SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2312.02134 · 2023

GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 8 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/2312.02134.json · badge: /api/ran/2312.02134.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
aipixel/gaussianavatar canonical (classifier label)
classifier basis: census
10 ran · 0 no recorded run
FunctionStatusWhere it lives
find_joint_kin_chain Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aipixel/gaussianavatar/submodules/smplx/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from wjrzm/VP-MoCap, is pointer only (licence: NONE) · get_code("3b5f8cde1dc64833")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from wjrzm/VP-MoCap, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

gaussian Ran
ran, honoured the contract we drafted
aipixel/gaussianavatar/utils/loss_utils.py
code served (permissive licence) · get_code("c56b7ef16f309a45")
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) and returned Tensor, ndim 1, dtype family float, shape (5,) 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 math import exp

torch.manual_seed(0)

ARGS = {
    "window_size": 5,
    "sigma": 1.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 1, dtype family float, shape (5,) — shape (5,) 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
gen_transf_mtx_from_vtransf Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aipixel/gaussianavatar/model/modules.py
code served (permissive licence) · get_code("412dda33006a6e27")
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 5, dtype family float, shape (2, 4, 4, 3, 3) 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)

# Small dimensions: batch=2, uv_size=4, num_verts=5, transf_size=3
batch_size = 2
num_verts = 5
uv_size = 4
transf_dim = 3

# vtransf: [batch, #verts, 3, 3] - per-vertex rotation matrices
vtransf = torch.randn(batch_size, num_verts, transf_dim, transf_dim)

# bary_coords: [uv_size, uv_size, 3] - barycentric coordinates
# Each pixel has 3 barycentric coords that sum to 1
bary_coords = torch.rand(uv_size, uv_size, 3)
bary_coords = bary_coords / bary_coords.sum(dim=-1, keepdim=True)  # normalize to sum to 1

# faces: [uv_size, uv_size, 3] - vertex indices for each pixel's triangle
# Each pixel references 3 vertex indices (0 to num_verts-1)
faces = torch.randint(0, num_verts, (uv_size, uv_size, 3))

# scaling parameter
scaling = 2.0

ARGS = {
    'vtransf': vtransf,
    'bary_coords': bary_coords,
    'faces': faces,
    'scaling': scaling
}
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 5, dtype family float, shape (2, 4, 4, 3, 3) — shape (2, 4, 4, 3, 3) 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
gen_transf_mtx_full_uv Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aipixel/gaussianavatar/model/modules.py
code served (permissive licence) · get_code("c661071dca6a8c5b")
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 5, dtype family float, shape (2, 2, 2, 3, 3) 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
import torch.nn.functional as F

torch.manual_seed(0)

# verts: [batch, Nverts, 3]
# batch=2, Nverts=4
verts = torch.randn(2, 4, 3)

# faces: [uv_size, uv_size, 3]
# uv_size=2, each element is an index into verts
uv_size = 2
faces = torch.tensor([
    [[0, 1, 2], [1, 2, 3]],
    [[1, 2, 3], [0, 1, 2]]
], dtype=torch.long)

ARGS = {
    "verts": verts,
    "faces": faces,
}
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 5, dtype family float, shape (2, 2, 2, 3, 3) — shape (2, 2, 2, 3, 3) 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
get_embedder Ran
ran; our contract draft was wrong, not the code
aipixel/gaussianavatar/model/modules.py
code served (permissive licence) · get_code("94d79c473e827470")
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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'multires': 4,
    'i': 0,
    'input_dims': 3
}
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
l1_loss_w Ran
ran; our contract draft was wrong, not the code
aipixel/gaussianavatar/utils/loss_utils.py
code served (permissive licence) · get_code("8c3f087d9f30037e")
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 0, dtype family float, shape () 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)

# The function takes two parameters: network_output and gt
# Both are used in torch.abs(network_output - gt).mean()
# So they need to be tensors that can be subtracted and have .mean() called on them

network_output = torch.randn(2, 4)
gt = torch.randn(2, 4)

ARGS = {
    "network_output": network_output,
    "gt": gt,
}
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 0, dtype family float, shape () — shape () 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
l2_loss Ran
ran, honoured the contract we drafted
aipixel/gaussianavatar/utils/loss_utils.py
code served (permissive licence) · get_code("8c3b0f873ba11813")
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 0, dtype family float, shape () 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
import torch.nn as nn

torch.manual_seed(0)

# Create small tensors for network output and ground truth
# Both should have the same shape for the loss computation
# Using batch size 2, feature dimension 4
network_output = torch.randn(2, 4)
gt = torch.randn(2, 4)

ARGS = {
    "network_output": network_output,
    "gt": gt,
}
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 0, dtype family float, shape () — shape () 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
rotate_camera_by_frame_idx Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aipixel/gaussianavatar/scene/dataset_mono.py
code served (permissive licence) · get_code("59b04ecdb5519178")
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 ndarray, ndim 2, dtype family float, shape (4, 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 numpy as np
import cv2

# Create extrinsics as a 4x4 matrix (homogeneous transformation matrix)
# Starting with identity and slight perturbations to make it realistic
extrinsics = np.array([
    [0.9, -0.1, 0.0, 1.0],
    [0.1, 0.9, 0.0, 2.0],
    [0.0, 0.0, 1.0, 3.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=np.float32)

# Frame index - an integer
frame_idx = 5

# Translation vector - optional, can be None or a 3-element array
trans = np.array([0.5, 0.2, 0.1], dtype=np.float32)

# Rotation axis - one of 'x', 'y', 'z'
rotate_axis = 'y'

# Period - number of frames for one full rotation
period = 196

# inv_angle - boolean flag for clockwise/counterclockwise
inv_angle = False

ARGS = {
    'extrinsics': extrinsics,
    'frame_idx': frame_idx,
    'trans': trans,
    'rotate_axis': rotate_axis,
    'period': period,
    'inv_angle': inv_angle
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (4, 4) — shape (4, 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
to_np Ran
ran; our contract draft was wrong, not the code
aipixel/gaussianavatar/submodules/smplx/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from wjrzm/VP-MoCap, is pointer only (licence: NONE) · get_code("43ebe3b786b2d59b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from wjrzm/VP-MoCap, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

to_tensor Ran
ran; our contract draft was wrong, not the code
aipixel/gaussianavatar/submodules/smplx/utils.py
code served (permissive licence) · get_code("9ad32763fbba1d53")
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
import numpy as np
from typing import NewType, Union, Optional

torch.manual_seed(0)

Tensor = NewType('Tensor', torch.Tensor)
Array = NewType('Array', np.ndarray)

# Test with a numpy array
numpy_array = np.array([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]], dtype=np.float32)

# Test with a torch tensor
torch_tensor = torch.randn(2, 4)

# Define ARGS with all parameters
ARGS = {
    "array": numpy_array,
    "dtype": torch.float32
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 3 the same month, 6 later) and 3 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
gaussian 160 other papers
53 dated earlier, 8 the same month, 97 later, 2 same year, order not determinable · 7 only as an assembly brick
1812.11941 (2018-12)
High Quality Monocular Depth Estimation via Transfer Learning
1904.01538 (2019-04)
Spatial Attentive Single-Image Deraining with a High Quality Real Rain Dataset
1907.11628 (2019-07)
Unsupervised Learning for Optical Flow Estimation Using Pyramid Convolution LSTM
and 157 more not named here
l2_loss 90 other papers
10 dated earlier, 8 the same month, 72 later
2308.04079 (2023-08)
3D Gaussian Splatting for Real-Time Radiance Field Rendering
2309.13101 (2023-09)
Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstruction
2310.08528 (2023-10)
4D Gaussian Splatting for Real-Time Dynamic Scene Rendering
and 87 more not named here
to_np 9 other papers
4 dated earlier, 1 the same month, 4 later · 1 only as an assembly brick
2104.03953 (2021-04)
SNARF: Differentiable Forward Skinning for Animating Non-Rigid Neural Implicit Shapes
2112.07471 (2021-12)
I M Avatar: Implicit Morphable Head Avatars from Videos
2203.12575 (2022-03)
NeuMan: Neural Human Radiance Field from a Single Video
and 6 more not named here
rotate_camera_by_frame_idx 2 other papers
1 dated earlier, 1 later
2201.04127 (2022-01)
HumanNeRF: Free-viewpoint Rendering of Moving People from Monocular Video
2404.07991 (2024-04)
GoMAvatar: Efficient Animatable Human Modeling from Monocular Video Using Gaussians-on-Mesh
to_tensor 2 other papers
2 later
2403.17610 (2024-03)
MMVP: A Multimodal MoCap Dataset with Vision and Pressure Sensors
2411.08128 (2024-11)
CameraHMR: Aligning People with Perspective
find_joint_kin_chain 1 other paper
1 later
2403.17610 (2024-03)
MMVP: A Multimodal MoCap Dataset with Vision and Pressure Sensors
get_embedder 1 other paper
1 dated earlier
2104.03313 (2021-04)
SCANimate: Weakly Supervised Learning of Skinned Clothed Avatar Networks

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 present GaussianAvatar, an efficient approach to creating realistic human avatars with dynamic 3D appearances from a single video. We start by introducing animatable 3D Gaussians to explicitly represent humans in various poses and clothing styles. Such an explicit and animatable representation can fuse 3D appearances more efficiently and consistently from 2D observations. Our representation is further augmented with dynamic properties to support pose-dependent appearance modeling, where a dynamic appearance network along with an optimizable feature tensor is designed to learn the motion-to-appearance mapping. Moreover, by leveraging the differentiable motion condition, our method enables a joint optimization of motions and appearances during avatar modeling, which helps to tackle the long-standing issue of inaccurate motion estimation in monocular settings. The efficacy of GaussianAvatar is validated on both the public dataset and our collected dataset, demonstrating its superior performances in terms of appearance quality and rendering efficiency.

For agents

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

get_harvested_code_for_paper("2312.02134")
get_code_for_paper("2312.02134")
have("2312.02134")

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

curl https://syntology.ai/api/ran/2312.02134.json

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

[![Syntology run record](https://syntology.ai/api/ran/2312.02134.svg)](https://syntology.ai/paper/2312.02134)

Connect an agent — have() is free.