IGEV++

• Original code: https://github.com/gangweix/IGEV-plusplus at 099dec979897
• Code licence: MIT (permissive); files: LICENSE
• Weights licence: MIT (assumed from code)
• Licence evidence: https://github.com/gangweix/IGEV-plusplus/blob/099dec97989780d151700459c682df1cc2c18e88/README.md#L23
• Training data: Freiburg SceneFlow (FlyingThings3D, Driving, Monkaa)
  • igev-plusplus-kitti12: Freiburg SceneFlow (FlyingThings3D, Driving, Monkaa), KITTI stereo 2012 / 2015 (SceneFlow, then fine-tuned on KITTI 2012 + 2015 (README recipe; KITTI 2012 model))
  • igev-plusplus-kitti15: Freiburg SceneFlow (FlyingThings3D, Driving, Monkaa), KITTI stereo 2012 / 2015 (SceneFlow, then fine-tuned on KITTI 2012 + 2015 (README recipe; KITTI 2015 model))
  • igev-plusplus-middlebury: Freiburg SceneFlow (FlyingThings3D, Driving, Monkaa), TartanAir, NVIDIA Falling Things (FAT), CREStereo synthetic stereo dataset (Megvii), InStereo2K, Middlebury stereo (2014), not stated by the authors (Middlebury submission model; original recipe (README, core/stereo_datasets.py): SceneFlow model, then TartanAir, SceneFlow, Falling Things, InStereo2K, CARLA, CREStereo and Middlebury 2005 / 2006 / 2014 / 2021 / MiddEval3, then a fine-tune on CREStereo, InStereo2K, CARLA, Falling Things and Middlebury)
  • igev-plusplus-eth3d: Freiburg SceneFlow (FlyingThings3D, Driving, Monkaa), TartanAir, MPI Sintel stereo, CREStereo synthetic stereo dataset (Megvii), InStereo2K, ETH3D two-view stereo (ETH3D submission model; original recipe (README, core/stereo_datasets.py): SceneFlow model, then TartanAir, SceneFlow, Sintel, CREStereo, InStereo2K and ETH3D, then a fine-tune on CREStereo, InStereo2K and ETH3D)
• Options: iters = 32 (>= 1), cuda_graphs = True (speed only)
• Kernel libraries: igev, igevpp, cutlass_conv, cudnn_hacc
• Variants: igev-plusplus-sceneflow, igev-plusplus-kitti12, igev-plusplus-kitti15, igev-plusplus-middlebury, igev-plusplus-eth3d
• Outputs: disparity
• Input size: any H x W >= 32, padded to a multiple of 32
• Tested up to (portable, fp16, RTX 5060 Ti 16 GB): igev-plusplus-sceneflow not measured yet; igev-plusplus-kitti12 not measured yet; igev-plusplus-kitti15 not measured yet; igev-plusplus-middlebury not measured yet; igev-plusplus-eth3d not measured yet
• igev-plusplus-kitti12, igev-plusplus-kitti15, igev-plusplus-middlebury, igev-plusplus-eth3d: same network and settings as igev-plusplus-sceneflow, measured once (rows not repeated).
• igev-plusplus-sceneflow original (fp32) 994x1482, original 994x1482, portable 994x1482, kernels 994x1482: measured on igev-plusplus-eth3d (same architecture, other checkpoint)
• igev-plusplus-sceneflow load / first call: original 0.42 s / 2.335 s (375x1242), portable 0.2 s / 1.099† s, kernels 0.16† s / 0.396† s
• igev-plusplus-kitti12 load / first call: original 0.42 s / 2.335 s (375x1242), portable 0.2 s / 1.099† s, kernels 0.16† s / 0.396† s
• igev-plusplus-kitti15 load / first call: original 0.42 s / 2.335 s (375x1242), portable 0.2 s / 1.099† s, kernels 0.16† s / 0.396† s
• igev-plusplus-middlebury load / first call: original 0.42 s / 2.335 s (375x1242), portable 0.2 s / 1.099† s, kernels 0.16† s / 0.396† s
• igev-plusplus-eth3d load / first call: original 0.42 s / 2.335 s (375x1242), portable 0.2 s / 1.099† s, kernels 0.16† s / 0.396† s
• † 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.
• Protocol, median of 9 after 3 warm-up, uint8 CUDA input (synthetic for bench_inputs sizes), card defaults; gc before each load; peak = max over the first call and the timed calls

Measured

Original = the original code named under Details, same checkpoint and protocol.

Measured on portable: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 468fba10, 2026-11-05; kernels: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 939adfe8, 2026-11-06.

Parity

CPU, float32: max |diff| <= 0.002 vs the original. GPU, float16 (the default), portable and kernels: mean |diff| <= 0.02 px and p99 <= 0.25 px. On real pairs up to 720x1280 the limits are mean 0.1 / p99 2.0 (precision-sensitive there, see Limitations).

Limitations

• CUDA output (float16, kernels, float32 with PyTorch's default TF32 convolutions) differs from the CPU by ~0.06 px mean / 1.2 px p99 at 720x1280, 0.005 px with full float32: rounding amplified by 32 GRU iterations. Portable and kernels reach the same ETH3D / Middlebury accuracy.