ms · kernels fp16	–	–	–	–	63.6 ms · kernels fp16
`raft-stereo-middlebury`	273 ms · kernels fp16	3634 ms · kernels fp16	416 ms · kernels fp16	121 ms · kernels fp16	313 ms · kernels fp16	110 ms · kernels fp16	63.6 ms · kernels fp16
`raft-stereo-sceneflow`	273 ms · kernels fp16	3637 ms · kernels fp16	–	–	–	–	63.6 ms · kernels fp16
`rt-monster-plusplus-mix-all`	94.1 ms · kernels fp16	663 ms · portable fp32	125 ms · kernels fp16	48.3 ms · portable fp16	105 ms · kernels fp16	49.4 ms · portable fp16	23.6 ms · kernels fp16
`s2m2-l`	376 ms · kernels fp16	6590 ms · portable fp16	–	–	–	–	91.7 ms · kernels fp16
`s2m2-m`	248 ms · kernels fp16	3620 ms · portable fp16	–	–	–	–	61.1 ms · kernels fp16
`s2m2-s`	113 ms · kernels fp16	1607 ms · portable fp16	163 ms · kernels fp16	56.8 ms · kernels fp16	137 ms · kernels fp16	62.4 ms · kernels fp16	29.0 ms · kernels fp16
`s2m2-xl`	682 ms · kernels fp16	13398 ms · portable fp16	1152 ms · kernels fp16	310 ms · kernels fp16	822 ms · kernels fp16	312 ms · kernels fp16	168 ms · kernels fp16
`selective-igev-eth3d-finetune`	349 ms · kernels fp16	5708 ms · kernels fp16	–	–	–	–	89.3 ms · kernels fp16
`selective-igev-eth3d-train`	349 ms · kernels fp16	5710 ms · kernels fp16	–	–	–	–	89.3 ms · kernels fp16
`selective-igev-kitti12`	349 ms · kernels fp16	5709 ms · kernels fp16	–	–	–	–	89.2 ms · kernels fp16
`selective-igev-kitti15`	349 ms · kernels fp16	5710 ms · kernels fp16	–	–	–	–	89.1 ms · kernels fp16
`selective-igev-middlebury-finetune`	349 ms · kernels fp16	5778 ms · kernels fp16	–	–	–	–	89.4 ms · kernels fp16
`selective-igev-middlebury-train`	349 ms · kernels fp16	5738 ms · kernels fp16	–	–	–	–	89.2 ms · kernels fp16
`selective-igev-sceneflow`	349 ms · kernels fp16	5710 ms · kernels fp16	515 ms · kernels fp16	163 ms · kernels fp16	410 ms · kernels fp16	168 ms · kernels fp16	89.2 ms · kernels fp16
`waft-stereo-b`	393 ms · kernels fp16	11022 ms · portable fp32	574 ms · kernels fp16	146 ms · kernels fp16	422 ms · kernels fp16	173 ms · kernels fp16	85.7 ms · kernels fp16
`waft-stereo-s`	284 ms · kernels fp16	3360 ms · portable fp32	358 ms · kernels fp16	100 ms · kernels fp16	283 ms · kernels fp16	125 ms · kernels fp16	60.1 ms · kernels fp16

Depth, 994x1482

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`da3-giant-1.1`	44.2 ms · kernels fp16	759 ms · portable fp16	63.2 ms · kernels fp16	22.5 ms · kernels fp16	60.1 ms · kernels fp16	25.0 ms · kernels fp16	14.2 ms · kernels fp16
`da3-large-1.1`	17.1 ms · kernels fp16	327 ms · portable fp16	22.4 ms · kernels fp16	8.49 ms · kernels fp16	21.1 ms · kernels fp16	–	6.59 ms · kernels fp16
`da3metric-large`	14.7 ms · kernels fp16	318 ms · portable fp16	19.5 ms · kernels fp16	7.78 ms · kernels fp16	18.6 ms · kernels fp16	–	6.04 ms · kernels fp16
`da3mono-large`	14.6 ms · kernels fp16	272 ms · portable fp16	19.2 ms · kernels fp16	7.66 ms · kernels fp16	18.5 ms · kernels fp16	12.8 ms · kernels fp16	5.81 ms · kernels fp16
`da3nested-giant-large-1.1`	58.2 ms · kernels fp16	1184 ms · portable fp16	81.9 ms · kernels fp16	30.0 ms · kernels fp16	78.1 ms · kernels fp16	–	19.7 ms · kernels fp16
`depth-anything-v2-b`	10.5 ms · kernels fp16	212 ms · portable fp32	14.6 ms · kernels fp16	4.67 ms · kernels fp16	13.7 ms · kernels fp16	–	4.00 ms · kernels fp16
`depth-anything-v2-l`	31.1 ms · kernels fp16	622 ms · portable fp32	42.7 ms · kernels fp16	13.6 ms · kernels fp16	40.1 ms · kernels fp16	16.1 ms · kernels fp16	8.61 ms · kernels fp16
`depth-anything-v2-s`	4.51 ms · kernels fp16	87.3 ms · portable fp32	5.92 ms · kernels fp16	2.24 ms · kernels fp16	5.43 ms · kernels fp16	4.63 ms · kernels fp16	2.02 ms · kernels fp16
`depth-pro`	271 ms · kernels fp16	5026 ms · portable fp16	472 ms · kernels fp16	133 ms · kernels fp16	359 ms · kernels fp16	119 ms · kernels fp16	68.9 ms · kernels fp16
`marigold-depth-v1-1`	383 ms · kernels fp16	2498 ms · portable fp16	346 ms · kernels fp16	128 ms · kernels fp16	260 ms · kernels fp16	177 ms · portable fp16	74.8 ms · kernels fp16
`metric3d-vit-large`	158 ms · kernels fp16	1822 ms · portable fp32	193 ms · kernels fp16	56.9 ms · kernels fp16	166 ms · kernels fp16	61.5 ms · kernels fp16	38.9 ms · kernels fp16
`midas-dpt-beit-large-512`	63.9 ms · kernels fp16	635 ms · portable fp16	91.5 ms · kernels fp16	26.8 ms · kernels fp16	73.1 ms · portable fp16	28.5 ms · kernels fp16	17.8 ms · portable fp16
`moge-2-vitl-normal`	73.4 ms · kernels fp16	1454 ms · kernels fp32	103 ms · kernels fp16	35.2 ms · kernels fp16	95.2 ms · kernels fp16	–	24.1 ms · kernels fp16
`moge-2-vits-normal`	29.7 ms · kernels fp16	363 ms · portable fp32	44.0 ms · kernels fp16	18.0 ms · kernels fp16	35.6 ms · kernels fp16	22.1 ms · kernels fp16	11.0 ms · kernels fp16
`moge-3-vitl`	–	2764 ms · kernels fp32	284 ms · kernels fp16	93.5 ms · kernels fp16	276 ms · kernels fp16	114 ms · kernels fp16	61.5 ms · kernels fp16
`unidepth-v2-vitl14`	66.4 ms · kernels fp16	1260 ms · portable fp32	85.8 ms · kernels fp16	25.4 ms · kernels fp16	77.0 ms · kernels fp16	26.6 ms · kernels fp16	18.4 ms · portable fp16
`unik3d-vitl`	62.9 ms · kernels fp16	1217 ms · portable fp32	88.9 ms · kernels fp16	27.7 ms · kernels fp16	78.4 ms · kernels fp16	29.8 ms · kernels fp16	19.1 ms · kernels fp16
`zoedepth-nk`	88.9 ms · kernels fp16	778 ms · portable fp32	116 ms · kernels fp16	32.1 ms · kernels fp16	92.9 ms · kernels fp16	41.2 ms · kernels fp16	19.2 ms · kernels fp16

Classification, 224x224

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`alexnet.tv_in1k`	0.660 ms · kernels fp16	3.04 ms · portable fp16	1.03 ms · kernels fp16	0.544 ms · kernels fp16	0.616 ms · kernels fp16	0.591 ms · kernels fp16	0.240 ms · kernels fp16
`convnextv2_base.fcmae_ft_in22k_in1k`	3.07 ms · kernels fp16	36.1 ms · portable fp32	3.68 ms · kernels fp16	2.64 ms · kernels fp16	3.79 ms · kernels fp16	5.03 ms · kernels fp16	2.83 ms · kernels fp16
`deit_base_patch16_224.fb_in1k`	1.79 ms · kernels fp16	16.9 ms · portable fp32	1.90 ms · kernels fp16	1.22 ms · kernels fp16	1.77 ms · kernels fp16	2.21 ms · kernels fp16	1.02 ms · kernels fp16
`densenet121.tv_in1k`	1.57 ms · kernels fp16	25.4 ms · portable fp32	1.80 ms · kernels fp16	1.40 ms · kernels fp16	1.91 ms · kernels fp16	2.69 ms · kernels fp16	1.31 ms · kernels fp16
`dinov2_vitl14_lc`	5.24 ms · kernels fp16	64.5 ms · portable fp32	6.65 ms · kernels fp16	2.76 ms · kernels fp16	5.94 ms · portable fp16	4.71 ms · kernels fp16	2.22 ms · portable fp16
`efficientnet_b0.tv_in1k`	0.870 ms · kernels fp16	13.0 ms · portable fp32	1.14 ms · kernels fp16	0.844 ms · kernels fp16	1.09 ms · kernels fp16	1.51 ms · kernels fp16	0.736 ms · kernels fp16
`efficientnet_b4.tv_in1k`	2.19 ms · kernels fp16	38.3 ms · portable fp32	2.45 ms · kernels fp16	1.70 ms · kernels fp16	2.55 ms · kernels fp16	3.30 ms · kernels fp16	1.61 ms · kernels fp16
`efficientnet_v2_s.tv_in1k`	2.11 ms · kernels fp16	34.0 ms · portable fp32	2.43 ms · kernels fp16	1.74 ms · kernels fp16	2.47 ms · kernels fp16	3.29 ms · kernels fp16	1.62 ms · kernels fp16
`efficientvit_b2.r288_in1k`	1.20 ms · kernels fp16	18.6 ms · portable fp32	1.55 ms · kernels fp16	1.01 ms · kernels fp16	1.40 ms · kernels fp16	1.89 ms · kernels fp16	0.926 ms · kernels fp16
`eva02_large_patch14_448.mim_m38m_ft_in22k_in1k`	14.2 ms · kernels fp16	235 ms · portable fp16	19.6 ms · kernels fp16	6.82 ms · kernels fp16	16.9 ms · kernels fp16	10.3 ms · kernels fp16	5.46 ms · kernels fp16
`fastvit_sa24.apple_dist_in1k`	1.49 ms · kernels fp16	16.8 ms · portable fp32	1.76 ms · kernels fp16	1.27 ms · kernels fp16	1.73 ms · kernels fp16	2.49 ms · kernels fp16	1.31 ms · kernels fp16
`inception_v3.tv_in1k`	1.56 ms · kernels fp16	17.4 ms · portable fp32	1.76 ms · kernels fp16	1.31 ms · kernels fp16	1.74 ms · kernels fp16	2.40 ms · kernels fp16	1.26 ms · kernels fp16
`mobilenet_v2.tv2_in1k`	0.520 ms · kernels fp16	6.48 ms · portable fp32	0.695 ms · kernels fp16	0.506 ms · kernels fp16	0.580 ms · kernels fp16	0.768 ms · kernels fp16	0.346 ms · kernels fp16
`mobilenet_v3_large.tv2_in1k`	0.630 ms · kernels fp16	8.54 ms · portable fp32	0.853 ms · kernels fp16	0.632 ms · kernels fp16	0.777 ms · kernels fp16	1.05 ms · kernels fp16	0.489 ms · kernels fp16
`mobilenetv1_100.ra4_e3600_r224_in1k`	0.430 ms · kernels fp16	4.29 ms · portable fp32	0.593 ms · kernels fp16	0.413 ms · kernels fp16	0.448 ms · kernels fp16	0.574 ms · kernels fp16	0.253 ms · kernels fp16
`mobilenetv4_conv_medium.e500_r256_in1k`	0.750 ms · kernels fp16	9.58 ms · portable fp32	0.955 ms · kernels fp16	0.722 ms · kernels fp16	0.900 ms · kernels fp16	1.19 ms · kernels fp16	0.592 ms · kernels fp16
`regnety_160.tv2_in1k`	1.96 ms · kernels fp16	33.1 ms · portable fp16	2.37 ms · kernels fp16	1.58 ms · kernels fp16	2.27 ms · kernels fp16	2.70 ms · kernels fp16	1.46 ms · kernels fp16
`resnet152.tv2_in1k`	2.17 ms · kernels fp16	25.7 ms · portable fp32	2.32 ms · kernels fp16	1.84 ms · kernels fp16	2.48 ms · kernels fp16	3.42 ms · kernels fp16	1.82 ms · kernels fp16
`resnet18.tv_in1k`	0.550 ms · kernels fp16	4.67 ms · portable fp32	0.718 ms · kernels fp16	0.519 ms · kernels fp16	0.609 ms · kernels fp16	0.797 ms · kernels fp16	0.370 ms · kernels fp16
`resnet50.tv2_in1k`	0.990 ms · kernels fp16	11.2 ms · portable fp32	–	–	–	–	0.666 ms · kernels fp16
`shufflenet_v2_x1_0.tv_in1k`	0.840 ms · kernels fp16	6.33 ms · portable fp32	1.13 ms · kernels fp16	0.874 ms · kernels fp16	1.22 ms · kernels fp16	1.63 ms · kernels fp16	0.811 ms · kernels fp16
`squeezenet1_1.tv_in1k`	0.420 ms · kernels fp16	3.72 ms · portable fp16	0.590 ms · kernels fp16	0.437 ms · kernels fp16	0.461 ms · kernels fp16	0.592 ms · kernels fp16	0.269 ms · kernels fp16
`swin_base_patch4_window7_224.ms_in22k_ft_in1k`	2.45 ms · kernels fp16	28.2 ms · portable fp32	2.92 ms · kernels fp16	2.02 ms · kernels fp16	2.80 ms · kernels fp16	3.79 ms · kernels fp16	1.91 ms · kernels fp16
`vgg16.tv_in1k`	1.56 ms · kernels fp16	11.7 ms · portable fp16	2.10 ms · kernels fp16	1.03 ms · kernels fp16	1.47 ms · kernels fp16	1.15 ms · kernels fp16	0.622 ms · kernels fp16
`vit_base_patch16_224.augreg2_in21k_ft_in1k`	1.78 ms · kernels fp16	16.9 ms · portable fp32	1.92 ms · kernels fp16	1.22 ms · kernels fp16	1.77 ms · kernels fp16	2.21 ms · kernels fp16	1.02 ms · kernels fp16
`vit_so400m_patch14_siglip_378.webli_ft_in1k`	12.9 ms · kernels fp16	197 ms · portable fp16	15.8 ms · kernels fp16	6.99 ms · kernels fp16	16.0 ms · kernels fp16	7.96 ms · portable fp16	4.80 ms · portable fp16

Segmentation, 1024x2048

model	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`deeplabv3-r101-d8-ade20k`	528 ms · portable fp16	–	–	–	–	6.03 ms · kernels fp16
`deeplabv3plus-r101-d8-ade20k`	430 ms · portable fp16	24.9 ms · kernels fp16	7.55 ms · kernels fp16	20.8 ms · kernels fp16	13.0 ms · kernels fp16	5.21 ms · kernels fp16
`dinov2-b-linear-ade20k`	409 ms · portable fp32	–	–	–	–	6.86 ms · kernels fp16
`dinov2-l-linear-ade20k`	1032 ms · portable fp32	81.8 ms · kernels fp16	22.8 ms · kernels fp16	67.1 ms · kernels fp16	22.3 ms · kernels fp16	13.8 ms · kernels fp16
`dinov2-s-linear-ade20k`	224 ms · portable fp32	16.6 ms · kernels fp16	6.60 ms · kernels fp16	13.1 ms · kernels fp16	7.64 ms · kernels fp16	3.90 ms · kernels fp16
`eomt-l-dinov2-ade20k`	540 ms · portable fp16	47.3 ms · kernels fp16	13.9 ms · kernels fp16	38.3 ms · kernels fp16	15.0 ms · kernels fp16	8.51 ms · kernels fp16
`eomt-l-dinov3-ade20k`	575 ms · portable fp16	54.3 ms · kernels fp16	15.4 ms · kernels fp16	44.6 ms · kernels fp16	17.8 ms · kernels fp16	9.54 ms · kernels fp16
`fcn-r101-d8-ade20k`	405 ms · portable fp16	21.8 ms · kernels fp16	6.92 ms · kernels fp16	19.5 ms · kernels fp16	11.8 ms · kernels fp16	5.21 ms · kernels fp16
`hrnet-w48-fcn-ade20k`	228 ms · portable fp16	–	–	–	–	4.64 ms · kernels fp16
`mask2former-swinl-ade20k`	909 ms · portable fp32	77.8 ms · kernels fp16	32.8 ms · kernels fp16	67.4 ms · kernels fp16	35.3 ms · kernels fp16	16.9 ms · kernels fp16
`oneformer-swinl-ade20k`	1127 ms · portable fp32	134 ms · kernels fp16	47.2 ms · kernels fp16	109 ms · kernels fp16	58.9 ms · kernels fp16	31.4 ms · kernels fp16
`pspnet-r101-d8-ade20k`	411 ms · portable fp16	–	–	–	–	10.3 ms · kernels fp16
`sapiens-seg-0.3b-goliath`	1517 ms · portable fp16	135 ms · kernels fp16	39.7 ms · kernels fp16	97.7 ms · kernels fp16	34.4 ms · kernels fp16	24.4 ms · portable fp16
`sapiens-seg-1b-goliath`	3113 ms · portable fp16	305 ms · kernels fp16	81.4 ms · kernels fp16	200 ms · kernels fp16	67.8 ms · kernels fp16	45.5 ms · portable fp16
`segformer-b0-ade20k`	106 ms · portable fp32	3.91 ms · kernels fp16	1.76 ms · kernels fp16	3.79 ms · kernels fp16	3.43 ms · kernels fp16	1.24 ms · kernels fp16
`segformer-b5-ade20k`	428 ms · portable fp32	34.5 ms · kernels fp16	9.82 ms · kernels fp16	28.5 ms · kernels fp16	18.3 ms · kernels fp16	8.22 ms · kernels fp16

Promptable segmentation, 1024x1024

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`sam-vit-b`	37.0 ms · kernels fp16	631 ms · portable fp32	40.7 ms · kernels fp16	15.7 ms · kernels fp16	35.1 ms · kernels fp16	19.2 ms · kernels fp16	10.3 ms · kernels fp16
`sam-vit-h`	143 ms · kernels fp16	2518 ms · portable fp32	239 ms · portable fp16	68.4 ms · portable fp16	182 ms · portable fp16	61.3 ms · kernels fp16	35.1 ms · kernels fp16
`sam-vit-l`	81.5 ms · kernels fp16	1373 ms · portable fp32	130 ms · portable fp16	38.6 ms · portable fp16	99.7 ms · portable fp16	37.5 ms · kernels fp16	19.6 ms · kernels fp16
`sam2.1-hiera-base-plus`	25.1 ms · kernels fp16	424 ms · portable fp32	36.8 ms · portable fp16	13.6 ms · portable fp16	34.2 ms · portable fp16	15.8 ms · portable fp16	7.78 ms · portable fp16
`sam2.1-hiera-large`	48.8 ms · kernels fp16	613 ms · portable fp32	80.7 ms · portable fp16	25.0 ms · portable fp16	65.2 ms · portable fp16	27.1 ms · portable fp16	13.8 ms · portable fp16
`sam2.1-hiera-small`	15.6 ms · kernels fp16	177 ms · portable fp32	–	–	–	–	5.43 ms · portable fp16
`sam2.1-hiera-tiny`	13.9 ms · kernels fp16	151 ms · portable fp32	17.6 ms · kernels fp16	8.14 ms · kernels fp16	15.6 ms · kernels fp16	12.0 ms · portable fp16	5.11 ms · kernels fp16
`sam3`	116 ms · kernels fp16	3620 ms · portable fp32	161 ms · kernels fp16	48.6 ms · kernels fp16	136 ms · kernels fp16	50.8 ms · kernels fp16	29.5 ms · kernels fp16

Detection, 800x1344

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`detr-r101-coco`	13.7 ms · kernels fp16	183 ms · portable fp32	12.8 ms · kernels fp16	9.91 ms · kernels fp16	11.6 ms · kernels fp16	10.6 ms · kernels fp16	3.52 ms · kernels fp16
`detr-r50-coco`	9.93 ms · kernels fp16	125 ms · portable fp32	11.8 ms · kernels fp16	8.21 ms · kernels fp16	8.64 ms · kernels fp16	8.82 ms · kernels fp16	3.14 ms · kernels fp16
`dfine-l-coco`	–	–	9.47 ms · kernels fp16	5.16 ms · kernels fp16	9.48 ms · kernels fp16	9.36 ms · kernels fp16	4.23 ms · kernels fp16
`dfine-m-coco`	–	–	–	–	–	–	3.30 ms · kernels fp16
`dfine-n-coco`	–	–	3.45 ms · kernels fp16	2.41 ms · kernels fp16	3.68 ms · kernels fp16	4.06 ms · kernels fp16	1.89 ms · kernels fp16
`dfine-s-coco`	–	–	–	–	–	–	2.37 ms · kernels fp16
`dfine-x-coco`	–	–	14.8 ms · kernels fp16	6.81 ms · kernels fp16	14.4 ms · kernels fp16	13.1 ms · kernels fp16	6.05 ms · kernels fp16
`faster-rcnn-r50-fpn-3x-coco`	19.5 ms · kernels fp16	214 ms · portable fp32	19.6 ms · kernels fp16	10.6 ms · kernels fp16	18.1 ms · kernels fp16	13.0 ms · kernels fp16	5.86 ms · kernels fp16
`mask-rcnn-r50-fpn-3x-coco`	21.0 ms · kernels fp16	222 ms · portable fp32	22.1 ms · kernels fp16	11.8 ms · kernels fp16	20.1 ms · kernels fp16	14.7 ms · kernels fp16	6.55 ms · kernels fp16
`rtdetrv2-l-coco`	6.61 ms · kernels fp16	82.6 ms · portable fp32	7.20 ms · kernels fp16	3.82 ms · kernels fp16	7.63 ms · kernels fp16	6.65 ms · kernels fp16	3.31 ms · kernels fp16
`rtdetrv2-m-coco`	5.27 ms · kernels fp16	55.3 ms · portable fp32	–	–	–	–	2.68 ms · kernels fp16
`rtdetrv2-ms-coco`	5.46 ms · kernels fp16	67.9 ms · portable fp32	–	–	–	–	2.57 ms · kernels fp16
`rtdetrv2-s-coco`	4.19 ms · kernels fp16	44.0 ms · portable fp32	4.56 ms · kernels fp16	2.76 ms · kernels fp16	4.88 ms · kernels fp16	4.74 ms · kernels fp16	2.21 ms · kernels fp16
`rtdetrv2-x-coco`	9.24 ms · kernels fp16	114 ms · portable fp32	10.1 ms · kernels fp16	4.89 ms · kernels fp16	10.4 ms · kernels fp16	8.11 ms · kernels fp16	4.07 ms · kernels fp16

Features, 224x224

model	RTX 5060 Ti 16 GB	Apple M1 Pro 16 GB	NVIDIA L4	NVIDIA L40S	NVIDIA A10	NVIDIA A100-SXM4-40GB	NVIDIA RTX PRO 6000
`dinov2-vitb14`	–	21.7 ms · portable fp16	–	–	–	–	0.985 ms · kernels fp16
`dinov2-vitb14-reg`	–	42.8 ms · portable fp16	–	–	–	–	0.987 ms · kernels fp16
`dinov2-vitg14`	–	193 ms · portable fp16	22.4 ms · kernels fp16	6.66 ms · kernels fp16	18.0 ms · kernels fp16	9.51 ms · kernels fp16	4.95 ms · kernels fp16
`dinov2-vitg14-reg`	13.0 ms · kernels fp16	235 ms · portable fp16	–	–	–	–	4.94 ms · kernels fp16
`dinov2-vitl14`	–	59.4 ms · portable fp16	6.34 ms · kernels fp16	2.39 ms · kernels fp16	5.34 ms · portable fp16	3.77 ms · kernels fp16	2.00 ms · kernels fp16
`dinov2-vitl14-reg`	–	86.3 ms · portable fp16	–	–	–	–	2.00 ms · kernels fp16
`dinov2-vits14`	–	11.2 ms · portable fp16	1.05 ms · kernels fp16	0.721 ms · kernels fp16	0.996 ms · kernels fp16	1.40 ms · kernels fp16	0.645 ms · kernels fp16
`dinov2-vits14-reg`	–	22.1 ms · portable fp16	–	–	–	–	0.639 ms · kernels fp16
`dinov3-convnext-base`	–	failed	–	–	–	–	1.78 ms · portable fp16
`dinov3-convnext-large`	–	failed	4.42 ms · kernels fp16	2.36 ms · kernels fp16	3.73 ms · kernels fp16	4.10 ms · kernels fp16	1.88 ms · portable fp16
`dinov3-convnext-small`	–	21.7 ms · portable fp16	–	–	–	–	1.78 ms · kernels fp16
`dinov3-convnext-tiny`	–	failed	1.46 ms · kernels fp16	0.965 ms · kernels fp16	1.34 ms · kernels fp16	1.79 ms · kernels fp16	0.973 ms · kernels fp16
`dinov3-vitb16`	–	22.7 ms · portable fp16	–	–	–	–	0.865 ms · kernels fp16
`dinov3-vith16plus`	–	failed	14.2 ms · kernels fp16	4.46 ms · kernels fp16	11.4 ms · kernels fp16	6.54 ms · kernels fp16	3.56 ms · kernels fp16
`dinov3-vitl16`	–	failed	5.22 ms · kernels fp16	2.27 ms · kernels fp16	3.76 ms · kernels fp16	3.64 ms · kernels fp16	1.99 ms · kernels fp16
`dinov3-vits16`	–	failed	1.02 ms · kernels fp16	0.727 ms · kernels fp16	1.02 ms · kernels fp16	1.47 ms · kernels fp16	0.654 ms · kernels fp16

RTX 5060 Ti 16 GB (reference)

Stereo, KITTI 2015, 375x1242

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
aanet-plus-sceneflow	65.7	55.4	55.9	32.6	1.7x	1.7x	1139 / 1210	469 / 518	287 / 380
aanet-sceneflow	75.4	47.9	45.9	30.6	1.6x	1.5x	363 / 428	310 / 366	208 / 264
bridgedepth-rvc-pretrain	169	121	114	94.7	1.3x	1.2x	1891 / 2006	1910 / 2006	1323 / 1446
bridgedepth-sf	164	118	111	92.1	1.3x	1.2x	1840 / 1970	1859 / 1954	1272 / 1394
c-fast-foundationstereo	203	143	97.7	50.9	2.8x	1.9x	1006 / 1138	733 / 914	438 / 554
c-foundationstereo-s	2499	1121	1074	431	2.6x	2.5x	2781 / 3204	3085 / 3450	1785 / 2036
cas-gwcnet-c-sceneflow	195	115	96.7	65.8	1.8x	1.5x	917 / 990	668 / 834	581 / 670
cas-gwcnet-gc-sceneflow	250	161	138	103	1.6x	1.3x	2634 / 2772	772 / 954	751 / 834
cfnet-sceneflow	194	155	112	89.8	1.7x	1.3x	2211 / 2584	1622 / 2016	1692 / 2116
coex-sceneflow	26.0	16.6	14.8	6.60	2.5x	2.2x	671 / 698	387 / 434	163 / 486
crestereo-eth3d	636	480	209	83.1	5.8x	2.5x	516 / 642	315 / 370	245 / 292
defom-stereo-vitl-sceneflow	947	509	404	184	2.8x	2.2x	2458 / 2610	2394 / 2538	1604 / 1846
dlnr-middlebury	292	168	117	71.5	2.4x	1.6x	1173 / 1334	915 / 1082	495 / 606
dlnr-sceneflow	292	168	117	71.5	2.4x	1.6x	1173 / 1334	915 / 1082	495 / 606
fast-acvnet-generalization	28.1	24.1	17.4	10.5	2.3x	1.7x	461 / 486	219 / 256	225 / 338
fast-acvnet-kitti12	28.1	24.1	17.4	10.5	2.3x	1.7x	461 / 486	219 / 256	225 / 338
fast-acvnet-kitti15	28.1	24.1	17.4	10.5	2.3x	1.7x	461 / 486	219 / 256	225 / 338
fast-acvnet-plus-generalization	35.4	26.4	19.5	12.0	2.2x	1.6x	468 / 506	218 / 256	246 / 358
fast-acvnet-plus-kitti12	35.4	26.4	19.5	12.0	2.2x	1.6x	468 / 506	218 / 256	246 / 358
fast-acvnet-plus-kitti15	35.4	26.4	19.5	12.0	2.2x	1.6x	468 / 506	218 / 256	246 / 358
fast-acvnet-plus-sceneflow	35.4	26.4	19.5	12.0	2.2x	1.6x	468 / 506	218 / 256	246 / 358
fast-acvnet-sceneflow	28.1	24.1	17.4	10.5	2.3x	1.7x	461 / 486	219 / 256	225 / 338
fast-foundationstereo	203	142	97.8	51.0	2.8x	1.9x	1052 / 1108	733 / 774	438 / 554
foundationstereo-vit-l	2801	1272	905	491	2.6x	1.8x	3614 / 3754	2154 / 2294	1588 / 1658
foundationstereo-vit-s	2139	1052	737	358	2.9x	2.1x	2388 / 2484	1542 / 1744	977 / 1090
great-igev-eth3d	659	425	301	126	3.4x	2.4x	2831 / 3074	1039 / 1258	572 / 704
great-igev-kitti	659	425	301	126	3.4x	2.4x	2831 / 3074	1039 / 1258	572 / 704
great-igev-middlebury	1055	705	492	175	4.0x	2.8x	10728 / 10974	1165 / 1298	1093 / 1184
great-igev-sceneflow	659	425	301	126	3.4x	2.4x	2831 / 3074	1039 / 1258	572 / 704
gwcnet-g-sceneflow	175	104	104	72.7	1.4x	1.4x	1622 / 1734	686 / 882	598 / 782
gwcnet-gc-sceneflow	183	107	107	78.3	1.4x	1.4x	1867 / 1976	687 / 982	567 / 642
hitnet-eth3d	63.1	46.0	22.4	7.41	6.2x	3.0x	746 / –	341 / 510	208 / 728
hitnet-middlebury-d400	177	130	52.6	17.0	7.7x	3.1x	2124 / –	698 / 1074	425 / 1428
igev-plusplus-eth3d	575	383	297	126	3.0x	2.4x	976 / 1054	540 / 720	453 / 720
igev-plusplus-kitti12	575	383	297	126	3.0x	2.4x	976 / 1054	540 / 720	453 / 720
igev-plusplus-kitti15	575	383	297	126	3.0x	2.4x	976 / 1054	540 / 720	453 / 720
igev-plusplus-middlebury	575	383	297	126	3.0x	2.4x	976 / 1054	540 / 720	453 / 720
igev-plusplus-sceneflow	575	383	297	126	3.0x	2.4x	976 / 1054	540 / 720	453 / 720
igev-stereo-eth3d	449	279	213	96.4	2.9x	2.2x	626 / 702	426 / 532	393 / 494
igev-stereo-kitti12	449	279	213	96.4	2.9x	2.2x	626 / 702	426 / 532	393 / 494
igev-stereo-kitti15	449	279	213	96.4	2.9x	2.2x	626 / 702	426 / 532	393 / 494
igev-stereo-middlebury	449	279	213	96.4	2.9x	2.2x	626 / 702	426 / 532	393 / 494
igev-stereo-sceneflow	449	279	213	96.4	2.9x	2.2x	626 / 702	426 / 532	393 / 494
las2-h	51.1	36.4	37.5	14.9	2.4x	2.5x	352 / 386	387 / 430	311 / 528
las2-l	46.9	30.1	30.9	8.96	3.4x	3.4x	422 / 488	424 / 458	169 / 380
las2-m	23.9	20.0	23.2	5.28	3.8x	4.4x	360 / 404	360 / 386	136 / 332
las2-s	19.3	17.3	20.5	3.90	4.4x	5.3x	345 / 382	343 / 368	128 / 338
mocha-v1-sceneflow	648	451	321	165	2.7x	1.9x	1430 / 1608	668 / 750	657 / 750
mocha-v2-sceneflow	615	398	315	164	2.4x	1.9x	1411 / 1588	668 / 750	656 / 750
monster-mix-all	993	995	419	187	5.3x	2.2x	2733 / 2958	1262 / 1594	1282 / 1634
monster-plusplus-mix-all	1012	1009	432	195	5.2x	2.2x	2810 / 3018	1339 / 1618	1308 / 1660
monster-sceneflow	993	995	419	187	5.3x	2.2x	2733 / 2958	1262 / 1594	1282 / 1634
nmrf-sceneflow	122	78.1	58.1	51.5	1.5x	1.1x	587 / 634	548 / 586	364 / 406
psmnet-sceneflow	248	204	107	73.5	2.8x	1.5x	1950 / 1990	700 / 822	681 / 862
raft-stereo-eth3d	543	291	215	87.7	3.3x	2.4x	816 / 872	476 / 552	282 / 362
raft-stereo-middlebury	543	291	215	87.7	3.3x	2.4x	816 / 872	476 / 552	282 / 362
raft-stereo-sceneflow	543	291	215	87.7	3.3x	2.4x	816 / 872	476 / 552	282 / 362
rt-monster-plusplus-mix-all	91.8	91.3	37.4	31.4	2.9x	1.2x	1051 / 1106	892 / 926	899 / 1116
s2m2-l	597	283	288	120	2.4x	2.4x	1747 / 1998	1510 / 1718	831 / 988
s2m2-m	356	184	184	78.2	2.4x	2.4x	1080 / 1284	952 / 1126	491 / 542
s2m2-s	163	99.1	91.5	38.3	2.6x	2.4x	676 / 748	633 / 750	287 / 400
s2m2-xl	1063	529	536	212	2.5x	2.5x	3315 / 3484	2722 / 2976	1502 / 1614
selective-igev-eth3d-finetune	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-eth3d-train	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-kitti12	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-kitti15	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-middlebury-finetune	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-middlebury-train	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
selective-igev-sceneflow	637	422	279	114	3.7x	2.4x	698 / 740	427 / 548	389 / 490
waft-stereo-b	380	208	208	93.6	2.2x	2.2x	1674 / 1860	1610 / 1732	851 / 936
waft-stereo-s	234	128	129	65.2	2.0x	2.0x	1326 / 1480	1281 / 1356	489 / 566

Stereo, Middlebury 2014 half, 994x1482

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
aanet-plus-sceneflow	263	194	159	88.9	2.2x	1.8x	3652 / 3784	1142 / 1278	813 / 920
aanet-sceneflow	289	178	140	101	1.8x	1.4x	1000 / 1126	845 / 986	536 / 644
bridgedepth-rvc-pretrain	629	470	442	354	1.3x	1.2x	2903 / 3200	2968 / 3246	2361 / 2686
bridgedepth-sf	616	460	437	348	1.3x	1.3x	2852 / 3144	2916 / 3214	2310 / 2634
c-fast-foundationstereo	741	527	330	135	3.9x	2.4x	2892 / 2996	2109 / 2336	1196 / 1516
c-foundationstereo-s	–	3874	3646	1362	2.8x	2.7x	8338 / 12870	6312 / 6590	3107 / 4376
cas-gwcnet-c-sceneflow	656	389	327	225	1.7x	1.5x	2791 / 2970	2006 / 2474	1725 / 1930
cas-gwcnet-gc-sceneflow	862	567	472	358	1.6x	1.3x	8317 / 8730	2338 / 2794	2272 / 2514
cfnet-sceneflow	659	545	484	401	1.4x	1.2x	6765 / 7880	3441 / 4376	3662 / 4718
coex-sceneflow	94.9	60.0	49.9	23.5	2.6x	2.1x	2098 / 2160	1136 / 1254	495 / 886
crestereo-eth3d	1980	1597	732	229	7.0x	3.2x	1541 / 1752	884 / 1030	644 / 732
defom-stereo-vitl-sceneflow	4175	2816	1531	659	4.3x	2.3x	8639 / 8766	4362 / 4838	3154 / 3566
dlnr-middlebury	978	598	398	214	2.8x	1.9x	3004 / 3552	2635 / 3142	1236 / 1606
dlnr-sceneflow	978	598	398	214	2.8x	1.9x	3004 / 3552	2635 / 3142	1236 / 1606
fast-acvnet-generalization	101	60.6	54.7	38.2	1.6x	1.4x	1421 / 1506	664 / 736	739 / 1042
fast-acvnet-kitti12	101	60.6	54.7	38.2	1.6x	1.4x	1421 / 1506	664 / 736	739 / 1042
fast-acvnet-kitti15	101	60.6	54.7	38.2	1.6x	1.4x	1421 / 1506	664 / 736	739 / 1042
fast-acvnet-plus-generalization	138	80.9	70.6	46.4	1.7x	1.5x	1441 / 1522	662 / 756	790 / 1120
fast-acvnet-plus-kitti12	138	80.9	70.6	46.4	1.7x	1.5x	1441 / 1522	662 / 756	790 / 1120
fast-acvnet-plus-kitti15	138	80.9	70.6	46.4	1.7x	1.5x	1441 / 1522	662 / 756	790 / 1120
fast-acvnet-plus-sceneflow	138	80.9	70.6	46.4	1.7x	1.5x	1441 / 1522	662 / 756	790 / 1120
fast-acvnet-sceneflow	101	60.6	54.7	38.2	1.6x	1.4x	1421 / 1506	664 / 736	739 / 1042
fast-foundationstereo	741	527	330	135	3.9x	2.4x	2939 / 3306	2109 / 2256	1196 / 1576
foundationstereo-vit-l	9228	4237	2990	1513	2.8x	2.0x	8286 / 8750	3141 / 3436	2818 / 3300
foundationstereo-vit-s	7129	3609	2514	1120	3.2x	2.2x	6987 / 7280	2493 / 2606	2169 / 2470
great-igev-eth3d	2252	1458	1085	393	3.7x	2.8x	8861 / 9206	1402 / 1678	1192 / 1564
great-igev-kitti	2252	1458	1085	393	3.7x	2.8x	8861 / 9206	1402 / 1678	1192 / 1564
great-igev-middlebury	OOM	OOM	1785	579	–	3.1x	–	1875 / 2078	1553 / 1744
great-igev-sceneflow	2252	1458	1085	393	3.7x	2.8x	8861 / 9206	1402 / 1678	1192 / 1564
gwcnet-g-sceneflow	604	352	344	238	1.5x	1.4x	5105 / 5414	2081 / 2402	1799 / 2102
gwcnet-gc-sceneflow	632	365	353	257	1.4x	1.4x	5887 / 6214	2082 / 2402	1768 / 1962
hitnet-eth3d	234	182	67.8	23.2	7.9x	2.9x	2433 / –	860 / 1090	647 / 1248
hitnet-middlebury-d400	637	483	176	52.7	9.2x	3.3x	6831 / –	1744 / 2414	1263 / 2368
igev-plusplus-eth3d	2012	1278	1081	393	3.3x	2.7x	2891 / 3108	1594 / 1840	1389 / 2302
igev-plusplus-kitti12	2012	1278	1081	393	3.3x	2.7x	2891 / 3108	1594 / 1840	1389 / 2302
igev-plusplus-kitti15	2012	1278	1081	393	3.3x	2.7x	2891 / 3108	1594 / 1840	1389 / 2302
igev-plusplus-middlebury	2012	1278	1081	393	3.3x	2.7x	2891 / 3108	1594 / 1840	1389 / 2302
igev-plusplus-sceneflow	2012	1278	1081	393	3.3x	2.7x	2891 / 3108	1594 / 1840	1389 / 2302
igev-stereo-eth3d	1518	941	763	288	3.3x	2.6x	1789 / 1936	1237 / 1492	1113 / 1374
igev-stereo-kitti12	1518	941	763	288	3.3x	2.6x	1789 / 1936	1237 / 1492	1113 / 1374
igev-stereo-kitti15	1518	941	763	288	3.3x	2.6x	1789 / 1936	1237 / 1492	1113 / 1374
igev-stereo-middlebury	1518	941	763	288	3.3x	2.6x	1789 / 1936	1237 / 1492	1113 / 1374
igev-stereo-sceneflow	1518	941	763	288	3.3x	2.6x	1789 / 1936	1237 / 1492	1113 / 1374
las2-h	205	135	123	53.3	2.5x	2.3x	1022 / 1084	1050 / 1110	1021 / 1388
las2-l	154	90.8	80.3	29.9	3.0x	2.7x	1010 / 1082	1036 / 1098	397 / 600
las2-m	79.1	52.4	54.1	18.3	2.9x	3.0x	948 / 1000	972 / 1026	329 / 572
las2-s	62.0	43.5	46.9	14.3	3.0x	3.3x	933 / 978	955 / 988	320 / 558
mocha-v1-sceneflow	2148	1456	1057	455	3.2x	2.3x	3637 / 3982	1632 / 1770	1588 / 1730
mocha-v2-sceneflow	2054	1295	1033	448	2.9x	2.3x	3576 / 3982	1632 / 1770	1587 / 1790
monster-mix-all	4275	4277	1585	668	6.4x	2.4x	6382 / 6430	2336 / 2494	2356 / 2714
monster-plusplus-mix-all	4339	4334	1633	698	6.2x	2.3x	6382 / 6430	2583 / 2918	2438 / 2760
monster-sceneflow	4275	4277	1585	668	6.4x	2.4x	6382 / 6430	2336 / 2494	2356 / 2714
nmrf-sceneflow	445	313	216	196	1.6x	1.1x	1656 / 2054	1510 / 1552	1022 / 1112
psmnet-sceneflow	760	608	340	232	2.6x	1.5x	6015 / 6090	2078 / 2442	1786 / 2122
raft-stereo-eth3d	1775	1037	746	273	3.8x	2.7x	2516 / 2628	1388 / 1492	765 / 802
raft-stereo-middlebury	1775	1037	746	273	3.8x	2.7x	2516 / 2628	1388 / 1492	765 / 802
raft-stereo-sceneflow	1775	1037	746	273	3.8x	2.7x	2516 / 2628	1388 / 1492	765 / 802
rt-monster-plusplus-mix-all	252	251	113	94.1	2.7x	1.2x	2922 / 3044	2592 / 2646	2704 / 3976
s2m2-l	1934	1005	1015	376	2.7x	2.7x	3248 / 4090	3245 / 3838	1249 / 1568
s2m2-m	1242	666	670	248	2.7x	2.7x	2371 / 2920	2300 / 2786	876 / 1162
s2m2-s	582	350	348	113	3.1x	3.1x	1759 / 1986	1733 / 1930	686 / 820
s2m2-xl	3510	1796	1815	682	2.6x	2.7x	5606 / 6080	5228 / 6016	2092 / 2274
selective-igev-eth3d-finetune	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-eth3d-train	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-kitti12	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-kitti15	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-middlebury-finetune	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-middlebury-train	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
selective-igev-sceneflow	2201	1455	998	349	4.2x	2.9x	2013 / 2152	1238 / 1488	1109 / 1390
waft-stereo-b	1898	1140	1137	393	2.9x	2.9x	5743 / 5876	5775 / 5852	1907 / 2316
waft-stereo-s	1252	689	689	284	2.4x	2.4x	3314 / 3398	3293 / 3456	1113 / 1286

Depth, Middlebury 2014 half, 994x1482

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
da3-giant-1.1	410	227	81.8	44.2	5.1x	1.9x	8590 / 8718	2638 / 4608	2628 / 2710
da3-large-1.1	208	166	37.4	17.1	9.7x	2.2x	3374 / 3472	992 / 1350	985 / 1090
da3metric-large	128	85.1	24.1	14.7	5.8x	1.6x	2996 / 3050	815 / 1328	811 / 868
da3mono-large	127	85.2	23.9	14.6	5.8x	1.6x	2997 / 3052	818 / 1328	814 / 868
da3nested-giant-large-1.1	533	308	108	58.2	5.3x	1.9x	10332 / 10402	3275 / 5914	3271 / 3378
depth-anything-v2-b	93.6	63.2	18.8	10.5	6.0x	1.8x	928 / 1026	450 / 516	387 / 436
depth-anything-v2-l	245	140	55.1	31.1	4.5x	1.8x	2560 / 2654	1026 / 1352	921 / 1034
depth-anything-v2-s	53.6	41.0	7.82	4.51	9.1x	1.7x	373 / 418	249 / 326	190 / 248
depth-pro	1776	569	518	271	2.1x	1.9x	3828 / 4294	3306 / 3498	3667 / 3778
marigold-depth-v1-1	963	439	408	383	1.1x	1.1x	3414 / 3580	2817 / 3760	2583 / 3762
metric3d-vit-large	629	438	200	158	2.8x	1.3x	3275 / 3350	1484 / 1696	1516 / 1688
midas-dpt-beit-large-512	386	333	78.8	63.9	5.2x	1.2x	2759 / 2898	992 / 1418	1003 / 1436
moge-2-vitl-normal	501	171	144	73.4	2.3x	2.0x	2339 / 2438	1661 / 1836	1498 / 1610
moge-2-vits-normal	162	82.6	64.2	29.7	2.8x	2.2x	1193 / 1268	1066 / 1336	884 / 1008
moge-3-vitl	843	486	514	215	2.3x	2.4x	2505 / 2630	1686 / 1924	1673 / 1884
unidepth-v2-vitl14	132	132	99.8	66.4	2.0x	1.5x	3230 / 3508	1208 / 1472	1211 / 1472
unik3d-vitl	134	134	108	62.9	2.1x	1.7x	3307 / 3570	1219 / 1492	1166 / 1266
zoedepth-nk	408	271	124	88.9	3.1x	1.4x	2040 / 2146	2037 / 2124	1771 / 1960

Depth, synthetic, 518x770

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
da3-giant-1.1	406	224	81.5	44.2	5.1x	1.8x	8590 / 8718	2621 / 4608	2527 / 2710
da3-large-1.1	205	163	36.9	16.7	9.7x	2.2x	3374 / 3472	975 / 1350	883 / 1090
da3metric-large	125	82.7	23.7	14.3	5.8x	1.7x	2996 / 3050	798 / 1328	794 / 868
da3mono-large	125	82.7	23.5	14.2	5.8x	1.7x	2997 / 3052	801 / 1328	797 / 868
da3nested-giant-large-1.1	529	307	105	58.3	5.3x	1.8x	10332 / 10402	3259 / 5914	3171 / 3378
depth-anything-v2-b	88.8	53.8	18.8	10.6	5.1x	1.8x	928 / 1026	433 / 516	370 / 436
depth-anything-v2-l	240	136	55.0	31.6	4.3x	1.7x	2560 / 2654	1009 / 1352	904 / 1034
depth-anything-v2-s	45.5	38.1	7.73	4.45	8.6x	1.7x	373 / 418	232 / 326	173 / 248
depth-pro	1769	566	519	271	2.1x	1.9x	3822 / 4354	3289 / 3498	3650 / 3778
marigold-depth-v1-1	962	429	408	384	1.1x	1.1x	3414 / 3580	2800 / 3760	2566 / 3762
metric3d-vit-large	630	439	200	158	2.8x	1.3x	3275 / 3370	1467 / 1696	1499 / 1688
midas-dpt-beit-large-512	399	320	75.6	63.9	5.0x	1.2x	2759 / 2858	975 / 1418	986 / 1436
moge-2-vitl-normal	500	169	145	73.9	2.3x	2.0x	2326 / 2418	1645 / 1836	1481 / 1610
moge-2-vits-normal	160	81.2	61.3	32.9	2.5x	1.9x	1180 / 1288	1049 / 1336	867 / 1008
moge-3-vitl	734	324	354	150	2.2x	2.4x	2493 / 2594	1669 / 1804	1506 / 1662
unidepth-v2-vitl14	78.5	78.6	59.6	40.7	1.9x	1.5x	2820 / 3030	1053 / 1472	1061 / 1472
unik3d-vitl	123	123	99.2	58.2	2.1x	1.7x	3220 / 3530	1172 / 1492	1130 / 1266
zoedepth-nk	295	250	114	80.0	3.1x	1.4x	1965 / 2124	1950 / 2124	1658 / 1960

Classification, synthetic, 224x224

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
alexnet.tv_in1k	1.17	1.82	0.970	0.604	1.9x	1.6x	383 / 404	151 / 224	184 / 228
convnextv2_base.fcmae_ft_in22k_in1k	10.7	14.6	10.2	2.94	3.7x	3.5x	542 / 566	299 / 376	318 / 380
deit_base_patch16_224.fb_in1k	6.39	5.46	2.98	1.64	3.3x	1.8x	531 / 548	203 / 386	300 / 370
densenet121.tv_in1k	10.7	13.6	9.41	1.28	8.4x	7.4x	80 / 102	53 / 90	83 / 132
dinov2_vitl14_lc	22.4	10.8	5.43	4.93	2.2x	1.1x	1782 / 1806	623 / 1228	890 / 950
efficientnet_b0.tv_in1k	5.64	7.91	4.98	0.769	7.3x	6.5x	64 / 84	48 / 96	78 / 118
efficientnet_b4.tv_in1k	11.1	15.7	10.0	2.00	5.5x	5.0x	147 / 174	89 / 114	111 / 180
efficientnet_v2_s.tv_in1k	12.1	16.6	10.2	1.89	6.4x	5.4x	159 / 176	81 / 106	114 / 188
efficientvit_b2.r288_in1k	10.7	14.7	11.9	1.10	9.7x	10.8x	173 / 194	85 / 124	120 / 212
eva02_large_patch14_448.mim_m38m_ft_in22k_in1k	83.7	32.1	18.0	14.1	2.3x	1.3x	1819 / 1864	648 / 1206	1236 / 1386
fastvit_sa24.apple_dist_in1k	11.9	16.5	6.88	1.39	8.6x	4.9x	157 / 176	81 / 108	112 / 190
inception_v3.tv_in1k	7.88	10.3	5.21	1.30	6.1x	4.0x	185 / 206	120 / 166	150 / 228
mobilenet_v2.tv2_in1k	3.60	5.14	2.95	0.439	8.2x	6.7x	54 / 72	43 / 70	75 / 92
mobilenet_v3_large.tv2_in1k	4.33	6.20	3.80	0.545	7.9x	7.0x	65 / 80	46 / 76	78 / 114
mobilenetv1_100.ra4_e3600_r224_in1k	2.19	2.96	1.77	0.370	5.9x	4.8x	58 / 74	44 / 72	76 / 114
mobilenetv4_conv_medium.e500_r256_in1k	5.78	7.77	3.91	0.641	9.0x	6.1x	90 / 106	54 / 86	87 / 124
regnety_160.tv2_in1k	12.1	12.2	8.34	1.86	6.5x	4.5x	512 / 524	208 / 250	229 / 316
resnet152.tv2_in1k	12.8	16.3	8.36	1.83	7.0x	4.6x	378 / 390	155 / 186	182 / 266
resnet18.tv_in1k	1.91	2.44	1.48	0.459	4.2x	3.2x	105 / 130	61 / 90	90 / 150
resnet50.tv2_in1k	4.74	5.84	3.15	0.826	5.7x	3.8x	180 / 192	89 / 142	116 / 162
shufflenet_v2_x1_0.tv_in1k	4.79	6.53	3.71	0.787	6.1x	4.7x	47 / 48	41 / 72	87 / 116
squeezenet1_1.tv_in1k	1.72	2.63	1.82	0.359	4.8x	5.1x	10 / 30	41 / 72	8 / 12
swin_base_patch4_window7_224.ms_in22k_ft_in1k	15.0	19.1	10.8	2.15	7.0x	5.0x	538 / 550	213 / 394	246 / 300
vgg16.tv_in1k	3.62	4.09	1.99	1.50	2.4x	1.3x	825 / 846	313 / 372	332 / 388
vit_base_patch16_224.augreg2_in21k_ft_in1k	6.33	5.48	3.02	1.65	3.3x	1.8x	531 / 548	205 / 390	300 / 370
vit_so400m_patch14_siglip_378.webli_ft_in1k	76.9	25.8	13.1	12.8	2.0x	1.0x	2495 / 2564	879 / 1708	1216 / 1346

Segmentation, synthetic, 1024x2048

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
deeplabv3-r101-d8-ade20k	108	54.8	47.9	20.1	2.7x	2.4x	2031 / 2162	453 / 532	362 / 518
deeplabv3plus-r101-d8-ade20k	104	53.4	47.1	18.0	3.0x	2.6x	1891 / 2024	413 / 492	365 / 560
dinov2-b-linear-ade20k	115	40.1	33.3	22.6	1.8x	1.5x	2063 / 2426	765 / 884	771 / 904
dinov2-l-linear-ade20k	332	95.7	83.9	56.8	1.7x	1.5x	3308 / 3726	1181 / 1244	1189 / 1244
dinov2-s-linear-ade20k	55.8	24.7	18.3	12.4	2.0x	1.5x	1694 / 2046	642 / 696	645 / 676
eomt-l-dinov2-ade20k	178	63.7	50.8	35.0	1.8x	1.5x	3827 / 4086	1515 / 1584	1515 / 1584
eomt-l-dinov3-ade20k	186	72.1	54.6	38.8	1.9x	1.4x	3822 / 4086	1514 / 1584	1514 / 1584
fcn-r101-d8-ade20k	94.2	47.0	40.7	16.6	2.8x	2.4x	1925 / 2102	417 / 492	294 / 478
hrnet-w48-fcn-ade20k	48.1	54.8	26.9	11.9	4.1x	2.3x	1915 / 2010	424 / 490	274 / 430
mask2former-swinl-ade20k	244	145	103	60.4	2.4x	1.7x	3572 / 3878	1164 / 1400	1164 / 1400
oneformer-swinl-ade20k	307	205	139	100.0	2.1x	1.4x	7395 / 7792	2054 / 2476	2054 / 2476
pspnet-r101-d8-ade20k	92.5	47.2	48.8	25.0	1.9x	2.0x	1922 / 2222	417 / 492	341 / 494
sapiens-seg-0.3b-goliath	466	173	132	92.6	1.9x	1.4x	2520 / 2724	1643 / 1762	1643 / 1762
sapiens-seg-1b-goliath	1189	354	275	185	1.9x	1.5x	5721 / 5906	3259 / 3384	3259 / 3382
segformer-b0-ade20k	17.2	17.4	11.4	3.48	5.0x	3.3x	1590 / 1766	339 / 376	148 / 238
segformer-b5-ade20k	132	72.0	38.7	25.3	2.8x	1.5x	2531 / 2620	546 / 590	407 / 534

Promptable segmentation, synthetic, 1024x1024

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
sam-vit-b	168	98.2	48.3	37.0	2.7x	1.3x	1769 / 1802	378 / 420	377 / 420
sam-vit-h	734	306	214	143	2.1x	1.5x	5295 / 5404	1484 / 2524	1480 / 1524
sam-vit-l	406	197	116	81.5	2.4x	1.4x	3411 / 3464	833 / 1262	831 / 862
sam2.1-hiera-base-plus	107	47.2	32.0	25.1	1.9x	1.3x	823 / 870	412 / 464	426 / 524
sam2.1-hiera-large	242	86.4	68.0	48.8	1.8x	1.4x	1656 / 1724	742 / 768	760 / 828
sam2.1-hiera-small	74.8	41.8	18.8	15.6	2.7x	1.2x	618 / 662	318 / 366	330 / 386
sam2.1-hiera-tiny	66.3	38.0	16.6	13.9	2.7x	1.2x	577 / 636	304 / 382	316 / 362
sam3	695	281	179	116	2.4x	1.5x	5255 / 5334	1922 / 2004	2019 / 2264

Detection, synthetic, 800x1344

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
detr-r101-coco	78.9	80.8	18.7	13.7	5.8x	1.4x	531 / 650	316 / 368	254 / 328
detr-r50-coco	55.5	62.0	12.4	9.93	5.6x	1.2x	459 / 558	280 / 332	218 / 272
dfine-l-coco	37.1	44.9	16.2	8.70	4.3x	1.9x	278 / 350	152 / 172	179 / 300
dfine-m-coco	30.8	38.2	13.5	5.97	5.2x	2.3x	209 / 256	126 / 142	156 / 252
dfine-n-coco	24.3	29.1	10.3	2.72	8.9x	3.8x	92 / 126	88 / 116	124 / 202
dfine-s-coco	25.7	31.9	11.1	4.04	6.4x	2.7x	152 / 184	105 / 128	138 / 212
dfine-x-coco	44.9	54.6	18.5	14.1	3.2x	1.3x	476 / 536	227 / 258	237 / 364
faster-rcnn-r50-fpn-3x-coco	83.6	41.4	22.8	19.5	2.1x	1.2x	626 / 708	390 / 490	391 / 450
mask-rcnn-r50-fpn-3x-coco	85.4	43.5	24.6	21.0	2.1x	1.2x	604 / 650	352 / 440	353 / 400
rtdetrv2-l-coco	30.3	38.4	12.1	6.61	4.6x	1.8x	354 / 412	177 / 200	200 / 306
rtdetrv2-m-coco	17.1	20.9	10.4	5.27	3.2x	2.0x	254 / 270	156 / 170	177 / 274
rtdetrv2-ms-coco	24.7	22.5	9.60	5.46	4.1x	1.8x	309 / 358	166 / 182	189 / 306
rtdetrv2-s-coco	14.1	16.8	8.04	4.19	3.4x	1.9x	192 / 220	134 / 166	155 / 250
rtdetrv2-x-coco	36.3	46.8	14.6	9.24	3.9x	1.6x	589 / 622	244 / 278	262 / 384

Features, synthetic, 224x224

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
dinov2-vitb14	6.90	4.50	2.92	1.48	3.0x	2.0x	530 / 548	205 / 226	237 / 290
dinov2-vitb14-reg	7.00	4.30	2.94	1.50	2.9x	2.0x	530 / 548	207 / 226	237 / 288
dinov2-vitg14	86.4	33.9	18.9	13.2	2.6x	1.4x	6542 / 6548	2216 / 2230	2246 / 2318
dinov2-vitg14-reg	86.6	34.2	18.7	13.0	2.6x	1.4x	6542 / 6568	2217 / 2228	2222 / 4430
dinov2-vitl14	22.8	9.80	6.12	4.59	2.1x	1.3x	1777 / 1784	622 / 646	655 / 714
dinov2-vitl14-reg	22.5	10.7	6.17	4.60	2.3x	1.3x	1777 / 1786	625 / 646	655 / 714
dinov2-vits14	3.30	4.40	2.95	0.848	3.9x	3.5x	160 / 166	79 / 98	112 / 148
dinov2-vits14-reg	4.11	5.59	2.98	0.814	5.0x	3.7x	160 / 168	80 / 98	112 / 142
dinov3-convnext-base	7.38	8.24	5.71	2.18	3.4x	2.6x	535 / 552	208 / 232	200 / 214
dinov3-convnext-large	10.5	9.59	5.45	3.59	2.7x	1.5x	1163 / 1252	419 / 430	441 / 494
dinov3-convnext-small	5.97	8.09	5.42	1.99	3.0x	2.7x	328 / 348	134 / 150	127 / 152
dinov3-convnext-tiny	2.96	4.47	3.04	1.18	2.5x	2.6x	198 / 206	93 / 108	86 / 112
dinov3-vitb16	6.95	9.33	4.86	1.40	5.0x	3.5x	527 / 550	201 / 206	201 / 252
dinov3-vith16plus	48.0	25.9	13.3	9.30	2.8x	1.4x	4849 / 4876	1645 / 1668	1644 / 1694
dinov3-vitl16	19.2	17.7	9.13	3.28	5.4x	2.8x	1772 / 1812	617 / 646	617 / 674
dinov3-vits16	6.47	9.16	4.67	0.903	7.2x	5.2x	158 / 166	77 / 80	77 / 106

Multiview, synthetic

model	input	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
mapanything-apache	336x518	2875	1967	636	544	3.6x	1.2x	10227 / 10952	3697 / 4092	3697 / 4072
mvsanywhere	448x672	273	219	93.2	35.6	6.1x	2.6x	1504 / 1576	931 / 1226	761 / 884
patchmatchnet	1536x1024	248	220	202	29.6	7.4x	6.8x	1228 / 1272	1806 / 2096	740 / 856
vggt-1b	336x518	2502	965	651	458	2.1x	1.4x	8688 / 9158	3691 / 4674	3702 / 4654
vggt-omega-1b-512	336x518	2650	906	726	464	2.0x	1.6x	7495 / 7654	3421 / 3518	3422 / 3518

Startup and first call

load: from the load call until the model is on the GPU; first call: the whole first inference call, including one normal inference (see the speed table for the warm time). Seconds, at each task's standard size (another size is shown in brackets).

model	load (s): original	portable	kernels	first call (s): original	portable	kernels
aanet-plus-sceneflow	0.15 (375x1242)	0.16	0.20†	0.70 (375x1242)	0.17†	0.09†
aanet-sceneflow	0.12 (375x1242)	0.19	0.27†	0.07 (375x1242)	0.14†	0.10†
bridgedepth-rvc-pretrain	5.87 (375x1242)	1.84	0.76†	1.36 (375x1242)	0.45†	0.36†
bridgedepth-sf	4.86 (375x1242)	1.53	0.72†	0.12 (375x1242)	0.44†	0.35†
c-fast-foundationstereo	0.17 (375x1242)	0.24	0.16†	0.17 (375x1242)	0.33†	0.14†
c-foundationstereo-s	1.67 (375x1242)	0.62	0.40†	2.73 (375x1242)	3.65†	1.36†
cas-gwcnet-c-sceneflow	0.48 (375x1242)	1.29	0.18†	0.41 (375x1242)	0.33†	0.23†
cas-gwcnet-gc-sceneflow	0.39	0.97	0.17†	0.64	0.47†	0.36†
cfnet-sceneflow	0.55	0.94	0.21†	0.56	0.48†	0.40†
coex-sceneflow	–	0.16	0.09†	0.63 (375x1242)	0.06†	0.04†
crestereo-eth3d	0.05	0.03	0.03†	1.50	0.75†	0.23†
defom-stereo-vitl-sceneflow	4.24	2.31	1.18†	2.83	1.53†	0.66†
dlnr-middlebury	0.83 (375x1242)	0.63	0.26†	1.71 (375x1242)	0.40†	0.21†
dlnr-sceneflow	0.83 (375x1242)	0.63	0.26†	1.71 (375x1242)	0.40†	0.21†
fast-acvnet-generalization	0.24 (375x1242)	0.18	0.20†	1.73 (375x1242)	0.08†	0.04†
fast-acvnet-kitti12	0.24 (375x1242)	0.18	0.20†	1.73 (375x1242)	0.08†	0.04†
fast-acvnet-kitti15	0.24 (375x1242)	0.18	0.20†	1.73 (375x1242)	0.08†	0.04†
fast-acvnet-plus-generalization	0.23 (375x1242)	0.10	0.10†	0.04 (375x1242)	0.07†	0.05†
fast-acvnet-plus-kitti12	0.23 (375x1242)	0.10	0.10†	0.04 (375x1242)	0.07†	0.05†
fast-acvnet-plus-kitti15	0.23 (375x1242)	0.10	0.10†	0.04 (375x1242)	0.07†	0.05†
fast-acvnet-plus-sceneflow	0.23 (375x1242)	0.10	0.10†	0.04 (375x1242)	0.07†	0.05†
fast-acvnet-sceneflow	0.24 (375x1242)	0.18	0.20†	1.73 (375x1242)	0.08†	0.04†
fast-foundationstereo	0.16 (375x1242)	0.25	0.14†	0.14 (375x1242)	0.33†	0.14†
foundationstereo-vit-l	10.86 (375x1242)	1.98	1.17†	2.56 (375x1242)	2.99†	1.52†
foundationstereo-vit-s	1.44 (375x1242)	1.03	0.33†	1.06 (375x1242)	2.51†	1.12†
great-igev-eth3d	0.44 (375x1242)	0.32	0.23†	4.60 (375x1242)	1.10†	0.40†
great-igev-kitti	0.44 (375x1242)	0.32	0.23†	4.60 (375x1242)	1.10†	0.40†
great-igev-middlebury	0.39 (375x1242)	3.91	0.61†	4.59 (375x1242)	1.81†	0.59†
great-igev-sceneflow	0.44 (375x1242)	0.32	0.23†	4.60 (375x1242)	1.10†	0.40†
gwcnet-g-sceneflow	0.26 (375x1242)	0.40	0.14†	0.46 (375x1242)	0.34†	0.24†
gwcnet-gc-sceneflow	0.29 (375x1242)	0.39	0.44†	0.12 (375x1242)	0.35†	0.26†
hitnet-eth3d	0.31	0.09	0.07†	2.06	0.08†	0.03†
hitnet-middlebury-d400	0.57	0.13	0.10†	3.51	0.18†	0.05†
igev-plusplus-eth3d	0.42 (375x1242)	0.20	0.16†	2.33 (375x1242)	1.10†	0.40†
igev-plusplus-kitti12	0.42 (375x1242)	0.20	0.16†	2.33 (375x1242)	1.10†	0.40†
igev-plusplus-kitti15	0.42 (375x1242)	0.20	0.16†	2.33 (375x1242)	1.10†	0.40†
igev-plusplus-middlebury	0.42 (375x1242)	0.20	0.16†	2.33 (375x1242)	1.10†	0.40†
igev-plusplus-sceneflow	0.42 (375x1242)	0.20	0.16†	2.33 (375x1242)	1.10†	0.40†
igev-stereo-eth3d	0.32 (375x1242)	0.47	0.15†	2.56 (375x1242)	0.76†	0.29†
igev-stereo-kitti12	0.32 (375x1242)	0.47	0.15†	2.56 (375x1242)	0.76†	0.29†
igev-stereo-kitti15	0.32 (375x1242)	0.47	0.15†	2.56 (375x1242)	0.76†	0.29†
igev-stereo-middlebury	0.32 (375x1242)	0.47	0.15†	2.56 (375x1242)	0.76†	0.29†
igev-stereo-sceneflow	0.32 (375x1242)	0.47	0.15†	2.56 (375x1242)	0.76†	0.29†
las2-h	0.26 (375x1242)	0.18	0.13†	1.21 (375x1242)	0.14†	0.06†
las2-l	0.39 (375x1242)	0.23	0.18†	0.04 (375x1242)	0.08†	0.03†
las2-m	0.17 (375x1242)	0.12	0.10†	0.02 (375x1242)	0.05†	0.02†
las2-s	0.11 (375x1242)	0.09	0.07†	0.02 (375x1242)	0.05†	0.01†
mocha-v1-sceneflow	1.95 (375x1242)	0.79	0.48†	1.85 (375x1242)	1.06†	0.45†
mocha-v2-sceneflow	2.37 (375x1242)	1.30	0.45†	0.40 (375x1242)	1.03†	0.45†
monster-mix-all	5.85 (375x1242)	4.83	0.95†	1.26 (375x1242)	1.59†	0.67†
monster-plusplus-mix-all	8.37 (375x1242)	2.45	0.79†	1.23 (375x1242)	1.63†	0.70†
monster-sceneflow	5.85 (375x1242)	4.83	0.95†	1.26 (375x1242)	1.59†	0.67†
nmrf-sceneflow	0.21 (375x1242)	0.10	0.07†	0.77 (375x1242)	0.22†	0.20†
psmnet-sceneflow	0.20 (375x1242)	0.10	0.08†	0.53 (375x1242)	0.35†	0.23†
raft-stereo-eth3d	0.27 (375x1242)	0.25	0.05†	1.45 (375x1242)	0.75†	0.27†
raft-stereo-middlebury	0.27 (375x1242)	0.25	0.05†	1.45 (375x1242)	0.75†	0.27†
raft-stereo-sceneflow	0.27 (375x1242)	0.25	0.05†	1.45 (375x1242)	0.75†	0.27†
rt-monster-plusplus-mix-all	0.99 (375x1242)	0.26	0.23†	0.66 (375x1242)	0.13†	0.10†
s2m2-l	1.38	0.99	0.58†	1.01	1.01†	0.37†
s2m2-m	3.95 (375x1242)	0.50	0.31†	0.27 (375x1242)	0.68†	0.25†
s2m2-s	0.22 (375x1242)	0.22	0.20†	0.14 (375x1242)	0.35†	0.11†
s2m2-xl	4.50 (375x1242)	2.73	0.90†	0.53 (375x1242)	1.81†	0.68†
selective-igev-eth3d-finetune	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-eth3d-train	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-kitti12	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-kitti15	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-middlebury-finetune	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-middlebury-train	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
selective-igev-sceneflow	0.36 (375x1242)	0.50	0.14†	2.25 (375x1242)	1.02†	0.35†
waft-stereo-b	3.07 (375x1242)	1.65	0.42†	0.44 (375x1242)	1.15†	0.39†
waft-stereo-s	1.60 (375x1242)	0.58	0.30†	0.14 (375x1242)	0.69†	0.28†
da3-giant-1.1	12.39	6.29	0.90†	0.68	0.16	0.06
da3-large-1.1	3.46	3.65	0.31†	0.45	0.25	0.03
da3metric-large	2.70	0.35	0.30†	0.28	0.05	0.02
da3mono-large	2.96	4.25	1.44†	0.28	0.09	0.02
da3nested-giant-large-1.1	28.67	13.25	2.73†	1.63	0.55	0.07
depth-anything-v2-b	3.94	0.85	0.20†	0.08	0.02	0.01
depth-anything-v2-l	5.66	2.39	0.91†	0.15	0.08	0.06
depth-anything-v2-s	0.71	0.11	0.08†	0.38	0.03	0.01
depth-pro	17.34	4.67	1.11†	0.74	0.58	0.37
marigold-depth-v1-1	1.32	8.20	1.10†	0.60	0.56	0.52
metric3d-vit-large	9.96	3.45	0.31†	0.88	0.23	0.16
midas-dpt-beit-large-512	5.16	2.49	0.28†	1.14	0.09	0.07
moge-2-vitl-normal	7.24	1.37	0.63†	0.17	0.14	0.13
moge-2-vits-normal	0.88	0.35	0.11†	1.50	0.08	0.10
moge-3-vitl	10.60	1.02	1.66	2.92	0.56	0.30
unidepth-v2-vitl14	3.80	4.16	0.90†	0.38	0.15	0.07
unik3d-vitl	7.28	4.16	0.84†	0.57	0.14	0.07
zoedepth-nk	8.35	1.92	0.33†	1.53	0.28	0.12
alexnet.tv_in1k	2.51	1.30	1.40	0.81	0.02	0.06
convnextv2_base.fcmae_ft_in22k_in1k	2.70	1.23	2.12	0.63	0.03	0.16
deit_base_patch16_224.fb_in1k	2.52	0.72	1.96	0.31	0.00	0.07
densenet121.tv_in1k	1.22	0.41	1.89	0.56	0.09	0.04
dinov2_vitl14_lc	5.80	3.34	3.94	0.32	0.02	0.10
efficientnet_b0.tv_in1k	1.25	0.17	1.53	2.86	0.38	0.06
efficientnet_b4.tv_in1k	0.41	0.37	1.94	1.72	0.06	0.07
efficientnet_v2_s.tv_in1k	1.48	0.35	1.90	0.97	0.04	0.05
efficientvit_b2.r288_in1k	1.84	0.40	1.92	1.89	0.04	0.05
eva02_large_patch14_448.mim_m38m_ft_in22k_in1k	6.98	2.80	3.61	0.39	0.02	0.10
fastvit_sa24.apple_dist_in1k	2.49	0.62	2.42	1.67	0.17	0.05
inception_v3.tv_in1k	1.62	0.28	1.86	0.40	0.04	0.05
mobilenet_v2.tv2_in1k	1.23	0.11	1.51	1.65	0.03	0.05
mobilenet_v3_large.tv2_in1k	1.25	0.34	1.46	1.97	0.03	0.05
mobilenetv1_100.ra4_e3600_r224_in1k	1.34	0.06	1.70	1.50	0.01	0.05
mobilenetv4_conv_medium.e500_r256_in1k	1.56	0.36	1.89	2.29	0.04	0.05
regnety_160.tv2_in1k	3.07	1.08	2.22	0.48	0.04	0.06
resnet152.tv2_in1k	2.34	0.43	1.88	0.24	0.02	0.04
resnet18.tv_in1k	0.27	0.07	1.48	0.00	0.00	0.04
resnet50.tv2_in1k	0.51	0.16	1.48	0.01	0.00	0.04
shufflenet_v2_x1_0.tv_in1k	1.13	0.14	1.55	1.15	0.02	0.08
squeezenet1_1.tv_in1k	1.09	0.05	1.44	0.37	0.02	0.03
swin_base_patch4_window7_224.ms_in22k_ft_in1k	2.71	0.94	2.28	0.41	0.03	0.09
vgg16.tv_in1k	3.22	1.74	1.90	0.30	0.00	0.06
vit_base_patch16_224.augreg2_in21k_ft_in1k	2.49	0.73	2.05	0.28	0.00	0.07
vit_so400m_patch14_siglip_378.webli_ft_in1k	7.52	7.29	5.82	0.35	0.07	0.09
deeplabv3-r101-d8-ade20k	1.64 (512x512)	0.73	0.91†	0.03 (512x512)	0.08	0.02†
deeplabv3plus-r101-d8-ade20k	4.72 (512x512)	0.62	0.75†	0.66 (512x512)	0.10	0.02†
dinov2-b-linear-ade20k	12.01 (512x512)	0.27	0.40†	0.02 (512x512)	0.10	0.07†
dinov2-l-linear-ade20k	7.86 (512x512)	0.81	0.92†	0.38 (512x512)	0.16	0.10†
dinov2-s-linear-ade20k	4.02 (512x512)	0.12	0.23†	0.30 (512x512)	0.09	0.06†
eomt-l-dinov2-ade20k	4.80 (512x512)	3.13	3.27†	0.54 (512x512)	0.05	0.04†
eomt-l-dinov3-ade20k	1.83 (512x512)	3.20	3.24†	0.07 (512x512)	0.08	0.04†
fcn-r101-d8-ade20k	4.72 (512x512)	0.53	0.91†	0.39 (512x512)	0.07	0.02†
hrnet-w48-fcn-ade20k	1.48 (512x512)	0.62	0.85†	0.08 (512x512)	0.07	0.01†
mask2former-swinl-ade20k	4.86 (512x512)	3.03	3.17†	0.76 (512x512)	0.13	0.07†
oneformer-swinl-ade20k	5.08 (512x512)	4.10	3.27†	0.81 (512x512)	0.15†	0.11†
pspnet-r101-d8-ade20k	1.35 (512x512)	0.47	0.77†	0.09 (512x512)	0.08	0.03†
sapiens-seg-0.3b-goliath	4.92 (512x512)	1.22	0.96†	1.02 (512x512)	0.13†	0.09†
sapiens-seg-1b-goliath	11.33 (512x512)	3.54	2.27†	0.57 (512x512)	0.27†	0.18†
segformer-b0-ade20k	4.83 (512x512)	2.55	2.92	2.23 (512x512)	0.08	0.01
segformer-b5-ade20k	1.24 (512x512)	3.09	3.43	0.50 (512x512)	0.08	0.03
sam-vit-b	0.75	0.46	0.17†	0.18	0.05†	0.04†
sam-vit-h	4.61	2.45	0.89†	0.38	0.21†	0.14†
sam-vit-l	1.80	1.21	0.43†	0.27	0.12†	0.08†
sam2.1-hiera-base-plus	0.74	0.43	0.19†	0.14	0.03†	0.03†
sam2.1-hiera-large	1.95	1.10	0.39†	0.18	0.07†	0.05†
sam2.1-hiera-small	0.48	0.36	0.13†	0.14	0.02†	0.02†
sam2.1-hiera-tiny	0.41	0.25	0.11†	0.15	0.02†	0.01†
sam3	1.12 (840x1297)	2.95	0.70†	0.96 (840x1297)	0.18†	0.12†
detr-r101-coco	0.95	0.81	0.22†	0.06	0.02	0.01
detr-r50-coco	2.55	0.70	0.28†	0.83	0.44	0.03
dfine-l-coco	3.68	1.09	1.46	1.41	0.67	0.47
dfine-m-coco	3.63	1.04	1.20	1.55	0.69	0.56
dfine-n-coco	5.24	1.25	1.21	4.07	0.85	0.62
dfine-s-coco	2.99	1.05	1.16	1.44	0.71	0.52
dfine-x-coco	3.88	1.39	1.41	1.45	0.65	0.45
faster-rcnn-r50-fpn-3x-coco	0.73	0.53	0.14†	0.04	0.25	0.04
mask-rcnn-r50-fpn-3x-coco	0.97	1.50	0.22†	0.84	0.42	0.04
rtdetrv2-l-coco	0.66	0.50	0.25†	0.04	0.01	0.02
rtdetrv2-m-coco	0.45	0.33	0.16†	0.02	0.01	0.01
rtdetrv2-ms-coco	0.52	0.45	0.17†	0.06	0.03	0.09
rtdetrv2-s-coco	1.32	0.31	0.46†	0.42	0.10	0.53
rtdetrv2-x-coco	1.13	1.09	0.29†	0.08	0.05	0.02
dinov2-vitb14	1.51	0.27	0.19	0.01	0.20	0.19
dinov2-vitb14-reg	1.71	0.59	0.21	0.04	0.04	0.20
dinov2-vitg14	19.29	4.37	2.22	0.05	0.10	0.16
dinov2-vitg14-reg	14.67	4.32	2.91†	0.09	0.03	0.02
dinov2-vitl14	7.06	0.57	0.57	0.01	0.15	0.18
dinov2-vitl14-reg	4.60	0.66	0.56	0.07	0.11	0.20
dinov2-vits14	0.41	0.07	0.07	0.17	0.12	0.15
dinov2-vits14-reg	1.32	0.21	0.08	0.11	0.13	0.17
dinov3-convnext-base	0.42	0.43	0.19	0.01	0.27	0.10
dinov3-convnext-large	1.64	0.85	0.38	0.02	0.24	0.20
dinov3-convnext-small	0.27	0.25	0.11	0.01	0.33	0.07
dinov3-convnext-tiny	0.43	0.48	0.08	1.03	0.42	0.12
dinov3-vitb16	0.68	0.38	0.16	0.01	0.20	0.29
dinov3-vith16plus	5.95	2.02	1.40	0.05	0.10	0.20
dinov3-vitl16	2.37	0.54	0.52	0.02	0.18	0.25
dinov3-vits16	0.48	0.07	0.06	0.95	0.27	0.26
mapanything-apache	6.12 (336x518)	2.89 (336x518)	0.57† (336x518)	2.62 (336x518)	0.66 (336x518)	0.54 (336x518)
mvsanywhere	3.21 (448x672)	1.88 (448x672)	0.52† (448x672)	1.04 (448x672)	0.22 (448x672)	0.05 (448x672)
patchmatchnet	0.04 (1536x1024)	0.06 (1536x1024)	0.05† (1536x1024)	0.92 (1536x1024)	0.25 (1536x1024)	0.04 (1536x1024)
vggt-1b	19.37 (336x518)	4.98 (336x518)	5.99† (336x518)	1.29 (336x518)	0.69 (336x518)	0.46 (336x518)
vggt-omega-1b-512	16.31 (336x518)	3.31 (336x518)	2.27† (336x518)	1.10 (336x518)	0.73 (336x518)	0.46 (336x518)

† measured in the speed sweep after the portable load (or an earlier size) in the same process: file cache and kernels are warm, so the time is lower than a fresh start.

torch.compile and TensorRT

Portable with model.compile() (Inductor, default mode) and the fp16 TensorRT engine of the portable model (export_tensorrt), one variant per family at the task's standard size, next to portable and kernels (export has the methods). The compile time is the cold first call (empty Inductor cache); the engine runs module() on synthetic inputs of the shape the timed call gives the network (classifier crop, all views of a scene; "only" marks engines of part of the call) and its memory is the device memory it holds. Rows measured by tools/speed_check.py run --modes compile tensorrt; a failure is shown as fails: reason.

model	input	time: portable	compile	TensorRT fp16	kernels	preparation: compile (cold first call)	ONNX + engine build	peak MiB (alloc / reserved): portable	compile	kernels	TensorRT (device)
aanet-sceneflow	994x1482	140 ms	91.8 ms	91.2 ms	101 ms	79.9 s	211 s	845 / 986	754 / 928	536 / 644	482
bridgedepth-sf	994x1482	437 ms	325 ms	fails: TensorRT bug: Myelin internal error	348 ms	111 s	–	2916 / 3214	2280 / 2308	2310 / 2634	–
c-fast-foundationstereo	994x1482	330 ms	339 ms	451 ms	135 ms	75.4 s	240 s	2109 / 2336	1649 / 1792	1196 / 1516	1252
cas-gwcnet-c-sceneflow	994x1482	327 ms	295 ms	220 ms	225 ms	61.7 s	131 s	2006 / 2474	2098 / 2448	1725 / 1930	1440
cfnet-sceneflow	994x1482	484 ms	249 ms	171 ms	401 ms	53.0 s	139 s	3441 / 4376	1026 / 1398	3662 / 4718	1238
coex-sceneflow	994x1482	49.9 ms	29.4 ms	20.6 ms	23.5 ms	22.4 s	56.6 s	1136 / 1254	351 / 452	495 / 886	360
crestereo-eth3d	994x1482	732 ms	398 ms	298 ms	229 ms	120 s	224 s	884 / 1030	695 / 748	644 / 732	584
defom-stereo-vitl-sceneflow	994x1482	1531 ms	1002 ms	fails: TensorRT bug: Myelin internal error	659 ms	122 s	–	4362 / 4838	3021 / 3292	3154 / 3566	–
dlnr-sceneflow	994x1482	398 ms	301 ms	241 ms	214 ms	69.0 s	161 s	2635 / 3142	1343 / 1672	1236 / 1606	1492
fast-acvnet-sceneflow	994x1482	54.7 ms	36.2 ms	36.7 ms	38.2 ms	41.8 s	71.3 s	664 / 736	584 / 698	739 / 1042	484
foundationstereo-vit-s	994x1482	2514 ms	2044 ms	1196 ms	1120 ms	212 s	545 s	2493 / 2606	2822 / 3646	2169 / 2470	2570
great-igev-eth3d	994x1482	1085 ms	829 ms	940 ms	393 ms	224 s	685 s	1402 / 1678	2831 / 3362	1192 / 1564	6018
gwcnet-g-sceneflow	994x1482	344 ms	417 ms	275 ms	238 ms	52.6 s	66.0 s	2081 / 2402	1527 / 1864	1799 / 2102	3548
hitnet-eth3d	994x1482	67.8 ms	27.7 ms	31.1 ms	23.2 ms	39.1 s	136 s	860 / 1090	719 / 728	647 / 1248	836
igev-plusplus-eth3d	994x1482	1081 ms	716 ms	494 ms	393 ms	159 s	425 s	1594 / 1840	952 / 1074	1389 / 2302	3632
igev-stereo-eth3d	994x1482	763 ms	527 ms	339 ms	288 ms	111 s	213 s	1237 / 1492	807 / 1052	1113 / 1374	1220
las2-s	994x1482	46.9 ms	23.7 ms	10.3 ms	14.3 ms	22.4 s	86.9 s	955 / 988	238 / 264	320 / 558	162
mocha-v2-sceneflow	994x1482	1033 ms	713 ms	437 ms	448 ms	148 s	361 s	1632 / 1770	1122 / 1308	1587 / 1790	1098
monster-plusplus-mix-all	994x1482	1633 ms	1126 ms	697 ms	698 ms	109 s	416 s	2583 / 2918	2328 / 2532	2438 / 2760	2642
monster-sceneflow	994x1482	1585 ms	1096 ms	671 ms	668 ms	103 s	384 s	2336 / 2494	2081 / 2272	2356 / 2714	2018
nmrf-sceneflow	994x1482	216 ms	161 ms	184 ms	196 ms	84.8 s	142 s	1510 / 1552	854 / 1104	1022 / 1112	826
psmnet-sceneflow	994x1482	340 ms	440 ms	669 ms	232 ms	60.0 s	88.9 s	2078 / 2442	1760 / 2082	1786 / 2122	1528
raft-stereo-eth3d	994x1482	746 ms	506 ms	333 ms	273 ms	99.9 s	132 s	1388 / 1492	815 / 988	765 / 802	826
rt-monster-plusplus-mix-all	994x1482	113 ms	77.0 ms	44.1 ms	94.1 ms	69.4 s	150 s	2592 / 2646	2750 / 2942	2704 / 3976	952
s2m2-s	994x1482	348 ms	186 ms	137 ms	113 ms	95.9 s	363 s	1733 / 1930	1579 / 1710	686 / 820	980
selective-igev-eth3d-finetune	994x1482	998 ms	691 ms	461 ms	349 ms	126 s	256 s	1238 / 1488	919 / 1088	1109 / 1390	1148
waft-stereo-s	994x1482	689 ms	357 ms	170 ms	284 ms	95.1 s	152 s	3293 / 3456	2024 / 2058	1113 / 1286	662
da3mono-large	994x1482	23.9 ms	21.2 ms	14.4 ms	14.6 ms	25.0 s	52.3 s	818 / 1328	792 / 828	814 / 868	754
depth-anything-v2-s	994x1482	7.82 ms	6.65 ms	4.46 ms	4.51 ms	15.7 s	29.9 s	249 / 326	201 / 224	190 / 248	186
depth-pro	994x1482	518 ms	488 ms	299 ms	271 ms	39.2 s	163 s	3306 / 3498	3053 / 3198	3667 / 3778	2926
marigold-depth-v1-1	994x1482	408 ms	342 ms	172 ms	383 ms	127 s	285 s	2817 / 3760	2341 / 2418	2583 / 3762	2368
metric3d-vit-large	994x1482	200 ms	149 ms	91.4 ms	158 ms	54.7 s	96.6 s	1484 / 1696	1126 / 1148	1516 / 1688	1010
midas-dpt-beit-large-512	994x1482	78.8 ms	58.8 ms	52.2 ms	63.9 ms	32.6 s	53.5 s	992 / 1418	1088 / 1116	1003 / 1436	800
moge-2-vits-normal	994x1482	64.2 ms	43.9 ms	24.5 ms	29.7 ms	124 s	131 s	1066 / 1336	836 / 992	884 / 1008	620
unidepth-v2-vitl14	994x1482	99.8 ms	96.4 ms	64.3 ms	66.4 ms	31.1 s	119 s	1208 / 1472	1304 / 1392	1211 / 1472	1050
unik3d-vitl	994x1482	108 ms	94.4 ms	58.0 ms	62.9 ms	32.8 s	105 s	1219 / 1492	1258 / 1312	1166 / 1266	846
zoedepth-nk	994x1482	124 ms	71.7 ms	73.3 ms	88.9 ms	93.8 s	138 s	2037 / 2124	1813 / 1880	1771 / 1960	852
alexnet.tv_in1k	224x224	0.970 ms	0.824 ms	0.401 ms	0.604 ms	4.33 s	13.4 s	151 / 224	183 / 204	184 / 228	128
convnextv2_base.fcmae_ft_in22k_in1k	224x224	10.2 ms	3.85 ms	2.13 ms	2.94 ms	14.4 s	35.1 s	299 / 376	324 / 356	318 / 380	204
deit_base_patch16_224.fb_in1k	224x224	2.98 ms	2.09 ms	1.12 ms	1.64 ms	7.27 s	28.3 s	203 / 386	238 / 266	300 / 370	196
densenet121.tv_in1k	224x224	9.41 ms	5.49 ms	0.686 ms	1.28 ms	20.4 s	33.6 s	53 / 90	89 / 112	83 / 132	36
dinov2_vitl14_lc	224x224	5.43 ms	5.09 ms	4.22 ms	4.93 ms	12.1 s	38.2 s	623 / 1228	660 / 688	890 / 950	618
efficientnet_b0.tv_in1k	224x224	4.98 ms	2.88 ms	0.536 ms	0.769 ms	9.12 s	25.8 s	48 / 96	80 / 96	78 / 118	38
efficientnet_v2_s.tv_in1k	224x224	10.2 ms	5.39 ms	1.35 ms	1.89 ms	12.1 s	32.0 s	81 / 106	116 / 146	114 / 188	74
efficientvit_b2.r288_in1k	224x224	11.9 ms	6.91 ms	0.790 ms	1.10 ms	11.0 s	35.3 s	85 / 124	120 / 164	120 / 212	74
eva02_large_patch14_448.mim_m38m_ft_in22k_in1k	224x224	18.0 ms	14.9 ms	12.4 ms	14.1 ms	21.0 s	43.5 s	648 / 1206	674 / 684	1236 / 1386	642
fastvit_sa24.apple_dist_in1k	224x224	6.88 ms	3.87 ms	0.808 ms	1.39 ms	9.39 s	44.9 s	81 / 108	113 / 148	112 / 190	72
inception_v3.tv_in1k	224x224	5.21 ms	3.03 ms	0.685 ms	1.30 ms	8.61 s	22.4 s	120 / 166	152 / 184	150 / 228	70
mobilenet_v2.tv2_in1k	224x224	2.95 ms	1.82 ms	0.261 ms	0.439 ms	6.56 s	16.3 s	43 / 70	77 / 92	75 / 92	36
mobilenet_v3_large.tv2_in1k	224x224	3.80 ms	2.19 ms	0.393 ms	0.545 ms	10.2 s	36.5 s	46 / 76	79 / 90	78 / 114	36
mobilenetv1_100.ra4_e3600_r224_in1k	224x224	1.77 ms	1.21 ms	0.176 ms	0.370 ms	4.59 s	14.1 s	44 / 72	77 / 90	76 / 114	36
mobilenetv4_conv_medium.e500_r256_in1k	224x224	3.91 ms	2.48 ms	0.345 ms	0.641 ms	6.98 s	19.6 s	54 / 86	87 / 102	87 / 124	36
regnety_160.tv2_in1k	224x224	8.34 ms	5.42 ms	1.78 ms	1.86 ms	8.01 s	28.1 s	208 / 250	236 / 290	229 / 316	172
resnet18.tv_in1k	224x224	1.48 ms	1.04 ms	0.220 ms	0.459 ms	4.31 s	14.0 s	61 / 90	91 / 108	90 / 150	34
resnet50.tv2_in1k	224x224	3.15 ms	2.10 ms	0.466 ms	0.826 ms	8.31 s	24.0 s	89 / 142	119 / 138	116 / 162	70
shufflenet_v2_x1_0.tv_in1k	224x224	3.71 ms	1.98 ms	0.329 ms	0.787 ms	7.73 s	16.9 s	41 / 72	71 / 88	87 / 116	32
squeezenet1_1.tv_in1k	224x224	1.82 ms	1.13 ms	0.214 ms	0.359 ms	4.49 s	9.10 s	41 / 72	39 / 48	8 / 12	34
swin_base_patch4_window7_224.ms_in22k_ft_in1k	224x224	10.8 ms	4.19 ms	1.48 ms	2.15 ms	26.9 s	64.2 s	213 / 394	243 / 270	246 / 300	202
vgg16.tv_in1k	224x224	1.99 ms	1.86 ms	1.17 ms	1.50 ms	4.08 s	15.9 s	313 / 372	342 / 368	332 / 388	302
vit_base_patch16_224.augreg2_in21k_ft_in1k	224x224	3.02 ms	2.10 ms	1.14 ms	1.65 ms	7.26 s	28.6 s	205 / 390	238 / 266	300 / 370	196
vit_so400m_patch14_siglip_378.webli_ft_in1k	224x224	13.1 ms	13.3 ms	11.4 ms	12.8 ms	14.2 s	60.7 s	879 / 1708	909 / 944	1216 / 1346	848
dinov2-s-linear-ade20k	1024x2048	18.3 ms	20.7 ms	40.3 ms	12.4 ms	10.6 s	36.2 s	642 / 696	1331 / 1376	645 / 676	1794
eomt-l-dinov2-ade20k	1024x2048	50.8 ms	49.4 ms	54.3 ms	35.0 ms	16.5 s	52.9 s	1515 / 1584	1585 / 1624	1515 / 1584	1798
hrnet-w48-fcn-ade20k	1024x2048	26.9 ms	23.9 ms	20.9 ms	11.9 ms	35.7 s	29.6 s	424 / 490	421 / 448	274 / 430	888
mask2former-swinl-ade20k	1024x2048	103 ms	fails: Inductor: Triton kernel does not compile (torch bug)	55.6 ms	60.4 ms	–	129 s	1164 / 1400	–	1164 / 1400	1816
oneformer-swinl-ade20k	1024x2048	139 ms	fails: Inductor: Triton kernel does not compile (torch bug)	83.3 ms	100.0 ms	–	140 s	2054 / 2476	–	2054 / 2476	3258
sapiens-seg-0.3b-goliath	1024x2048	132 ms	118 ms	156 ms	92.6 ms	12.5 s	52.9 s	1643 / 1762	1373 / 1462	1643 / 1762	1564
segformer-b0-ade20k	1024x2048	11.4 ms	9.27 ms	5.51 ms	3.48 ms	12.3 s	60.9 s	339 / 376	336 / 356	148 / 238	792
sam-vit-b	1024x1024	48.3 ms	40.9 ms	27.3 ms (image encoder only)	37.0 ms	14.5 s	36.4 s	378 / 420	362 / 400	377 / 420	322
sam2.1-hiera-tiny	1024x1024	16.6 ms	14.2 ms	8.56 ms (image encoder only)	13.9 ms	13.4 s	34.1 s	304 / 382	270 / 322	316 / 362	266
sam3	1024x1024	179 ms	168 ms	89.5 ms (image encoder only)	116 ms	20.0 s	54.2 s	1922 / 2004	1910 / 2004	2019 / 2264	1122
detr-r50-coco	800x1344	12.4 ms	9.27 ms	4.40 ms	9.93 ms	21.7 s	33.4 s	280 / 332	233 / 292	218 / 272	182
dfine-n-coco	800x1344	10.3 ms	5.43 ms	0.942 ms	2.72 ms	26.7 s	51.5 s	88 / 116	101 / 118	124 / 202	56
faster-rcnn-r50-fpn-3x-coco	800x1344	22.8 ms	18.9 ms	5.35 ms (backbone + FPN only)	19.5 ms	26.0 s	15.6 s	390 / 490	413 / 438	391 / 450	258
rtdetrv2-s-coco	800x1344	8.04 ms	4.22 ms	1.81 ms	4.19 ms	37.0 s	42.2 s	134 / 166	152 / 182	155 / 250	98
dinov2-vitb14	224x224	2.92 ms	2.20 ms	1.30 ms	1.48 ms	9.52 s	25.2 s	205 / 226	240 / 266	237 / 290	198
dinov3-convnext-tiny	224x224	3.04 ms	1.96 ms	0.681 ms	1.18 ms	7.76 s	22.8 s	93 / 108	123 / 148	86 / 112	72
dinov3-vits16	224x224	4.67 ms	2.18 ms	0.564 ms	0.903 ms	8.81 s	25.0 s	77 / 80	109 / 138	77 / 106	64
mapanything-apache	336x518	636 ms	596 ms	367 ms	544 ms	46.2 s	137 s	3697 / 4092	3347 / 3792	3697 / 4072	3880
mvsanywhere	448x672	93.2 ms	70.8 ms	fails: TensorRT: NaN output (onnxruntime on the same ONNX is right)	35.6 ms	153 s	–	931 / 1226	818 / 860	761 / 884	–
patchmatchnet	1536x1024	202 ms	175 ms	51.0 ms	29.6 ms	29.4 s	82.1 s	1806 / 2096	1798 / 1888	740 / 856	844
vggt-1b	336x518	651 ms	557 ms	424 ms	458 ms	74.3 s	199 s	3691 / 4674	3611 / 3856	3702 / 4654	3642
vggt-omega-1b-512	336x518	726 ms	589 ms	377 ms	464 ms	65.7 s	187 s	3421 / 3518	3082 / 3154	3422 / 3518	2994

Original code

The original columns are the authors' code with the same checkpoint and settings, in fp16 where the code supports it and the output stays finite, else fp32 (marked (fp32)). Original fp32 is the same code at its default fp32 precision, where that differs. The kernels speed-up is against the faster of original fp16 and fp32. Notes per row are on the model pages. A dash means not measured or no runnable original. Where kernels was not timed at the original's size, the overview shows the speed-up against portable instead.

Fixed-size speed and peak memory, all models

Every model, portable and kernels, on a synthetic input of the size in the column header, median of 10 warm runs (3 for calls over 2 s). Cells: median ms / peak allocated MiB / peak reserved MiB, weights included. OOM = out of memory on the 16 GB GPU. Fixed-input models (classification, DINO features) run only at their native size: see their speed table above. The 12 MP column probes the largest size that fits. Depth-anything-v2, midas-dpt-beit-large-512 and zoedepth-nk resize to their working resolution: time depends on the aspect ratio, not the input pixels (12 MP is 4:3, so it is faster than 1080p at 16:9). Measured 2026-10/11 on main at the time; the fixed-input table and the GREAT-IGEV and IGEV++ rows are from 2026-11-03, and the other columns have not been re-measured since.

Cells: median ms / peak allocated MiB / peak reserved MiB (incl. weights). OOM = out of memory; SKIPPED = a smaller size already failed; TIMEOUT = the call did not finish in a 170 s GPU-lock step; - = not run (the size is outside the model's range).

model	mode	480x640	720x1280	1080x1920	12 MP 3000x4000
aanet-plus-sceneflow	portable	58.4 ms / 1109 / 1292	170.2 ms / 3400 / 3858	389.0 ms / 7214 / 8226	SKIPPED
aanet-plus-sceneflow	kernels	28.8 ms / 178 / 244	53.9 ms / 501 / 696	128.5 ms / 1049 / 1430	793.2 ms / 5752 / 8230
aanet-sceneflow	portable	40.1 ms / 300 / 340	94.2 ms / 781 / 830	231.2 ms / 1711 / 1820	1591.3 ms / 9603 / 12836
aanet-sceneflow	kernels	25.9 ms / 148 / 196	59.2 ms / 344 / 602	152.4 ms / 728 / 1252	967.9 ms / 3967 / 6854
bridgedepth-rvc-pretrain	portable	76.7 ms / 1727 / 1798	247.2 ms / 2359 / 2628	676.5 ms / 3562 / 4102	7918.7 ms (n=3) / 13620 / 14590
bridgedepth-rvc-pretrain	kernels	62.2 ms / 1147 / 1266	200.9 ms / 1764 / 2096	534.7 ms / 2939 / 3514	5573.1 ms (n=3) / 12950 / 13592
bridgedepth-sf	portable	73.2 ms / 1676 / 1748	242.8 ms / 2308 / 2560	668.4 ms / 3511 / 4032	7849.6 ms (n=3) / 13569 / 14540
bridgedepth-sf	kernels	59.1 ms / 1096 / 1214	196.3 ms / 1714 / 2044	526.9 ms / 2888 / 3462	5498.6 ms (n=3) / 12898 / 13540
c-fast-foundationstereo	portable	64.6 ms / 600 / 784	193.1 ms / 1626 / 2144	464.9 ms / 3485 / 4540	OOM
c-fast-foundationstereo	kernels	38.5 ms / 306 / 386	85.9 ms / 762 / 960	182.0 ms / 1582 / 2060	1072.8 ms / 8699 / 11372
c-foundationstereo-s	portable	681.8 ms / 2075 / 3130	2218.0 ms (n=3) / 4355 / 6278	4704.5 ms (n=3) / 8457 / 12000	SKIPPED
c-foundationstereo-s	kernels	291.3 ms / 1262 / 1430	838.6 ms / 2380 / 3312	1826.4 ms / 4137 / 6164	SKIPPED
fast-foundationstereo	portable	64.6 ms / 600 / 784	193.1 ms / 1626 / 2144	464.8 ms / 3485 / 4540	OOM
fast-foundationstereo	kernels	38.7 ms / 306 / 386	85.8 ms / 762 / 960	181.9 ms / 1582 / 2060	1073.0 ms / 8699 / 11372
cas-gwcnet-c-sceneflow	portable	61.3 ms / 465 / 738	195.5 ms / 1386 / 2160	452.8 ms / 3046 / 4720	OOM
cas-gwcnet-c-sceneflow	kernels	44.8 ms / 442 / 568	135.5 ms / 1242 / 1586	316.1 ms / 2692 / 3408	OOM
cas-gwcnet-gc-sceneflow	portable	86.7 ms / 533 / 1166	281.6 ms / 1588 / 3308	647.4 ms / 3494 / 7262	OOM
cas-gwcnet-gc-sceneflow	kernels	66.9 ms / 510 / 956	213.7 ms / 1449 / 2606	493.8 ms / 3147 / 5634	OOM
cfnet-sceneflow	portable	73.6 ms / 1070 / 1588	284.0 ms / 2640 / 4342	758.8 ms / 4508 / 5956	OOM
cfnet-sceneflow	kernels	59.3 ms / 1118 / 1618	229.9 ms / 2778 / 4362	634.1 ms / 4801 / 5976	OOM
coex-sceneflow	portable	11.2 ms / 225 / 312	28.5 ms / 671 / 890	70.0 ms / 1485 / 1916	554.7 ms / 8660 / 10964
coex-sceneflow	kernels	4.3 ms / 103 / 322	13.7 ms / 256 / 400	32.9 ms / 559 / 804	211.1 ms / 3190 / 4410
crestereo-eth3d	portable	142.5 ms / 214 / 284	427.8 ms / 556 / 740	1022.9 ms / 1172 / 1568	6466.7 ms (n=3) / 6508 / 8990
crestereo-eth3d	kernels	61.8 ms / 216 / 292	145.2 ms / 552 / 792	311.2 ms / 1162 / 1638	1767.7 ms / 6427 / 9362
defom-stereo-vitl-sceneflow	portable	263.9 ms / 2070 / 2368	860.4 ms / 3248 / 4044	2169.2 ms (n=3) / 5378 / 7082	OOM
defom-stereo-vitl-sceneflow	kernels	120.5 ms / 1357 / 1702	376.5 ms / 2275 / 3190	936.5 ms / 2991 / 3578	5950.0 ms (n=3) / 11923 / 15526
dlnr-middlebury	portable	81.8 ms / 623 / 774	235.4 ms / 1651 / 2134	548.9 ms / 3617 / 4726	OOM
dlnr-middlebury	kernels	52.8 ms / 372 / 470	126.7 ms / 815 / 1070	297.2 ms / 1617 / 1958	1775.1 ms / 8572 / 12080
dlnr-sceneflow	portable	81.8 ms / 623 / 774	235.2 ms / 1651 / 2134	549.1 ms / 3617 / 4726	OOM
dlnr-sceneflow	kernels	52.9 ms / 372 / 470	126.8 ms / 815 / 1070	297.4 ms / 1617 / 1958	1775.1 ms / 8572 / 12080
fast-acvnet-generalization	portable	13.9 ms / 137 / 208	31.1 ms / 402 / 664	78.3 ms / 890 / 1418	492.1 ms / 5077 / 8074
fast-acvnet-generalization	kernels	6.8 ms / 169 / 582	22.0 ms / 435 / 608	55.7 ms / 954 / 1372	357.3 ms / 5445 / 7618
fast-acvnet-kitti12	portable	14.0 ms / 137 / 208	31.1 ms / 402 / 664	78.2 ms / 890 / 1418	492.4 ms / 5077 / 8074
fast-acvnet-kitti12	kernels	6.8 ms / 169 / 582	22.0 ms / 435 / 608	55.6 ms / 954 / 1372	357.3 ms / 5445 / 7618
fast-acvnet-kitti15	portable	13.9 ms / 137 / 208	31.1 ms / 402 / 664	78.2 ms / 890 / 1418	492.4 ms / 5077 / 8074
fast-acvnet-kitti15	kernels	6.8 ms / 169 / 582	22.0 ms / 435 / 608	55.6 ms / 954 / 1372	357.3 ms / 5445 / 7618
fast-acvnet-plus-generalization	portable	14.3 ms / 136 / 208	38.0 ms / 401 / 664	99.4 ms / 887 / 1418	623.6 ms / 5054 / 8076
fast-acvnet-plus-generalization	kernels	7.8 ms / 169 / 546	25.3 ms / 457 / 598	66.2 ms / 1006 / 1226	422.5 ms / 5743 / 7342
fast-acvnet-plus-kitti12	portable	14.4 ms / 136 / 208	38.0 ms / 401 / 664	99.4 ms / 887 / 1418	623.3 ms / 5054 / 8076
fast-acvnet-plus-kitti12	kernels	7.7 ms / 169 / 546	25.3 ms / 457 / 598	66.2 ms / 1006 / 1226	422.4 ms / 5743 / 7342
foundationstereo-vit-l	portable	577.7 ms / 1654 / 2216	1846.4 ms / 2676 / 3370	3776.7 ms (n=3) / 3570 / 5048	OOM
foundationstereo-vit-l	kernels	320.0 ms / 1296 / 1642	961.0 ms / 2232 / 2800	1788.5 ms / 3254 / 4506	9521.9 ms (n=3) / 12653 / 14398
foundationstereo-vit-s	portable	481.4 ms / 1042 / 1352	1508.2 ms / 2047 / 2678	3424.1 ms (n=3) / 2900 / 4592	OOM
foundationstereo-vit-s	kernels	245.6 ms / 685 / 862	687.1 ms / 1604 / 2008	1499.8 ms / 2585 / 3440	8960.9 ms (n=3) / 12058 / 13800
gwcnet-g-sceneflow	portable	69.2 ms / 418 / 586	219.1 ms / 1254 / 1776	502.0 ms / 2764 / 3874	2911.8 ms (n=3) / 13647 / 15104
gwcnet-g-sceneflow	kernels	50.9 ms / 393 / 526	156.0 ms / 1112 / 1590	358.5 ms / 2412 / 3464	OOM
gwcnet-gc-sceneflow	portable	71.2 ms / 418 / 442	224.5 ms / 1255 / 1688	514.1 ms / 2766 / 3680	2981.7 ms (n=3) / 13924 / 14002
gwcnet-gc-sceneflow	kernels	54.9 ms / 394 / 468	167.5 ms / 1114 / 1534	385.9 ms / 2415 / 3300	OOM
hitnet-eth3d	portable	18.1 ms / 221 / 336	40.2 ms / 659 / 1004	92.6 ms / 1134 / 2438	579.4 ms / 6558 / 11110
hitnet-eth3d	kernels	5.2 ms / 140 / 470	14.5 ms / 392 / 482	30.8 ms / 836 / 1004	181.7 ms / 4814 / 5758
hitnet-middlebury-d400	portable	35.8 ms / 602 / 838	108.8 ms / 1067 / 2222	251.7 ms / 2392 / 3814	OOM
hitnet-middlebury-d400	kernels	11.7 ms / 266 / 966	33.1 ms / 759 / 978	74.2 ms / 1695 / 2248	426.2 ms / 9604 / 12664
igev-stereo-eth3d	portable	148.4 ms / 280 / 352	442.8 ms / 768 / 1034	1089.2 ms / 1689 / 2272	6940.4 ms (n=3) / 10547 / 14300
igev-stereo-eth3d	kernels	70.2 ms / 263 / 372	177.6 ms / 698 / 1000	391.1 ms / 1506 / 2044	2328.3 ms (n=3) / 6315 / 8190
igev-stereo-kitti12	portable	148.6 ms / 280 / 352	442.9 ms / 768 / 1034	1089.5 ms / 1689 / 2272	6944.6 ms (n=3) / 10547 / 14300
igev-stereo-kitti12	kernels	70.1 ms / 263 / 372	177.7 ms / 698 / 1000	391.5 ms / 1506 / 2044	2330.4 ms (n=3) / 6315 / 8190
igev-stereo-kitti15	portable	148.3 ms / 280 / 352	442.9 ms / 768 / 1034	1089.7 ms / 1689 / 2272	6944.1 ms (n=3) / 10547 / 14300
igev-stereo-kitti15	kernels	70.3 ms / 263 / 372	177.8 ms / 698 / 1000	391.3 ms / 1506 / 2044	2330.4 ms (n=3) / 6315 / 8190
igev-stereo-middlebury	portable	148.4 ms / 280 / 352	442.4 ms / 768 / 1034	1089.3 ms / 1689 / 2272	6939.3 ms (n=3) / 10547 / 14300
igev-stereo-middlebury	kernels	70.2 ms / 263 / 372	177.9 ms / 698 / 1000	391.5 ms / 1506 / 2044	2328.7 ms (n=3) / 6315 / 8190
igev-stereo-sceneflow	portable	148.6 ms / 280 / 352	442.9 ms / 768 / 1034	1089.8 ms / 1689 / 2272	6943.4 ms (n=3) / 10547 / 14300
igev-stereo-sceneflow	kernels	70.2 ms / 263 / 372	178.2 ms / 698 / 1000	392.0 ms / 1506 / 2044	2334.4 ms (n=3) / 6315 / 8190
las2-h	portable	29.8 ms / 269 / 328	69.1 ms / 668 / 904	174.8 ms / 1386 / 2168	1408.7 ms / 8708 / 12018
las2-h	kernels	10.2 ms / 218 / 596	28.5 ms / 604 / 1636	73.6 ms / 1344 / 1648	512.5 ms / 9695 / 12606
las2-l	portable	25.5 ms / 286 / 328	51.1 ms / 653 / 700	110.1 ms / 1317 / 1404	824.0 ms / 7201 / 7598
las2-l	kernels	6.6 ms / 129 / 270	16.8 ms / 241 / 676	38.6 ms / 422 / 588	254.6 ms / 2174 / 2896
las2-m	portable	18.9 ms / 219 / 252	35.4 ms / 586 / 648	73.9 ms / 1249 / 1358	602.8 ms / 7135 / 7584
las2-m	kernels	4.0 ms / 96 / 258	9.2 ms / 187 / 618	21.5 ms / 346 / 516	146.5 ms / 1870 / 2684
las2-s	portable	16.1 ms / 202 / 234	31.0 ms / 569 / 632	64.4 ms / 1232 / 1342	532.1 ms / 7119 / 7566
las2-s	kernels	3.0 ms / 88 / 246	6.9 ms / 179 / 602	16.6 ms / 338 / 500	109.3 ms / 1859 / 2598
mocha-v1-sceneflow	portable	231.1 ms / 500 / 662	628.3 ms / 1078 / 1438	1484.8 ms / 2164 / 2688	10361.4 ms (n=3) / 12466 / 15520
mocha-v1-sceneflow	kernels	135.4 ms / 495 / 694	317.0 ms / 1055 / 1338	718.5 ms / 2092 / 2484	4396.8 ms (n=3) / 11664 / 14958
mocha-v2-sceneflow	portable	227.0 ms / 500 / 662	613.9 ms / 1078 / 1438	1451.9 ms / 2164 / 2688	10135.5 ms (n=3) / 12464 / 15164
mocha-v2-sceneflow	kernels	133.6 ms / 494 / 654	312.9 ms / 1054 / 1456	709.3 ms / 2091 / 2482	4316.6 ms (n=3) / 11664 / 14404
monster-mix-all	portable	284.1 ms / 1089 / 1184	894.5 ms / 1732 / 2154	2291.3 ms (n=3) / 2935 / 3668	OOM
monster-mix-all	kernels	127.2 ms / 1112 / 1250	379.5 ms / 1752 / 2288	956.8 ms / 2939 / 3460	OOM
monster-sceneflow	portable	284.1 ms / 1089 / 1184	894.5 ms / 1732 / 2154	2290.9 ms (n=3) / 2935 / 3668	OOM
monster-sceneflow	kernels	127.1 ms / 1112 / 1250	379.4 ms / 1752 / 2288	956.9 ms / 2939 / 3460	OOM
monster-plusplus-mix-all	portable	294.3 ms / 1138 / 1246	928.8 ms / 1880 / 2318	2370.0 ms (n=3) / 3269 / 3812	OOM
monster-plusplus-mix-all	kernels	132.9 ms / 1130 / 1270	396.3 ms / 1804 / 2344	998.1 ms / 3051 / 3724	OOM
nmrf-sceneflow	portable	35.9 ms / 390 / 454	124.3 ms / 953 / 1094	315.1 ms / 2040 / 2402	2021.2 ms (n=3) / 11263 / 13666
nmrf-sceneflow	kernels	31.7 ms / 247 / 298	112.1 ms / 626 / 800	286.8 ms / 1356 / 1814	1834.3 ms / 7543 / 10866
psmnet-sceneflow	portable	69.0 ms / 426 / 606	206.2 ms / 1256 / 1780	479.2 ms / 2834 / 3966	OOM
psmnet-sceneflow	kernels	49.6 ms / 415 / 518	140.3 ms / 1108 / 1414	329.2 ms / 2459 / 3104	OOM
raft-stereo-eth3d	portable	143.6 ms / 320 / 376	433.7 ms / 864 / 1058	1036.0 ms / 1852 / 2258	6289.1 ms (n=3) / 10406 / 12814
raft-stereo-eth3d	kernels	57.8 ms / 201 / 214	166.5 ms / 569 / 590	370.4 ms / 1233 / 1264	2138.5 ms (n=3) / 7295 / 7436
raft-stereo-middlebury	portable	143.7 ms / 320 / 376	433.7 ms / 864 / 1058	1035.8 ms / 1852 / 2258	6287.7 ms (n=3) / 10406 / 12814
raft-stereo-middlebury	kernels	57.8 ms / 201 / 214	165.9 ms / 569 / 590	370.5 ms / 1233 / 1264	2139.9 ms (n=3) / 7295 / 7436
raft-stereo-sceneflow	portable	143.7 ms / 320 / 376	433.6 ms / 864 / 1058	1035.8 ms / 1852 / 2258	6286.9 ms (n=3) / 10406 / 12814
raft-stereo-sceneflow	kernels	57.6 ms / 201 / 214	166.0 ms / 569 / 590	370.4 ms / 1233 / 1264	2138.4 ms (n=3) / 7295 / 7436
rt-monster-plusplus-mix-all	portable	31.2 ms / 609 / 720	67.1 ms / 1625 / 1968	168.0 ms / 3478 / 4148	OOM
rt-monster-plusplus-mix-all	kernels	18.3 ms / 630 / 1934	57.5 ms / 1660 / 1998	140.1 ms / 3545 / 4188	OOM
s2m2-l	portable	182.7 ms / 1304 / 1460	618.6 ms / 2499 / 2956	1459.4 ms / 4660 / 5680	OOM
s2m2-l	kernels	83.3 ms / 719 / 782	228.8 ms / 988 / 1224	534.8 ms / 1530 / 2250	4189.4 ms (n=3) / 6355 / 9246
s2m2-m	portable	114.5 ms / 780 / 904	402.5 ms / 1690 / 1954	973.5 ms / 3330 / 3896	OOM
s2m2-m	kernels	55.3 ms / 395 / 492	153.2 ms / 655 / 940	355.5 ms / 1119 / 1760	2793.3 ms (n=3) / 5151 / 7876
s2m2-s	portable	56.0 ms / 456 / 536	202.3 ms / 1141 / 1450	511.1 ms / 2427 / 3054	OOM
s2m2-s	kernels	32.7 ms / 212 / 294	70.2 ms / 452 / 692	161.9 ms / 887 / 1534	1203.3 ms / 4637 / 7166
s2m2-xl	portable	338.0 ms / 2476 / 2710	1112.0 ms / 4248 / 4842	2594.0 ms (n=3) / 7447 / 8684	SKIPPED
s2m2-xl	kernels	140.8 ms / 1441 / 1644	414.8 ms / 1748 / 2040	956.5 ms / 2493 / 3446	6933.0 ms (n=3) / 9200 / 11578
selective-igev-eth3d-finetune	portable	190.5 ms / 282 / 362	575.3 ms / 769 / 1036	1420.2 ms / 1690 / 2030	8918.7 ms (n=3) / 10549 / 14320
selective-igev-eth3d-finetune	kernels	82.5 ms / 260 / 368	212.8 ms / 693 / 1010	472.5 ms / 1502 / 2046	2806.7 ms (n=3) / 6928 / 8800
selective-igev-eth3d-train	portable	190.5 ms / 282 / 362	575.3 ms / 769 / 1036	1420.4 ms / 1690 / 2030	8919.8 ms (n=3) / 10549 / 14320
selective-igev-eth3d-train	kernels	82.4 ms / 260 / 368	212.8 ms / 693 / 1010	472.6 ms / 1502 / 2046	2807.0 ms (n=3) / 6928 / 8800
selective-igev-kitti12	portable	190.6 ms / 282 / 362	575.3 ms / 769 / 1036	1420.5 ms / 1690 / 2030	8917.7 ms (n=3) / 10549 / 14320
selective-igev-kitti12	kernels	82.3 ms / 260 / 368	212.9 ms / 693 / 1010	471.9 ms / 1502 / 2046	2803.4 ms (n=3) / 6928 / 8800
selective-igev-kitti15	portable	190.6 ms / 282 / 362	575.3 ms / 769 / 1036	1420.8 ms / 1690 / 2030	8918.8 ms (n=3) / 10549 / 14320
selective-igev-kitti15	kernels	82.5 ms / 260 / 368	212.8 ms / 693 / 1010	471.9 ms / 1502 / 2046	2803.9 ms (n=3) / 6928 / 8800
selective-igev-middlebury-finetune	portable	190.6 ms / 282 / 362	575.6 ms / 769 / 1036	1420.8 ms / 1690 / 2030	8923.1 ms (n=3) / 10549 / 14320
selective-igev-middlebury-finetune	kernels	82.3 ms / 260 / 368	212.6 ms / 693 / 1010	472.1 ms / 1502 / 2046	2804.4 ms (n=3) / 6928 / 8800
waft-stereo-b	portable	SKIPPED	SKIPPED	SKIPPED	SKIPPED
waft-stereo-b	kernels	62.9 ms / 676 / 908	214.9 ms / 1310 / 1880	607.1 ms / 2523 / 3712	OOM
waft-stereo-s	portable	83.4 ms / 941 / 1146	357.7 ms / 2165 / 3356	1234.1 ms / 5368 / 8448	SKIPPED
waft-stereo-s	kernels	54.4 ms / 381 / 516	148.0 ms / 760 / 1120	441.0 ms / 1473 / 2150	6423.1 ms (n=3) / 7556 / 11364
da3-giant-1.1	portable	112.6 ms / 5032 / 5332	90.1 ms / 4930 / 5150	90.3 ms / 4943 / 5150	118.5 ms / 5206 / 5524
da3-giant-1.1	kernels	49.5 ms / 2630 / 3032	36.0 ms / 2562 / 2854	36.7 ms / 2575 / 2886	54.5 ms / 3011 / 3386
da3-large-1.1	portable	45.6 ms / 1733 / 1992	37.8 ms / 1639 / 1814	38.8 ms / 1651 / 1826	51.5 ms / 1938 / 2214
da3-large-1.1	kernels	18.3 ms / 967 / 1218	14.2 ms / 901 / 1068	14.8 ms / 914 / 1108	24.1 ms / 1354 / 1608
da3metric-large	portable	32.3 ms / 1542 / 1652	25.4 ms / 1488 / 1576	26.0 ms / 1502 / 1600	38.1 ms / 1896 / 2184
da3metric-large	kernels	15.7 ms / 805 / 880	12.1 ms / 782 / 838	12.8 ms / 795 / 858	21.5 ms / 1266 / 1570
da3mono-large	portable	32.2 ms / 1542 / 1652	25.2 ms / 1488 / 1576	25.9 ms / 1502 / 1600	38.0 ms / 1896 / 2184
da3mono-large	kernels	15.6 ms / 805 / 880	12.0 ms / 782 / 838	12.6 ms / 795 / 858	21.3 ms / 1266 / 1570
da3nested-giant-large-1.1	portable	145.3 ms / 6309 / 6624	114.9 ms / 6206 / 6432	115.5 ms / 6218 / 6440	151.5 ms / 6482 / 6810
da3nested-giant-large-1.1	kernels	65.6 ms / 3276 / 3658	48.0 ms / 3206 / 3490	48.7 ms / 3220 / 3528	71.2 ms / 3658 / 4168
depth-anything-v2-b	portable	18.8 ms / 643 / 776	26.2 ms / 733 / 876	26.3 ms / 745 / 908	20.0 ms / 787 / 924
depth-anything-v2-b	kernels	9.5 ms / 325 / 408	13.6 ms / 376 / 466	13.7 ms / 390 / 486	10.7 ms / 606 / 622
depth-anything-v2-l	portable	55.9 ms / 1752 / 1946	78.6 ms / 1915 / 2170	78.8 ms / 1928 / 2180	57.1 ms / 1885 / 2216
depth-anything-v2-l	kernels	28.3 ms / 875 / 1000	38.7 ms / 965 / 1104	38.8 ms / 977 / 1112	29.3 ms / 1058 / 1174
depth-anything-v2-s	portable	8.1 ms / 262 / 326	10.9 ms / 314 / 376	11.0 ms / 329 / 422	9.5 ms / 509 / 536
depth-anything-v2-s	kernels	4.0 ms / 129 / 166	5.7 ms / 164 / 218	5.8 ms / 177 / 214	5.1 ms / 462 / 480
depth-pro	portable	558.6 ms / 3317 / 4044	558.7 ms / 3323 / 4080	559.0 ms / 3337 / 4084	562.3 ms / 3451 / 4120
depth-pro	kernels	272.9 ms / 3621 / 4372	272.1 ms / 3628 / 4388	271.8 ms / 3641 / 4386	274.9 ms / 3754 / 4500
marigold-depth-v1-1	portable	455.3 ms / 2898 / 3650	351.7 ms / 2648 / 3306	351.8 ms / 2661 / 3378	458.1 ms / 3033 / 3844
marigold-depth-v1-1	kernels	433.0 ms / 2636 / 3464	324.3 ms / 2452 / 3070	324.9 ms / 2465 / 3142	435.6 ms / 2770 / 3552
metric3d-vit-large	portable	207.0 ms / 2219 / 2422	207.0 ms / 2228 / 2440	207.3 ms / 2242 / 2428	211.9 ms / 2401 / 2926
metric3d-vit-large	kernels	158.4 ms / 1471 / 1796	158.4 ms / 1478 / 1812	158.8 ms / 1492 / 1822	162.9 ms / 1616 / 2022
midas-dpt-beit-large-512	portable	79.3 ms / 1729 / 1880	141.1 ms / 1870 / 2052	141.2 ms / 1883 / 2080	81.6 ms / 1863 / 1990
midas-dpt-beit-large-512	kernels	52.1 ms / 922 / 1048	96.0 ms / 1012 / 1220	96.0 ms / 1025 / 1280	53.1 ms / 1077 / 1178
moge-2-vitl-normal	portable	175.5 ms / 2275 / 2508	178.8 ms / 2286 / 2518	179.2 ms / 2299 / 2528	191.5 ms / 2410 / 2562
moge-2-vitl-normal	kernels	74.3 ms / 1435 / 1572	77.5 ms / 1444 / 1588	76.0 ms / 1459 / 1644	90.5 ms / 1755 / 2174
moge-2-vits-normal	portable	88.6 ms / 1123 / 1418	85.4 ms / 1132 / 1392	86.7 ms / 1146 / 1446	98.6 ms / 1257 / 1464
moge-2-vits-normal	kernels	33.9 ms / 818 / 970	33.1 ms / 826 / 976	31.0 ms / 839 / 1010	46.9 ms / 1144 / 1566
moge-3-vitl	portable	657.4 ms / 2580 / 3070	633.4 ms / 2577 / 2792	620.0 ms / 2588 / 2804	655.1 ms / 2701 / 3066
moge-3-vitl	kernels	386.2 ms / 1570 / 1836	374.2 ms / 1557 / 1840	362.3 ms / 1566 / 1890	390.6 ms / 1838 / 2300
unidepth-v2-vitl14	portable	61.2 ms / 1993 / 2266	130.9 ms / 2561 / 3154	131.1 ms / 2575 / 3150	134.4 ms / 2674 / 3266
unidepth-v2-vitl14	kernels	32.5 ms / 1011 / 1110	69.1 ms / 1214 / 1378	69.4 ms / 1227 / 1430	72.6 ms / 1334 / 1706
unik3d-vitl	portable	115.1 ms / 2101 / 2474	124.2 ms / 2153 / 2570	124.8 ms / 2166 / 2582	128.7 ms / 2271 / 2486
unik3d-vitl	kernels	57.9 ms / 1076 / 1202	62.6 ms / 1101 / 1272	63.2 ms / 1115 / 1240	65.9 ms / 1226 / 1512
zoedepth-nk	portable	107.5 ms / 1857 / 1966	147.2 ms / 2046 / 2184	147.6 ms / 2107 / 2220	113.0 ms / 2451 / 2658
zoedepth-nk	kernels	74.9 ms / 1548 / 1802	105.6 ms / 1847 / 2138	106.2 ms / 1907 / 2230	80.3 ms / 2142 / 2486
deeplabv3-r101-d8-ade20k	portable	33.5 ms / 550 / 784	45.5 ms / 978 / 1282	52.4 ms / 1650 / 1966	103.8 ms / 7442 / 7698
deeplabv3-r101-d8-ade20k	kernels	20.0 ms / 583 / 718	28.5 ms / 1012 / 1196	35.5 ms / 1684 / 1880	90.5 ms / 7476 / 7708
deeplabv3plus-r101-d8-ade20k	portable	35.4 ms / 513 / 766	48.8 ms / 944 / 1272	55.8 ms / 1616 / 1944	105.7 ms / 7396 / 7644
deeplabv3plus-r101-d8-ade20k	kernels	23.8 ms / 546 / 716	34.0 ms / 977 / 1222	41.0 ms / 1649 / 1902	94.4 ms / 7429 / 7668
fcn-r101-d8-ade20k	portable	28.6 ms / 515 / 720	38.9 ms / 943 / 1230	46.0 ms / 1614 / 1894	99.2 ms / 7407 / 7662
fcn-r101-d8-ade20k	kernels	17.6 ms / 549 / 692	25.5 ms / 977 / 1154	32.5 ms / 1649 / 1854	88.2 ms / 7441 / 7682
hrnet-w48-fcn-ade20k	portable	31.0 ms / 523 / 696	36.1 ms / 954 / 1194	42.8 ms / 1626 / 1836	101.5 ms / 7406 / 7632
hrnet-w48-fcn-ade20k	kernels	23.0 ms / 523 / 652	27.7 ms / 955 / 1114	35.1 ms / 1627 / 1824	93.3 ms / 7406 / 7694
pspnet-r101-d8-ade20k	portable	31.8 ms / 513 / 810	46.8 ms / 941 / 1296	53.8 ms / 1613 / 1976	102.4 ms / 7406 / 7632
pspnet-r101-d8-ade20k	kernels	20.5 ms / 548 / 722	32.6 ms / 975 / 1178	39.5 ms / 1647 / 1838	91.2 ms / 7439 / 7668
dinov2-b-linear-ade20k	portable	22.5 ms / 922 / 1040	35.5 ms / 1217 / 1398	40.0 ms / 1230 / 1416	151.7 ms / 1289 / 1960
dinov2-b-linear-ade20k	kernels	18.3 ms / 929 / 1080	28.4 ms / 1224 / 1438	34.1 ms / 1237 / 1456	147.9 ms / 1296 / 2000
dinov2-l-linear-ade20k	portable	56.5 ms / 1330 / 1524	86.3 ms / 1624 / 1876	91.1 ms / 1638 / 1890	184.9 ms / 1697 / 2266
dinov2-l-linear-ade20k	kernels	39.6 ms / 1338 / 1526	62.8 ms / 1632 / 1884	67.7 ms / 1645 / 1894	168.5 ms / 1704 / 2264
dinov2-s-linear-ade20k	portable	12.6 ms / 792 / 834	20.1 ms / 1088 / 1154	25.1 ms / 1101 / 1168	141.7 ms / 1160 / 1522
dinov2-s-linear-ade20k	kernels	11.5 ms / 796 / 854	18.5 ms / 1092 / 1168	23.5 ms / 1105 / 1174	140.4 ms / 1163 / 1774
eomt-l-dinov2-ade20k	portable	50.3 ms / 1917 / 2484	55.4 ms / 2412 / 3506	63.2 ms / 3417 / 4834	129.2 ms / 14890 / 15016
eomt-l-dinov2-ade20k	kernels	38.5 ms / 1917 / 2484	43.8 ms / 2412 / 3506	51.5 ms / 3417 / 4834	117.4 ms / 14890 / 15016
eomt-l-dinov3-ade20k	portable	54.2 ms / 1916 / 2486	59.2 ms / 2411 / 3502	67.0 ms / 3415 / 4836	133.1 ms / 14889 / 15016
eomt-l-dinov3-ade20k	kernels	42.5 ms / 1916 / 2486	47.7 ms / 2411 / 3502	55.3 ms / 3415 / 4836	121.2 ms / 14889 / 15016
mask2former-swinl-ade20k	portable	71.9 ms / 1194 / 1912	96.1 ms / 1698 / 3182	104.1 ms / 2985 / 4500	152.8 ms / 14457 / 15480
mask2former-swinl-ade20k	kernels	45.3 ms / 1194 / 1816	60.5 ms / 1699 / 3056	68.5 ms / 2986 / 4374	126.0 ms / 14457 / 15344
oneformer-swinl-ade20k	portable	86.2 ms / 1295 / 1940	120.8 ms / 2220 / 3960	138.9 ms / 3772 / 6702	OOM
oneformer-swinl-ade20k	kernels	59.0 ms / 1295 / 1858	85.4 ms / 2222 / 3854	103.3 ms / 3772 / 6642	OOM
sapiens-seg-0.3b-goliath	portable	131.3 ms / 1621 / 1960	132.2 ms / 1627 / 1964	133.7 ms / 1641 / 1950	146.0 ms / 2254 / 3302
sapiens-seg-0.3b-goliath	kernels	92.6 ms / 1621 / 1960	93.5 ms / 1627 / 1964	95.2 ms / 1641 / 1950	107.5 ms / 2254 / 3302
sapiens-seg-1b-goliath	portable	274.9 ms / 3238 / 3718	275.7 ms / 3245 / 3734	277.0 ms / 3259 / 3730	289.7 ms / 3869 / 5076
sapiens-seg-1b-goliath	kernels	185.7 ms / 3238 / 3718	186.6 ms / 3245 / 3734	187.6 ms / 3259 / 3730	199.5 ms / 3869 / 5076
segformer-b0-ade20k	portable	7.5 ms / 237 / 290	11.9 ms / 454 / 518	18.5 ms / 798 / 864	74.7 ms / 3888 / 4048
segformer-b0-ade20k	kernels	7.3 ms / 237 / 290	11.6 ms / 454 / 518	18.2 ms / 798 / 864	74.5 ms / 3888 / 4048
segformer-b5-ade20k	portable	28.3 ms / 469 / 668	36.7 ms / 707 / 958	43.4 ms / 1050 / 1304	96.1 ms / 4063 / 4380
segformer-b5-ade20k	kernels	26.6 ms / 470 / 668	33.7 ms / 708 / 958	40.4 ms / 1051 / 1304	94.4 ms / 4064 / 4380
sam-vit-b	portable	48.6 ms / 367 / 482	48.6 ms / 374 / 482	49.0 ms / 387 / 484	-
sam-vit-b	kernels	37.3 ms / 364 / 424	37.4 ms / 370 / 404	38.1 ms / 384 / 432	-
sam-vit-h	portable	216.1 ms / 1490 / 1882	215.4 ms / 1496 / 1902	215.7 ms / 1510 / 1916	-
sam-vit-h	kernels	143.6 ms / 1459 / 1712	144.4 ms / 1466 / 1704	145.1 ms / 1479 / 1714	-
sam-vit-l	portable	117.2 ms / 814 / 1052	116.5 ms / 820 / 1060	117.5 ms / 834 / 1048	-
sam-vit-l	kernels	82.0 ms / 812 / 930	81.6 ms / 818 / 946	83.2 ms / 833 / 922	-
sam2.1-hiera-base-plus	portable	32.0 ms / 395 / 540	31.9 ms / 402 / 558	32.1 ms / 415 / 568	-
sam2.1-hiera-base-plus	kernels	25.1 ms / 408 / 472	25.1 ms / 415 / 490	25.3 ms / 428 / 472	-
sam2.1-hiera-large	portable	68.2 ms / 724 / 936	68.6 ms / 732 / 948	68.3 ms / 745 / 954	-
sam2.1-hiera-large	kernels	48.7 ms / 742 / 824	48.7 ms / 750 / 836	49.0 ms / 763 / 842	-
sam2.1-hiera-small	portable	18.7 ms / 298 / 420	19.1 ms / 305 / 422	18.9 ms / 319 / 426	-
sam2.1-hiera-small	kernels	15.4 ms / 310 / 356	15.5 ms / 317 / 358	15.6 ms / 331 / 346	-
sam2.1-hiera-tiny	portable	16.7 ms / 285 / 410	16.6 ms / 292 / 410	16.8 ms / 305 / 426	-
sam2.1-hiera-tiny	kernels	13.8 ms / 296 / 346	13.8 ms / 304 / 346	14.0 ms / 317 / 346	-
sam3	portable	179.8 ms / 1928 / 2196	181.5 ms / 1935 / 2208	182.6 ms / 1948 / 2192	-
sam3	kernels	116.3 ms / 2052 / 2460	116.5 ms / 2059 / 2420	116.9 ms / 2072 / 2490	-
detr-r101-coco	portable	14.4 ms / 278 / 332	17.6 ms / 307 / 356	18.3 ms / 320 / 416	23.8 ms / 636 / 676
detr-r101-coco	kernels	10.5 ms / 228 / 278	13.0 ms / 248 / 294	13.3 ms / 262 / 354	19.8 ms / 637 / 678
detr-r50-coco	portable	9.5 ms / 241 / 296	11.6 ms / 270 / 320	12.2 ms / 283 / 380	18.7 ms / 599 / 640
detr-r50-coco	kernels	7.7 ms / 191 / 240	8.9 ms / 211 / 256	9.8 ms / 225 / 316	16.7 ms / 600 / 640
faster-rcnn-r50-fpn-3x-coco	portable	18.5 ms / 309 / 420	21.5 ms / 338 / 426	22.2 ms / 352 / 454	27.2 ms / 602 / 734
faster-rcnn-r50-fpn-3x-coco	kernels	16.1 ms / 310 / 422	18.7 ms / 339 / 460	18.4 ms / 353 / 488	24.6 ms / 602 / 736
mask-rcnn-r50-fpn-3x-coco	portable	18.1 ms / 314 / 422	23.1 ms / 348 / 440	25.1 ms / 413 / 536	27.1 ms / 607 / 736
mask-rcnn-r50-fpn-3x-coco	kernels	15.6 ms / 315 / 422	19.6 ms / 349 / 472	22.3 ms / 414 / 568	24.6 ms / 607 / 736
rtdetrv2-l-coco	portable	24.1 ms / 171 / 202	15.4 ms / 176 / 204	13.2 ms / 204 / 226	18.2 ms / 574 / 698
rtdetrv2-l-coco	kernels	6.4 ms / 205 / 328	6.5 ms / 213 / 338	7.1 ms / 240 / 338	16.4 ms / 610 / 758
rtdetrv2-m-coco	portable	9.4 ms / 147 / 178	11.9 ms / 155 / 202	12.8 ms / 181 / 216	16.1 ms / 552 / 688
rtdetrv2-m-coco	kernels	5.1 ms / 182 / 312	5.3 ms / 189 / 324	5.8 ms / 216 / 336	15.1 ms / 586 / 752
rtdetrv2-ms-coco	portable	9.9 ms / 160 / 190	9.5 ms / 166 / 200	9.7 ms / 193 / 228	16.6 ms / 563 / 700
rtdetrv2-ms-coco	kernels	5.3 ms / 195 / 314	5.4 ms / 202 / 326	6.0 ms / 229 / 338	15.2 ms / 600 / 760
rtdetrv2-s-coco	portable	7.7 ms / 128 / 154	17.8 ms / 135 / 164	10.2 ms / 160 / 192	14.8 ms / 530 / 664
rtdetrv2-s-coco	kernels	4.0 ms / 161 / 288	4.1 ms / 168 / 300	4.7 ms / 194 / 312	14.0 ms / 564 / 728
rtdetrv2-x-coco	portable	15.4 ms / 237 / 264	23.5 ms / 244 / 274	18.4 ms / 266 / 302	22.8 ms / 636 / 774
rtdetrv2-x-coco	kernels	9.0 ms / 272 / 390	9.2 ms / 279 / 402	9.8 ms / 302 / 414	19.0 ms / 673 / 836
dinov2-vitb14	portable	10.3 ms / 231 / 242	39.8 ms / 295 / 342	137.6 ms / 417 / 442	3232.4 ms (n=3) / 1460 / 1562
dinov2-vitb14	kernels	7.5 ms / 243 / 262	28.0 ms / 324 / 362	94.7 ms / 480 / 542	2147.8 ms (n=3) / 1821 / 2002
dinov2-vitb14-reg	portable	10.1 ms / 238 / 254	39.9 ms / 303 / 368	140.2 ms / 423 / 544	3244.0 ms (n=3) / 1467 / 2244
dinov2-vitb14-reg	kernels	7.6 ms / 254 / 374	28.1 ms / 337 / 452	94.7 ms / 492 / 704	2147.6 ms (n=3) / 1833 / 2404
dinov2-vitg14	portable	101.5 ms / 2273 / 2304	375.9 ms / 2410 / 2464	1169.8 ms / 2673 / 2764	23035.0 ms (n=3) / 4929 / 5404
dinov2-vitg14	kernels	69.7 ms / 2296 / 2322	257.4 ms / 2469 / 2522	797.8 ms / 2801 / 2924	15248.1 ms (n=3) / 5650 / 6264
dinov2-vitg14-reg	portable	101.8 ms / 2294 / 2454	377.1 ms / 2432 / 2656	1184.7 ms / 2694 / 2984	23131.1 ms (n=3) / 4951 / 6020
dinov2-vitg14-reg	kernels	70.3 ms / 2298 / 2634	257.6 ms / 2472 / 2864	802.9 ms / 2803 / 3352	15244.2 ms (n=3) / 5654 / 6720
dinov2-vitl14	portable	29.8 ms / 656 / 682	118.1 ms / 736 / 782	395.5 ms / 889 / 902	8770.0 ms (n=3) / 2205 / 2482
dinov2-vitl14	kernels	19.6 ms / 671 / 702	77.1 ms / 775 / 802	252.1 ms / 974 / 1042	5331.0 ms (n=3) / 2686 / 2862
dinov3-convnext-base	portable	8.3 ms / 253 / 290	22.8 ms / 357 / 410	55.2 ms / 565 / 650	323.2 ms / 2310 / 2670
dinov3-convnext-base	kernels	6.5 ms / 279 / 360	15.1 ms / 372 / 572	33.5 ms / 502 / 548	190.9 ms / 1914 / 2048
dinov3-convnext-large	portable	14.8 ms / 484 / 510	42.5 ms / 633 / 766	99.9 ms / 932 / 1048	587.2 ms / 3435 / 4008
dinov3-convnext-large	kernels	11.0 ms / 505 / 642	26.7 ms / 635 / 950	61.0 ms / 837 / 926	357.8 ms / 2887 / 2946
dinov3-convnext-small	portable	5.9 ms / 169 / 188	15.6 ms / 250 / 266	37.9 ms / 413 / 486	220.9 ms / 1778 / 2106
dinov3-convnext-small	kernels	4.5 ms / 197 / 276	10.6 ms / 272 / 410	23.3 ms / 366 / 406	133.7 ms / 1474 / 1526
dinov3-convnext-tiny	portable	3.6 ms / 127 / 148	9.8 ms / 209 / 226	23.9 ms / 372 / 446	136.6 ms / 1737 / 2066
dinov3-convnext-tiny	kernels	2.7 ms / 156 / 236	6.5 ms / 231 / 368	14.6 ms / 324 / 366	82.1 ms / 1433 / 1484
dinov3-vitb16	portable	8.8 ms / 224 / 242	31.6 ms / 281 / 304	102.4 ms / 389 / 464	2039.3 ms (n=3) / 1308 / 1502
dinov3-vitb16	kernels	5.3 ms / 233 / 326	19.4 ms / 297 / 324	64.4 ms / 424 / 462	1322.5 ms / 1515 / 1702
dinov3-vith16plus	portable	60.7 ms / 1689 / 1724	218.8 ms / 1795 / 1846	636.2 ms / 1996 / 2106	10005.0 ms (n=3) / 3712 / 3944
dinov3-vith16plus	kernels	38.6 ms / 1690 / 1786	138.3 ms / 1789 / 1806	393.4 ms / 1984 / 2024	6021.9 ms (n=3) / 3643 / 3804
dinov3-vitl16	portable	25.9 ms / 645 / 662	95.3 ms / 715 / 764	298.3 ms / 847 / 924	5549.3 ms (n=3) / 1975 / 2142
dinov3-vitl16	kernels	15.6 ms / 656 / 726	54.9 ms / 738 / 764	174.5 ms / 899 / 922	3262.4 ms (n=3) / 2274 / 2402
dinov3-vits16	portable	4.8 ms / 92 / 104	12.9 ms / 130 / 158	44.4 ms / 201 / 238	990.9 ms / 807 / 1036
dinov3-vits16	kernels	2.3 ms / 97 / 168	8.6 ms / 145 / 242	29.7 ms / 213 / 236	644.8 ms / 876 / 1076
fast-acvnet-plus-kitti15	portable	14.3 ms / 136 / 208	38.0 ms / 401 / 664	99.4 ms / 887 / 1418	623.7 ms / 5054 / 8076
fast-acvnet-plus-kitti15	kernels	7.8 ms / 169 / 546	25.3 ms / 457 / 598	66.2 ms / 1006 / 1226	422.8 ms / 5743 / 7342
fast-acvnet-plus-sceneflow	portable	14.3 ms / 136 / 208	38.0 ms / 401 / 664	99.4 ms / 887 / 1418	623.7 ms / 5054 / 8076
fast-acvnet-plus-sceneflow	kernels	7.8 ms / 169 / 546	25.3 ms / 457 / 598	66.1 ms / 1006 / 1226	422.7 ms / 5743 / 7342
fast-acvnet-sceneflow	portable	14.0 ms / 137 / 208	31.1 ms / 402 / 664	78.2 ms / 890 / 1418	492.4 ms / 5077 / 8074
fast-acvnet-sceneflow	kernels	6.8 ms / 169 / 582	22.0 ms / 435 / 608	55.6 ms / 954 / 1372	357.6 ms / 5445 / 7618
selective-igev-middlebury-train	portable	190.7 ms / 282 / 362	575.4 ms / 769 / 1036	1420.5 ms / 1690 / 2030	8921.4 ms (n=3) / 10549 / 14320
selective-igev-middlebury-train	kernels	82.5 ms / 260 / 368	213.1 ms / 693 / 1010	472.4 ms / 1502 / 2046	2804.8 ms (n=3) / 6928 / 8800
selective-igev-sceneflow	portable	190.6 ms / 282 / 362	575.4 ms / 769 / 1036	1420.3 ms / 1690 / 2030	8917.6 ms (n=3) / 10549 / 14320
selective-igev-sceneflow	kernels	82.5 ms / 260 / 368	213.5 ms / 693 / 1010	473.3 ms / 1502 / 2046	2808.4 ms (n=3) / 6928 / 8800
dinov2-vitl14-reg	portable	30.1 ms / 656 / 682	118.2 ms / 736 / 782	396.1 ms / 889 / 982	8778.1 ms (n=3) / 2205 / 2622
dinov2-vitl14-reg	kernels	19.6 ms / 671 / 702	77.1 ms / 775 / 802	252.5 ms / 974 / 1042	5326.8 ms (n=3) / 2686 / 2982
dinov2-vits14	portable	4.3 ms / 95 / 116	16.8 ms / 133 / 176	61.4 ms / 207 / 256	1574.5 ms / 843 / 976
dinov2-vits14	kernels	3.3 ms / 100 / 116	12.0 ms / 148 / 176	39.0 ms / 239 / 276	957.1 ms / 1023 / 1116
dinov2-vits14-reg	portable	4.3 ms / 96 / 108	16.9 ms / 135 / 164	61.3 ms / 210 / 248	1582.1 ms / 846 / 1240
dinov2-vits14-reg	kernels	3.3 ms / 106 / 148	12.0 ms / 153 / 176	39.0 ms / 244 / 308	960.9 ms / 1029 / 1344
vggt-1b	portable	793.4 ms / 3868 / 5308	561.3 ms / 3710 / 4930	566.5 ms / 3920 / 5178	849.5 ms / 8219 / 10306
vggt-1b	kernels	559.0 ms / 3880 / 5474	398.6 ms / 3721 / 5156	404.0 ms / 3931 / 5390	615.0 ms / 8232 / 10518
vggt-omega-1b-512	portable	722.8 ms / 3449 / 3866	723.2 ms / 3559 / 4038	725.7 ms / 3771 / 4366	757.0 ms / 5594 / 6894
vggt-omega-1b-512	kernels	461.8 ms / 3449 / 3796	461.1 ms / 3560 / 3968	463.9 ms / 3771 / 4332	493.0 ms / 5594 / 6894
mapanything-apache	portable	680.9 ms / 3914 / 5848	493.1 ms / 3635 / 5196	498.1 ms / 3846 / 5442	734.6 ms / 7240 / 9442
mapanything-apache	kernels	569.7 ms / 3914 / 5846	415.3 ms / 3635 / 5196	419.9 ms / 3846 / 5442	622.3 ms / 7240 / 9440
mvsanywhere	portable	107.5 ms / 969 / 1500	75.8 ms / 954 / 1398	77.4 ms / 1166 / 1756	124.0 ms / 3110 / 4342
mvsanywhere	kernels	40.6 ms / 769 / 1006	31.0 ms / 838 / 1068	32.5 ms / 1049 / 1414	57.1 ms / 2908 / 4006
patchmatchnet	portable	43.1 ms / 439 / 522	154.4 ms / 1242 / 1534	408.6 ms / 2753 / 3346	OOM
patchmatchnet	kernels	9.2 ms / 253 / 328	24.4 ms / 693 / 942	56.7 ms / 1520 / 2016	378.6 ms / 8628 / 11332
great-igev-eth3d	portable	206.7 ms / 681 / 1030	627.7 ms / 1216 / 2262	1532.6 ms / 1688 / 2780	13067.4 ms (n=3) / 11364 / 15106
great-igev-eth3d	kernels	99.0 ms / 385 / 476	241.2 ms / 1041 / 1338	544.2 ms / 1506 / 2354	3753.5 ms (n=3) / 11736 / 15500
great-igev-kitti	portable	217.1 ms / 681 / 1030	628.2 ms / 1216 / 2262	1532.7 ms / 1688 / 2780	13065.3 ms (n=3) / 11364 / 15106
great-igev-kitti	kernels	93.2 ms / 385 / 476	240.9 ms / 1041 / 1338	544.6 ms / 1506 / 2356	3755.4 ms (n=3) / 11736 / 15500
great-igev-middlebury	portable	328.5 ms / 1082 / 1888	1025.5 ms / 1344 / 2262	2554.0 ms (n=3) / 2549 / 4266	OOM
great-igev-middlebury	kernels	131.4 ms / 942 / 1258	356.0 ms / 1279 / 1906	804.4 ms / 1825 / 2656	6324.4 ms (n=3) / 11920 / 15500
great-igev-sceneflow	portable	200.9 ms / 681 / 1030	624.1 ms / 1216 / 2262	1531.8 ms / 1688 / 2780	13071.3 ms (n=3) / 11364 / 15106
great-igev-sceneflow	kernels	93.5 ms / 385 / 476	241.1 ms / 1041 / 1338	545.3 ms / 1506 / 2354	3759.3 ms (n=3) / 11736 / 15500
igev-plusplus-eth3d	portable	201.5 ms / 354 / 614	606.5 ms / 987 / 1662	1534.8 ms / 2172 / 3496	OOM
igev-plusplus-eth3d	kernels	88.7 ms / 357 / 1092	241.2 ms / 830 / 1210	534.3 ms / 1799 / 2464	3135.8 ms (n=3) / 8419 / 9752
igev-plusplus-kitti12	portable	201.4 ms / 354 / 614	607.2 ms / 987 / 1662	1535.8 ms / 2172 / 3496	OOM
igev-plusplus-kitti12	kernels	89.1 ms / 357 / 1092	244.8 ms / 830 / 1210	535.9 ms / 1799 / 2464	3143.3 ms (n=3) / 8419 / 9750
igev-plusplus-kitti15	portable	202.7 ms / 354 / 614	606.9 ms / 987 / 1662	1535.8 ms / 2172 / 3496	OOM
igev-plusplus-kitti15	kernels	88.7 ms / 357 / 1092	240.2 ms / 830 / 1210	534.9 ms / 1799 / 2464	3141.8 ms (n=3) / 8419 / 9752
igev-plusplus-middlebury	portable	201.2 ms / 354 / 612	606.8 ms / 987 / 1662	1535.6 ms / 2172 / 3496	OOM
igev-plusplus-middlebury	kernels	88.7 ms / 357 / 1092	240.5 ms / 830 / 1210	535.0 ms / 1799 / 2464	3140.6 ms (n=3) / 8419 / 9752
igev-plusplus-sceneflow	portable	201.0 ms / 354 / 612	606.2 ms / 987 / 1662	1534.9 ms / 2172 / 3496	OOM
igev-plusplus-sceneflow	kernels	88.6 ms / 357 / 1092	240.6 ms / 830 / 1210	535.4 ms / 1799 / 2464	3143.9 ms (n=3) / 8419 / 9752