SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.03307 · 2024

4D-Rotor Gaussian Splatting: Towards Efficient Novel View Synthesis for Dynamic Scenes

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 4 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2402.03307.json · badge: /api/ran/2402.03307.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
weify627/4D-Rotor-Gaussians canonical (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
lossfun_outer Ran
ran; our contract draft was wrong, not the code
weify627/4D-Rotor-Gaussians/nerfstudio/model_components/losses.py
code served (permissive licence) · get_code("f1e7d68796e89ab2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import Tensor

torch.manual_seed(0)

# Create small tensors for testing lossfun_outer
# Shape: (*batch, num_samples)
# batch = 2, num_samples = 4

# t: interval edges - shape (2, 5) - one more than weights
t = torch.linspace(0, 1, 5, dtype=torch.float32).unsqueeze(0).expand(2, -1)

# w: weights - shape (2, 4)
w = torch.ones(2, 4, dtype=torch.float32) * 0.5

# t_env: interval edges of upper bound envelope - shape (2, 6)
t_env = torch.linspace(0, 1.2, 6, dtype=torch.float32).unsqueeze(0).expand(2, -1)

# w_env: weights for envelope - shape (2, 5)
w_env = torch.ones(2, 5, dtype=torch.float32) * 0.8

ARGS = {
    't': t,
    'w': w,
    't_env': t_env,
    'w_env': w_env,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jaxtyping, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
merge_methods Ran
ran; our contract draft was wrong, not the code
weify627/4D-Rotor-Gaussians/nerfstudio/configs/method_configs.py
code served (permissive licence) · get_code("c64c9a3a66a1d0c8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from collections import OrderedDict

torch.manual_seed(0)

# Create sample methods - these are callables/functions
def method_1():
    return "result_1"

def method_2():
    return "result_2"

def method_3():
    return "result_3"

# Create existing methods and descriptions
existing_methods = OrderedDict({
    'method_1': method_1,
    'method_2': method_2,
})

existing_descriptions = OrderedDict({
    'method_1': 'Description for method 1',
    'method_2': 'Description for method 2',
})

# Create new methods and descriptions to merge in
new_methods = {
    'method_3': method_3,
    'method_2': lambda: "updated_result_2",  # This will overwrite method_2 if overwrite=True
}

new_descriptions = {
    'method_3': 'Description for method 3',
    'method_2': 'Updated description for method 2',
}

ARGS = {
    'methods': existing_methods,
    'method_descriptions': existing_descriptions,
    'new_methods': new_methods,
    'new_descriptions': new_descriptions,
    'overwrite': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
outer Ran
ran; our contract draft was wrong, not the code
weify627/4D-Rotor-Gaussians/nerfstudio/model_components/losses.py
code served (permissive licence) · get_code("de818de3a5e2a784")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from jaxtyping import Float
from torch import Tensor

torch.manual_seed(0)

# Small batch size of 2, different dimension sizes for clarity
batch_size = 2
num_samples_0 = 4
num_samples_1 = 3

# Create small tensors with the required shapes
t0_starts = torch.randn(batch_size, num_samples_0)
t0_ends = torch.randn(batch_size, num_samples_0)
t1_starts = torch.randn(batch_size, num_samples_1)
t1_ends = torch.randn(batch_size, num_samples_1)
y1 = torch.rand(batch_size, num_samples_1)

# Sort t0_starts, t0_ends, t1_starts, t1_ends to ensure valid interval semantics
# (starts < ends is expected for searchsorted to work correctly)
t0_starts, _ = torch.sort(t0_starts, dim=-1)
t0_ends = t0_starts + torch.rand(batch_size, num_samples_0) + 0.1
t1_starts, _ = torch.sort(t1_starts, dim=-1)
t1_ends = t1_starts + torch.rand(batch_size, num_samples_1) + 0.1

ARGS = {
    't0_starts': t0_starts,
    't0_ends': t0_ends,
    't1_starts': t1_starts,
    't1_ends': t1_ends,
    'y1': y1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jaxtyping, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ray_samples_to_sdist Ran
ran, honoured the contract we drafted
weify627/4D-Rotor-Gaussians/nerfstudio/model_components/losses.py
code served (permissive licence) · get_code("34a1673bf755a59b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 2) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

class RaySamples:
    """Stand-in class for ray_samples with spacing_starts and spacing_ends attributes"""
    def __init__(self, spacing_starts, spacing_ends):
        self.spacing_starts = spacing_starts
        self.spacing_ends = spacing_ends

# Create small tensors
# spacing_starts should have shape (..., 1, 1) where we extract [..., 0]
# spacing_ends should have shape (..., 1, 1) where we extract [..., -1:, 0]
batch_size = 2
spacing_starts = torch.randn(batch_size, 1, 1)
spacing_ends = torch.randn(batch_size, 4, 1)

ray_samples = RaySamples(spacing_starts=spacing_starts, spacing_ends=spacing_ends)

ARGS = {
    "ray_samples": ray_samples
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 2) — shape (2, 2) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
sort_methods Ran
ran; our contract draft was wrong, not the code
weify627/4D-Rotor-Gaussians/nerfstudio/configs/method_configs.py
code served (permissive licence) · get_code("06036a2535dc9b18")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from collections import OrderedDict

torch_manual_seed = 0

methods = OrderedDict([
    ("method_alpha", lambda x: x * 2),
    ("method_beta", lambda x: x + 1),
    ("method_gamma", lambda x: x - 1),
])

method_descriptions = OrderedDict([
    ("method_beta", "Description for beta"),
    ("method_alpha", "Description for alpha"),
    ("method_gamma", "Description for gamma"),
])

ARGS = {
    "methods": methods,
    "method_descriptions": method_descriptions,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 5 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 5 later) and 0 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
merge_methods 5 other papers
2 dated earlier, 3 later
1607.00653 (2016-07)
node2vec: Scalable Feature Learning for Networks
2103.12352 (2021-03)
iMAP: Implicit Mapping and Positioning in Real-Time
2403.09637 (2024-03)
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
and 2 more not named here
ray_samples_to_sdist 5 other papers
1 dated earlier, 4 later
2302.04264 (2023-02)
Nerfstudio: A Modular Framework for Neural Radiance Field Development
2403.09637 (2024-03)
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
2404.12378 (2024-04)
6Img-to-3D: Few-Image Large-Scale Outdoor Driving Scene Reconstruction
and 2 more not named here
sort_methods 5 other papers
2 dated earlier, 3 later
1607.00653 (2016-07)
node2vec: Scalable Feature Learning for Networks
2103.12352 (2021-03)
iMAP: Implicit Mapping and Positioning in Real-Time
2403.09637 (2024-03)
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
and 2 more not named here
lossfun_outer 3 other papers
3 later
2403.09637 (2024-03)
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
2409.12954 (2024-09)
GStex: Per-Primitive Texturing of 2D Gaussian Splatting for Decoupled Appearance and Geometry Modeling
2411.16816 (2024-11)
SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving
outer 3 other papers
3 later
2403.09637 (2024-03)
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
2409.12954 (2024-09)
GStex: Per-Primitive Texturing of 2D Gaussian Splatting for Decoupled Appearance and Geometry Modeling
2411.16816 (2024-11)
SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We consider the problem of novel-view synthesis (NVS) for dynamic scenes. Recent neural approaches have accomplished exceptional NVS results for static 3D scenes, but extensions to 4D time-varying scenes remain non-trivial. Prior efforts often encode dynamics by learning a canonical space plus implicit or explicit deformation fields, which struggle in challenging scenarios like sudden movements or generating high-fidelity renderings. In this paper, we introduce 4D Gaussian Splatting (4DRotorGS), a novel method that represents dynamic scenes with anisotropic 4D XYZT Gaussians, inspired by the success of 3D Gaussian Splatting in static scenes. We model dynamics at each timestamp by temporally slicing the 4D Gaussians, which naturally compose dynamic 3D Gaussians and can be seamlessly projected into images. As an explicit spatial-temporal representation, 4DRotorGS demonstrates powerful capabilities for modeling complicated dynamics and fine details--especially for scenes with abrupt motions. We further implement our temporal slicing and splatting techniques in a highly optimized CUDA acceleration framework, achieving real-time inference rendering speeds of up to 277 FPS on an RTX 3090 GPU and 583 FPS on an RTX 4090 GPU. Rigorous evaluations on scenes with diverse motions showcase the superior efficiency and effectiveness of 4DRotorGS, which consistently outperforms existing methods both quantitatively and qualitatively.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2402.03307")
get_code_for_paper("2402.03307")
have("2402.03307")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2402.03307.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2402.03307.svg)](https://syntology.ai/paper/2402.03307)

Connect an agent — have() is free.