skillsafe-ai/depth-anything-v2-small

Depth Anything V2 Small (monocular depth)

Browser-ready import artifacts for depth-estimation, produced by SkillSafe's reproducible converter (models/ in skillsafe.ai) from a pinned upstream source. Every byte here is derivable from that source plus the recipe below; nothing was edited by hand.

Provenance

Upstream	https://huggingface.co/onnx-community/depth-anything-v2-small/tree/4472b7362082ad9968fee890ca0f1e5aca36b93d
Upstream SHA-256 / commit	4472b7362082ad9968fee890ca0f1e5aca36b93d
Recipe	recipes/depth-anything-v2-small.yaml — sha256 f243268eb921e74d01f0623d06ad1d74435cd52d325fdf2de3745b60961f0757
Toolchain	Python 3.12.13, torch 2.10.0, onnx 1.23.0, onnxruntime 1.30.0 on Darwin 25.6.0 arm64
Converted	2026-09-22T20:41:12+00:00

Files

file	class	size	SHA-256
config.json	bundle	0.00 MB	3aee5b9bc4f711ee885c2526d871f0c8c6c8c4b26b8e04253d0167f6a83264f5
onnx/model.onnx	registry	94.47 MB	afb6a5c28f3b6bf1618c6e43f02073ef9dfdc70e937502d51603e57b0a1df10c
onnx/model_fp16.onnx	registry	47.34 MB	2df6223f206b5164e21f664ace61dabeb9bb6a49b8b5a3e00510b4807d0f5b04
onnx/model_quantized.onnx	registry	26.00 MB	fcf51f1b230362b28690bb9d1809bf0431f29cad20534e3f589bd7285547f20d
preprocessor_config.json	bundle	0.00 MB	03576db3c13dd0471fdf5f5e1428befcb95de063fe699879150b293dc9e0a2c6

registry files are parameter files served from models.skillsafe.ai once vetted; bundle files ship inside an app; registry-shared is a runtime library reused by every model of the same architecture.

Verification

Imported as published upstream (no conversion). Each file is pinned by SHA-256 to its source; every ONNX file passed onnx.checker and a CPU smoke run under onnxruntime with zero-filled inputs at the declared shapes:

file	inputs	outputs	ms
onnx/model.onnx	pixel_values[1, 3, 518, 518]	predicted_depth[1, 518, 518]	121.6
onnx/model_fp16.onnx	pixel_values[1, 3, 518, 518]	predicted_depth[1, 518, 518]	157.7
onnx/model_quantized.onnx	pixel_values[1, 3, 518, 518]	predicted_depth[1, 518, 518]	106.7

Use in the browser

import * as ort from "onnxruntime-web";
const session = await ort.InferenceSession.create("https://huggingface.co/skillsafe-ai/depth-anything-v2-small/resolve/main/onnx/model.onnx", { executionProviders: ["webgpu", "wasm"] });

Contract (onnx/model.onnx): input pixel_values float32 ['batch_size', 3, 'height', 'width'] → output predicted_depth float32 ['floor(1.0*batch_size)', '14*floor(height/14)', '14*floor(width/14)']. Opset 14.

Licence and attribution

Depth Anything V2 Small: Lihe Yang et al. (HKU / TikTok), Apache License 2.0. https://github.com/DepthAnything/Depth-Anything-V2 — ONNX export by onnx-community.

Licence: Apache-2.0 — notice: https://github.com/DepthAnything/Depth-Anything-V2/blob/main/LICENSE. The conversion recipe and this model card are part of the SkillSafe repository and carry its licence; the weights remain under the upstream licence above.

The full manifest.json in this repo records the recipe, sources, toolchain (including the uv.lock hash) and per-file verification numbers.