MVSplat
MVSplat (Chen et al.): 3D Gaussian splats and depth from several posed, calibrated images in one pass; saves a 3DGS .ply. MIT.
experimental Fork of donydchen/mvsplat@01f9a28 · backends: portable, kernels
- Paper MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images ECCV 2024 · first released 21 Mar 2024
Checkpoint mvsplat-re10k
Paper 2024 Added in 0.2.0 Evaluated with 0.2.0
Use it
pip install fastforks Lean install with only what this model needs: pip install fastforks-core
from fastforks import multiview
model = multiview.load("mvsplat-re10k") # kernels on a supported GPU, else portable PyTorch
result = model(images)
result.depth # also: confidence, gaussians
result.backend # what ran: mode, device, dtype, kernels fastforks bench mvsplat-re10k Times the model on this machine. load(…, mode="portable") or mode="kernels" picks a backend; usage guide.
Speed and memory
Every backend on one GPU (RTX 5060 Ti 16 GB) and the same input, the authors’ code included. Kernels runs fp16. Portable runs fp16 on CUDA.
Input: synthetic, 336x518
| Backend | dtype | Time per call (ms) | vs original fp16 | Peak memory allocated / reserved | Load (s) | First call (s) |
|---|---|---|---|---|---|---|
| Original fp32 | fp32 | not measured | – | – | – | – |
| Original fp16the authors' mixed-precision option (float16); donydchen/mvsplat@01f9a28edb5e EncoderCostVolume, long side 256 | fp16 | 279 | – | 2,889 MiB/4,160 MiB | 0.2 | 0.56 |
| Portable fp32 | fp32 | not measured | – | – | – | – |
| Portablemeasured on 45b0e4f0 with the cudnn_hacc per-call workspace change of 600d8c2c | fp16 | 181 | 1.5× | 1,239 MiB/1,532 MiB | 0.1 | 1.02 |
| Kernelsmeasured on 45b0e4f0 with the cudnn_hacc per-call workspace change of 600d8c2c | fp16 | 172 | 1.6× | 1,239 MiB/1,532 MiB | 1.3 | 0.69 |
| Portable + torch.compile | 115 | 2.4× | 2,762 MiB/2,810 MiB | 102 (build) | – | |
| TensorRTplanned for 0.2.1 | fails | – | – | – | – |
Measured on RTX 5060 Ti 16 GB, torch 2.14.1+cu130, 2026-11-13. First call, 3 warm-up, peak reset, median of 3 rounds x median of 5. Ratios divide two
times of this table (same GPU, same input). Original fp16 is the authors’ own precision option, else torch.autocast. torch.compile and TensorRT: build time instead of load time. TensorRT memory (engine, not comparable) is the GPU memory the engine took, read from the driver because TensorRT allocates outside PyTorch, so it can't be compared with the PyTorch peaks of the other rows.
Other devices
| Device | Backend | dtype | ms / call | Peak memory | Output vs reference | Input |
|---|---|---|---|---|---|---|
| Apple M1 Pro 16 GB MPS, torch 2.14.1, 2026-11-13 | Original | fp32 | cannot run here | – | 336x518 | |
| Portable | fp16 | 1112 | 4,537 MiB process | matches | 336x518 | |
| Portable | fp32 | 1077 | 4,523 MiB process | matches | 336x518 | |
| Kernels | fp16 | not implemented | – | 336x518 | ||
CUDA memory is torch’s peak allocated / reserved; Apple has no such counter, so the Mac rows show the process footprint (weights, Python and torch included). “Output vs reference” compares the device’s output with the stored golden output under the card’s tolerance; – = not checked on that device. Kernels “not implemented” = no kernels for this model on that backend yet: it runs portable there.
Licence and weights
Commercial use allowed
- Only the RealEstate10K checkpoint is shipped (MIT code and weights); the ACID checkpoint has the same architecture.
Our reading of the licences, not legal advice.
Weights
re10k.ckpt 48.0 MB· hosted on Google Drivefastforks downloads and checks these files on first use (fastforks download mvsplat-re10k).
Accuracy
The test set through the authors’ code and through ours: same checkpoint, same protocol. The distance between the original’s fp32 and fp16 runs is what half precision alone costs; our runs should stay within it.
DTU scans 1
9, 24, 33, 75, view 24 + 7 sources, AbsRel / delta1 %
| Run | dtype | AbsRel ↓ | δ1.25 % ↑ | MAE mm ↓ | Δ AbsRel vs fp32 original | scans | ms / scan |
|---|---|---|---|---|---|---|---|
| Original fp32 | fp32 | 0.0438 | 97.31 | 31.17 | reference | 5 | 481 |
| Original fp16 | fp16 | 0.0439 | 97.30 | 31.21 | −0.0001 | 5 | 383 |
| Portable fp32 | fp32 | 0.0438 | 97.31 | 31.17 | ±0.0000 | 5 | 414 |
| Portable | fp16 | 0.0438 | 97.30 | 31.20 | ±0.0000 | 5 | 254 |
| Kernels | fp16 | 0.0438 | 97.30 | 31.20 | ±0.0000 | 5 | 243 |
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 authors’ code, same checkpoint and protocol. Measured on original fp32, original fp16: RTX 5060 Ti, torch 2.14.1+cu130, 2026-11-13; portable fp32: RTX 5060 Ti, torch 2.14.1+cu130, fastforks 30c7a8e9, 2026-11-17; portable, kernels: RTX 5060 Ti, torch 2.14.1+cu130, fastforks 28db82cb, 2026-11-13.
Closeness to the original
Beyond the test-set scores: the raw output of ours against the original’s fp32 output on the same input, pixel by pixel. 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).
| Check | Tolerance | Measured |
|---|---|---|
| CPU fp32, max |diff| | ≤ 0.002 | not recorded |
| GPU default dtypes, mean |diff| | ≤ 0.02 | 0.00157 |
| GPU default dtypes, p99 |diff| | ≤ 0.25 | 0.00705 |
Recorded for mvsplat-re10k on RTX 5060 Ti, torch 2.14.1+cu130, 2026-11-13, with these kernels: cudnn_hacc
Details
MVSplat (Chen et al. 2024): feed-forward 3D Gaussian splatting and depth from posed, calibrated images. A UniMatch CNN and multi-view swin transformer feed a plane-sweep cost volume (all other views), refined by 2D U-Nets with cross-view attention into every view's depth; a Gaussian head predicts one Gaussian per pixel (opacity, scale, rotation, degree-4 SH) in the given world frame (`result.gaussians`, `.save("scene.ply")`). Forward pass only: no renderer.
- Input / output
- 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.
- Input size
- any H x W >= 1, resized or padded inside the model; tested up to 3000x4000 (3,276 MiB allocated)
- Options
long_side=256>= 16· changes the output
- Outputs
depth,confidence,gaussians- CUDA kernels
cudnn_haccconvolutions with fp16 accumulation (cuDNN)
- Apple (MPS)
- fp32, fp16
Full reference, generated from the repository
MVSplat (Chen et al.): 3D Gaussian splats and depth from several posed, calibrated images in one pass; saves a 3DGS .ply. MIT.
Commercial use: yes.
| variant | input | backend | time (ms) | peak memory (MiB, allocated / reserved) |
|---|---|---|---|---|
mvsplat-re10k | synthetic (336x518) | original portable kernels portable + torch.compile TensorRT fp16 | 279 181 172 115 planned for 0.2.1 | 2889 / 4160 1239 / 1532 1239 / 1532 2762 / 2810 – |
Details
- Original code: https://github.com/donydchen/mvsplat at
01f9a28 - Code licence: MIT (permissive); files:
LICENSE - Weights licence: MIT
- Licence evidence: https://github.com/donydchen/mvsplat/blob/01f9a28edb5eb68416e7e63b01f8d90c3bdfbf01/LICENSE and https://github.com/donydchen/mvsplat/blob/01f9a28edb5eb68416e7e63b01f8d90c3bdfbf01/README.md and https://drive.google.com/drive/folders/14_E_5R6ojOWnLSrSVLVEMHnTiKsfddjU
- Training data: RealEstate10K (camera trajectories of YouTube real-estate videos) (RealEstate10K (YouTube videos, camera poses from ORB-SLAM; research use per its release), the backbone initialised from UniMatch (GMDepth, ScanNet).)
- Licence note: Only the RealEstate10K checkpoint is shipped (MIT code and weights); the ACID checkpoint has the same architecture.
- Options:
long_side= 256 (>= 16) - Kernel libraries:
cudnn_hacc - Variants:
mvsplat-re10k - Not in this release (planned for 0.2.1):
mvsplat-re10kTensorRT export: the TensorRT engine has not yet been validated against the portable model - Outputs: depth, confidence, gaussians; kind:
metric - Input size: any H x W >= 1, resized or padded inside the model
- Tested up to (portable, fp16, RTX 5060 Ti 16 GB):
mvsplat-re10k3000x4000 (3.2 GiB peak allocated, 4.4 reserved) mvsplat-re10k336x518, first call: portable + torch.compile 102 s (cold Inductor cache).mvsplat-re10koriginal 336x518: the authors' mixed-precision option (float16); donydchen/mvsplat@01f9a28edb5e EncoderCostVolume, long side 256mvsplat-re10kportable 336x518, kernels 336x518: measured on 45b0e4f0 with the cudnn_hacc per-call workspace change of 600d8c2cmvsplat-re10kload / first call: original 1.37 s / 2.817 s (336x518), portable 0.15 s / 1.019 s (336x518), kernels 1.27 s / 0.693 s (336x518)- Protocol, first call, 3 warm-up, peak reset, median of 3 rounds x median of 5
Measured
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 | portable fp32 | kernels | note |
|---|---|---|---|---|---|---|---|
mvsplat-re10k | DTU scans 1, 9, 24, 33, 75, view 24 + 7 sources, AbsRel / delta1 % / MAE mm | 0.0438 / 97.31 / 31.17 | 0.0439 / 97.30 / 31.21 | 0.0438 / 97.30 / 31.20 | 0.0438 / 97.31 / 31.17 | 0.0438 / 97.30 / 31.20 |
Measured on portable: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 28db82cb, 2026-11-13; portable32: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 30c7a8e9, 2026-11-17; kernels: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 28db82cb, 2026-11-13.
Parity
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).
| variant | kernels error vs original, mean / p99 | GPU, torch | kernels used |
|---|---|---|---|
mvsplat-re10k | 0.00157 / 0.00705 | RTX 5060 Ti, torch 2.14.1+cu130 | cudnn_hacc@cudnn92400 |
Example output
No example rendered for this checkpoint yet.