MVSAnywhere (fastforks)

MVSAnywhere (Izquierdo et al. 2025): zero-shot multi-view stereo, the depth of a reference view from posed, calibrated source views. A view-agnostic MLP cost volume (warped matching features + camera metadata, softmax over the views) is encoded by a DINOv2 ViT-B fused with a second DINOv2's image features and decoded by a DPT head; the depth range comes from the camera baselines and is refined once.

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/nianticlabs/mvsanywhere at 5bd49bb.

Task: multiview. Tier: stable. Outputs: depth, mask, sky, depth_range. Backends: portable, kernels.

Input: several images of one scene (posed and calibrated for multi-view stereo, unposed otherwise). Output: depth maps and, where the model predicts them, cameras and 3-D points.

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
mvsanywhere@1	mvsanywhere_hero.ckpt	00878bab903b	LicenseRef-noncommercial AND CC-BY-NC-SA-4.0 (assumed from code)	no

Licence

Original-code strings, recorded as they are (research-use package; check them before any other use). Code: LicenseRef-noncommercial AND CC-BY-NC-SA-4.0. Files: LICENSE, LICENSE-antialiased-cnns.md.

Code licence: LicenseRef-noncommercial AND CC-BY-NC-SA-4.0; the authors do not state a separate weights licence.

Training data: Synthetic data per the paper (arXiv 2411.14390): Hypersim, TartanAir, BlendedMVG, MatrixCity, VKITTI2, Dynamic Replica, MVS-Synth, SAIL-VOS 3D; DINOv2 / Depth Anything V2 initialisation.

• Licence note: Niantic's non-commercial licence covers the code and the checkpoint; the BlurPool stem comes from Adobe's antialiased-cnns (CC-BY-NC-SA-4.0).

Licence evidence: https://github.com/nianticlabs/mvsanywhere/blob/5bd49bb/LICENSE and README.md (the checkpoint is published under the repository's licence)

Input size

any H x W >= 1, resized or padded inside the model Tested up to (portable, fp16, RTX 5060 Ti 16 GB): mvsanywhere 3000x4000 (3.0 GiB peak allocated, 4.2 reserved)

Speed and memory

Multiview, 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
mvsanywhere	448x672	273	219	93.2	35.2	6.2x	2.7x	1504 / 1576	931 / 1226	702 / 760

Startup and first call

model	load (s): original	portable	kernels	first call (s): original	portable	kernels
mvsanywhere	3.21 (448x672)	1.88 (448x672)	1.76 (448x672)	1.04 (448x672)	0.22 (448x672)	0.48 (448x672)

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)
mvsanywhere	448x672	93.2 ms	70.8 ms	planned for 0.2.1	35.2 ms	153 s	–	931 / 1226	818 / 860	702 / 760	–

Fixed-size speed and peak memory, all models

model	mode	480x640	720x1280	1080x1920	12 MP 3000x4000
mvsanywhere	portable	107.5 ms / 969 / 1500	75.8 ms / 954 / 1398	77.4 ms / 1166 / 1756	124.0 ms / 3110 / 4342
mvsanywhere	kernels	40.6 ms / 769 / 1006	31.0 ms / 838 / 1068	32.5 ms / 1049 / 1414	57.1 ms / 2908 / 4006

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

On public datasets

640x512 crops, the 7 most co-visible views as sources; GT = depth rendered from the structured-light points. AbsRel / delta1 over GT pixels, after per-image median scaling unless the model takes the cameras (calibrated MVS: cameras, error in mm; PatchMatchNet also the depth range 425-935 mm, MVSAnywhere its own range from the camera baselines). Original = the original code named under Details, same checkpoint and protocol.

variant	dataset	original fp32	original fp16	portable	kernels	note
mvsanywhere	DTU scans 1, 9, 24, 33, 75, view 24 + 7 sources, AbsRel / delta1 % / MAE mm	0.0077 / 99.82 / 5.57	0.0110 / 99.84 / 7.85	0.0078 / 99.82 / 5.60	0.0077 / 99.82 / 5.59

Measured on NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 555d7b79, 2026-11-06.

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| <= 0.1 % and p99 <= 0.5 % of the reference's median |value| (absolute error of the output depth).

Usage

from fastforks import multiview

# non-commercial weights: run with FASTFORKS_NONCOMMERCIAL=1 in the environment
model = multiview.load("mvsanywhere")  # mode="auto": the preferred backend from recorded measurements
result = model(["ref.jpg", "src1.jpg", "src2.jpg"], intrinsics=K, cam_to_world=c2w)  # views[0] = reference
result.depth  # [1, H, W] float32 tensor; K [N, 3, 3] pixels, c2w [N, 4, 4] OpenCV camera-to-world
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: mvsanywhere.

Kernels backend

multiview.load("mvsanywhere", mode="kernels") runs our CUDA kernels (fastbb, igev, cutlass_conv, cudnn_hacc, mvsanywhere, 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.