UniDepth V2 (fastforks)
UniDepth V2 (Piccinelli et al. 2025): universal metric depth from a DINOv2 ViT-L encoder and a camera-prompted decoder; predicts depth, 3-D points, a confidence map and the pinhole intrinsics, or uses given intrinsics.
This repo is the model card only: it holds no weights. fastforks downloads the authors' weights from the links below and checks them against the SHA-256 in the model card. Code: https://github.com/kleinicke/fastforks. Kernels for the kernels backend: https://huggingface.co/fastforks/kernels. Original code: https://github.com/lpiccinelli-eth/UniDepth at 8d8cfe4.
Author weight repos: https://huggingface.co/lpiccinelli/unidepth-v2-vitl14.
Task: depth. Tier: experimental. Outputs: depth, points, intrinsics, confidence. Backends: portable, kernels.
Input: one RGB image. Output: a depth or inverse-depth map (metric for the models marked metric).
Variants
Weights are the authors' files (not hosted here). SHA-256 shortened to 12 hex digits; the full hashes are in card.toml.
| variant | weights (author link) | SHA-256 | weights licence | commercial use |
|---|---|---|---|---|
unidepth-v2-vitl14@1 |
model.safetensors | ba73d3de7353 |
CC-BY-NC-4.0 (assumed from code) | no |
Licence
Original-code strings, recorded as they are (research-use package; check them before any other use). Code: CC-BY-NC-4.0. Files: LICENSE.
Code licence: CC-BY-NC-4.0; the authors do not state a separate weights licence.
Training data: Mixture of public metric-depth datasets (indoor, outdoor, driving, synthetic) per the UniDepthV2 paper (arXiv 2502.20110).
Licence evidence: https://github.com/lpiccinelli-eth/UniDepth/blob/8d8cfe4c7ee15297099983607febf0d4f32eb3d6/README.md#license
Input size
any H x W >= 8, resized or padded inside the model
Tested up to (portable, fp16, RTX 5060 Ti 16 GB): unidepth-v2-vitl14 3000x4000 (2.6 GiB peak allocated, 3.2 reserved)
Speed and memory
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 |
|---|---|---|---|---|---|---|---|---|---|
unidepth-v2-vitl14 |
132.4 | 132.0 | 99.85 | 66.37 | 2.0x | 1.5x | 3230 / 3508 | 1208 / 1472 | 1211 / 1472 |
Startup and first call
| model | load (s): original | portable | kernels | first call (s): original | portable | kernels |
|---|---|---|---|---|---|---|
unidepth-v2-vitl14 |
3.80 | 4.16 | 0.90† | 0.38 | 0.15 | 0.07 |
Fixed-size speed and peak memory, all models
| model | mode | 640x480 | 1280x720 | 1920x1080 | 12 MP 4000x3000 |
|---|---|---|---|---|---|
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 |
Fixed-size cells: time (ms) / peak allocated / peak reserved memory (MiB).
Hardware, protocol and caveats: benchmarks (RTX 5060 Ti 16 GB, sm_120). Original = the authors' code at the precision of portable (fp16, or (fp32) where fp16 is unsupported or not finite); original fp32 = its default fp32 run. Peak memory is the torch allocator's.
Accuracy
On public datasets
AbsRel / delta1 after a least-squares scale + shift fit (in disparity space for inverse-depth models), mean over images. Original = the authors' code, same checkpoint and protocol.
| variant | dataset | portable | kernels | note |
|---|---|---|---|---|
unidepth-v2-vitl14 |
NYUv2 test, AbsRel / delta1 | 0.0355 / 0.9843 | 0.0355 / 0.9843 | |
unidepth-v2-vitl14 |
iBims-1, AbsRel / delta1 | 0.0298 / 0.9850 | 0.0298 / 0.9850 | |
unidepth-v2-vitl14 |
ETH3D two-view train left, AbsRel / delta1 | 0.0586 / 0.9802 | 0.0586 / 0.9802 | possibly in the training data |
unidepth-v2-vitl14 |
Middlebury 2014 trainingH left, AbsRel / delta1 | 0.0221 / 0.9970 | 0.0221 / 0.9970 | possibly in the training data |
Measured on NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, 2026-10-27.
Parity with the original code
Every variant is checked against golden outputs of the original code at its pinned commit. CPU, float32: max |diff| <= 0.002 vs the original. GPU, float16 (the default), portable and kernels: mean |diff| <= 1 % and p99 <= 2 % of the reference's median |value|.
Limitations
- The confidence head adds measurable time; the option confidence = false skips it.
Usage
from fastforks import depth
model = depth.load("unidepth-v2-vitl14") # portable PyTorch on the GPU if there is one, else the CPU
result = model("room.jpg")
result.depth # [H, W] float32 tensor; result.kind says what the maps mean
# models that take a camera: model("room.jpg", intrinsics=K), K [3, 3] in pixels
result.backend # what ran, e.g. Backend(mode='portable', device='cuda:0', dtype='float16')
Install with pip install fastforks (see https://github.com/kleinicke/fastforks). Other variants: unidepth-v2-vitl14.
Kernels backend
depth.load("unidepth-v2-vitl14", mode="kernels") runs our CUDA kernels (fastbb, cudnn_hacc, dpt_fused). It needs a prebuilt kernel for your GPU architecture (downloaded from https://huggingface.co/fastforks/kernels on first use) and a recent NVIDIA driver; if there is none it raises BackendUnavailableError with the reason and never falls back silently. mode="auto" picks kernels when it can run and logs the choice.