Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/Depth Anything

Depth Anything

Reported on 32 benchmarks across 4 tasks · 1 paper · 8 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision15 results

Depth EstimationonNYU-Depth V2
Delta < 1.25· uses extra data· 2024-01-19
0.984
best: 0.989 (UniK3D (FT, metric))
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonNYU-Depth V2
Delta < 1.25^2· uses extra data· 2024-01-19
0.998
best: 1 (HybridDepth)
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonNYU-Depth V2
Delta < 1.25^3· uses extra data· 2024-01-19
1
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonETH3D
Delta < 1.25· 2024-01-19
0.882
best: 0.981 (Distill Any Depth)
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonNYU-Depth V2
RMSE· uses extra data· 2024-01-19
0.206
best: 0.013 (Defocus/DepthNet (Normalized))
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonNYU-Depth V2
absolute relative error· uses extra data· 2024-01-19
0.056
best: 0.026 (HybridDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonNYU-Depth V2
log 10· uses extra data· 2024-01-19
0.024
best: 0.059 (SC-DepthV2)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonETH3D
absolute relative error· 2024-01-19
0.0127
best: 0.0121 (HDN)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
Delta < 1.25· uses extra data· 2024-01-19
0.982
best: 0.99 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
Delta < 1.25^2· uses extra data· 2024-01-19
0.998
best: 0.999 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
Delta < 1.25^3· uses extra data· 2024-01-19
1
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
RMSE· uses extra data· 2024-01-19
1.896
best: 1.394 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
RMSE log· uses extra data· 2024-01-19
0.069
best: 0.048 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
Sq Rel· uses extra data· 2024-01-19
0.121
best: 0.224 (SfM-Revisited)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
Depth EstimationonKITTI Eigen split
absolute relative error· uses extra data· 2024-01-19
0.046
best: 0.029 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891

Methodology15 results

3DonNYU-Depth V2
Delta < 1.25· uses extra data· 2024-01-19
0.984
best: 0.989 (UniK3D (FT, metric))
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonNYU-Depth V2
Delta < 1.25^2· uses extra data· 2024-01-19
0.998
best: 1 (HybridDepth)
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonNYU-Depth V2
Delta < 1.25^3· uses extra data· 2024-01-19
1
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonETH3D
Delta < 1.25· 2024-01-19
0.882
best: 0.981 (Distill Any Depth)
SOTA
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonNYU-Depth V2
RMSE· uses extra data· 2024-01-19
0.206
best: 0.013 (Defocus/DepthNet (Normalized))
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonNYU-Depth V2
absolute relative error· uses extra data· 2024-01-19
0.056
best: 0.026 (HybridDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonNYU-Depth V2
log 10· uses extra data· 2024-01-19
0.024
best: 0.059 (SC-DepthV2)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonETH3D
absolute relative error· 2024-01-19
0.0127
best: 0.0121 (HDN)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
Delta < 1.25· uses extra data· 2024-01-19
0.982
best: 0.99 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
Delta < 1.25^2· uses extra data· 2024-01-19
0.998
best: 0.999 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
Delta < 1.25^3· uses extra data· 2024-01-19
1
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
RMSE· uses extra data· 2024-01-19
1.896
best: 1.394 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
RMSE log· uses extra data· 2024-01-19
0.069
best: 0.048 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
Sq Rel· uses extra data· 2024-01-19
0.121
best: 0.224 (SfM-Revisited)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891
3DonKITTI Eigen split
absolute relative error· uses extra data· 2024-01-19
0.046
best: 0.029 (SPIDepth)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891

Medical1 result

Semantic SegmentationonCityscapes val
mIoU· 2024-01-19
86.2
best: 90.3 (EfficientPS (Cityscapes-fine))
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891

Audio1 result

10-shot image generationonCityscapes val
mIoU· 2024-01-19
86.2
best: 90.3 (EfficientPS (Cityscapes-fine))
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data arXiv:2401.10891