Prompt Depth Anything (fastforks)

Prompt Depth Anything (Lin et al., CVPR 2025): metric depth from an image and a low-resolution metric depth prompt (e.g. a LiDAR map), a DINOv2 ViT with a DPT head that fuses the prompt at every refinement stage.

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/DepthAnything/PromptDA at 9161547.

Author weight repos: https://huggingface.co/depth-anything/prompt-depth-anything-vitl, https://huggingface.co/depth-anything/prompt-depth-anything-vits.

Task: depth. Tier: experimental. Outputs: depth. Backends: portable, kernels.

Input: one RGB image. Output: a depth or inverse-depth map. kind (a card field, result.kind) says what the numbers mean: metric = metres; scale_invariant, affine_invariant, affine_invariant_log (log depth) and affine_invariant_inverse are relative (unknown scale or shift, not metres). Metric models that use the focal length take model(image, intrinsics=K) (K [3, 3], pixels of the input): metric3d-vit-large and da3metric-large need it for correct metres (da3metric-large: metres = output x focal / 300; without intrinsics= a 300 px focal length is assumed, wrong for a real camera). unidepth-v2-vitl14, unik3d-vitl, depth-pro and MoGe estimate the camera themselves (given intrinsics= replace fx); zoedepth-nk and the relative models (da3mono-large too) take none. Normals-only models (kind none) return result.normals [B, H, W, 3] (unit vectors, camera frame: x right, y down, z forward).

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
prompt-depth-anything-l@1 prompt_depth_anything_vitl.ckpt 4b3a3e0a60ea Apache-2.0 yes
prompt-depth-anything-s@1 prompt_depth_anything_vits.ckpt 51b99fef5e3c Apache-2.0 yes

Licence

Original-code strings, recorded as they are (research-use package; check them before any other use). Code: Apache-2.0. Licence files: LICENSE.

Training data: Synthetic and real RGB-D data (including pseudo-labelled real images) per the paper (arXiv 2412.14015).

Licence evidence: https://github.com/DepthAnything/PromptDA/blob/9161547ca7151e14fcc8f490ad3cc02fa9437483/README.md and the Hugging Face model cards (license: apache-2.0)

Input size

any H x W >= 1, resized or padded inside the model (ViT patch 14) Tested up to (portable, fp16, RTX 5060 Ti 16 GB): prompt-depth-anything-l 3000x4000 (1.4 GiB peak allocated, 1.7 reserved); prompt-depth-anything-s 3000x4000 (0.6 GiB peak allocated, 0.8 reserved)

Speed and memory

Depth, 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
prompt-depth-anything-l 756x1008 – – 125 69.3 – 1.8x – 1318 / 1488 1179 / 1348
prompt-depth-anything-s 756x1008 – – 18.4 11.4 – 1.6x – 363 / 442 297 / 384

Startup and first call

model load (s): original portable kernels first call (s): original portable kernels
prompt-depth-anything-l 4.80 (756x1008) 1.50 (756x1008) 2.78 (756x1008) 0.77 (756x1008) 0.30 (756x1008) 0.22 (756x1008)
prompt-depth-anything-s 0.42 (756x1008) 0.37 (756x1008) 1.20 (756x1008) 0.79 (756x1008) 0.24 (756x1008) 0.18 (756x1008)

torch.compile and TensorRT

model input time: portable compile TensorRT fp16 kernels preparation: compile (cold first call) ONNX + engine build peak MiB (alloc / reserved): portable compile kernels TensorRT (engine + torch)
prompt-depth-anything-l 756x1008 125 ms 117 ms 75.3 ms 69.3 ms 21.2 s 52.5 s 1318 / 1488 968 / 1148 1179 / 1348 1649 / 1714
prompt-depth-anything-s 756x1008 18.4 ms 16.3 ms 10.5 ms 11.4 ms 14.7 s 46.6 s 363 / 442 260 / 322 297 / 384 368 / 410

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 original code named above 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

Parity with the original code

Every variant is checked against reference 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| <= 0.02 and p99 <= 0.25 (absolute error of the output depth).

Limitations

  • The depth prompt is required: call the model with prompt_depth= a finite, positive map in metres (zero or missing LiDAR pixels must be filled first). Export is at a fixed size only.

Usage

from fastforks import depth

model = depth.load("prompt-depth-anything-l")  # mode="auto": the preferred backend from recorded measurements
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: prompt-depth-anything-l, prompt-depth-anything-s.

Kernels backend

depth.load("prompt-depth-anything-l", mode="kernels") runs our CUDA kernels (fastbb, 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.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for fastforks/prompt-depth-anything

Finetuned
(2)
this model

Collection including fastforks/prompt-depth-anything