SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.04166 · 2023

BAA-NGP: Bundle-Adjusting Accelerated Neural Graphics Primitives

arXiv · PDF · Open in the Atlas

Code that ran

9 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; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 1 failure of our instrument; 1 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/2306.04166.json · badge: /api/ran/2306.04166.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
IntelLabs/baa-ngp canonical (classifier label)
classifier basis: census
9 ran · 1 no recorded run
FunctionStatusWhere it lives
construct_pose Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/pose_utils.py
code served (permissive licence) · get_code("a17204ec6041d356")
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 3, dtype family float, shape (2, 3, 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

torch.manual_seed(0)

# Case 1: Both R and t provided
R_both = torch.eye(3, dtype=torch.float32).unsqueeze(0).repeat(2, 1, 1)  # shape: (2, 3, 3)
t_both = torch.randn(2, 3, dtype=torch.float32)  # shape: (2, 3)

# Case 2: Only R provided
R_only = torch.eye(3, dtype=torch.float32).unsqueeze(0).repeat(2, 1, 1)  # shape: (2, 3, 3)

# Case 3: Only t provided
t_only = torch.randn(2, 3, dtype=torch.float32)  # shape: (2, 3)

# Use Case 1 (both provided) as the primary test case
ARGS = {
    "R": R_both,
    "t": t_both,
}
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 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
img2cam Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/camera_utils.py
code served (permissive licence) · get_code("37cfc5dcf0b446d5")
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 3, dtype family float, shape (2, 4, 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 as nn

torch.manual_seed(0)

# Create a small camera intrinsics matrix
# This needs to support .inverse() and .transpose() methods
# So it should be a torch.Tensor with shape (3, 3) or (batch, 3, 3)
cam_intr = torch.tensor([
    [500.0, 0.0, 320.0],
    [0.0, 500.0, 240.0],
    [0.0, 0.0, 1.0]
], dtype=torch.float32)

# Create 3D points in image space
# Shape should be compatible with matrix multiplication
# X @ cam_intr.inverse().transpose(-1, -2)
# If cam_intr is (3, 3), then X should be (..., 3)
X = torch.randn(2, 4, 3, dtype=torch.float32)

ARGS = {
    "X": X,
    "cam_intr": cam_intr
}
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 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 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
invert_pose Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/pose_utils.py
code served (permissive licence) · get_code("96881278efbd0013")
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 3, dtype family float, shape (2, 3, 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

torch.manual_seed(0)

# Create a valid pose tensor with shape (2, 3, 4)
# This represents 2 poses, each with a 3x3 rotation matrix and 3x1 translation vector
batch_size = 2
R = torch.eye(3).unsqueeze(0).repeat(batch_size, 1, 1).float()
t = torch.randn(batch_size, 3).float()
pose = torch.cat([R, t.unsqueeze(-1)], dim=-1)

# Test with use_inverse=False (default)
ARGS = {
    "pose": pose,
    "use_inverse": False
}
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 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
load_ckpt Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/utils.py
code served (permissive licence) · get_code("b5c6b7e13c76ed78")
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
import os
import tempfile
import shutil

torch.manual_seed(0)

# Create a minimal model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create a minimal optimizer
model1 = SimpleModel()
model2 = SimpleModel()
optimizer = torch.optim.SGD(model1.parameters(), lr=0.01)
scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=10)

# Create a temporary directory with a checkpoint
temp_dir = tempfile.mkdtemp()
model_dir = os.path.join(temp_dir, 'model')
os.makedirs(model_dir, exist_ok=True)

# Create and save a checkpoint
checkpoint = {
    'model1': model1.state_dict(),
    'model2': [model2.state_dict()],
    'optimizer': optimizer.state_dict(),
    'scheduler': scheduler.state_dict(),
    'epoch': 5,
    'iter': 100
}
checkpoint_path = os.path.join(temp_dir, 'model.ckpt')
torch.save(checkpoint, checkpoint_path)

# Create fresh instances for loading
model1_load = SimpleModel()
model2_load = SimpleModel()
optimizer_load = torch.optim.SGD(model1_load.parameters(), lr=0.01)
scheduler_load = torch.optim.lr_scheduler.StepLR(optimizer_load, step_size=10)

ARGS = {
    'save_dir': temp_dir,
    'models': {'model1': model1_load, 'model2': [model2_load]},
    'optimizers': {'optimizer': optimizer_load},
    'schedulers': {'scheduler': scheduler_load},
    'iteration': None
}
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
namedtuple_map Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/utils.py
code served (permissive licence) · get_code("944dbd0d0b063699")
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 Point. The image it ran in was not recorded. If your copy returns the same type 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 namedtuple
import torch

torch.manual_seed(0)

# Create a simple namedtuple type
Point = namedtuple('Point', ['x', 'y', 'z'])

# Define a function that will be applied to each element
def scale_tensor(x):
    return x * 2

# Create a namedtuple instance with tensors and None values
tup = Point(
    x=torch.tensor([1.0, 2.0, 3.0, 4.0]),
    y=None,
    z=torch.tensor([5.0, 6.0, 7.0, 8.0])
)

ARGS = {
    'fn': scale_tensor,
    'tup': tup
}
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 Point — a Point output has no array shape of its own; none is stated
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
procrustes_analysis Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/camera_utils.py
code served (permissive licence) · get_code("a8a39408d263db64")
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 EasyDict. The image it ran in was not recorded. If your copy returns the same type 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
from easydict import EasyDict as edict

torch.manual_seed(0)

# Create two sets of 3D point clouds
# X0 and X1 should be tensors of shape (N, D) where N is number of points and D is dimension
# The function performs centering, scaling, and SVD operations on these tensors

X0 = torch.randn(2, 4, dtype=torch.float32)  # 2 points, 4 dimensions
X1 = torch.randn(2, 4, dtype=torch.float32)  # 2 points, 4 dimensions

ARGS = {
    "X0": X0,
    "X1": X1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
easydict, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EasyDict — a EasyDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a EasyDict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
se3_to_SE3 Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/lie_utils.py
code served (permissive licence) · get_code("3a11e486616a44f4")
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 3, dtype family float, shape (2, 3, 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

torch.manual_seed(0)

# The function se3_to_SE3 takes a single parameter 'wu' which should be a tensor
# of shape (..., 6) where the last dimension is split into w (3,) and u (3,)
# w represents angular velocity (omega) and u represents linear velocity

# Create a small batch of se3 elements
# Shape: (batch_size, 6) where 6 = 3 (angular) + 3 (linear)
wu = torch.randn(2, 6, dtype=torch.float32)

ARGS = {
    "wu": wu
}
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 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
so3_to_SO3 Ran
ran on a synthesized input
IntelLabs/baa-ngp/baangp/lie_utils.py
code served (permissive licence) · get_code("41ec182be4f256d1")
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 3, dtype family float, shape (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

torch.manual_seed(0)

# w is used as:
# - w.unbind(dim=-1) - unbind along last dimension
# - w.norm(dim=-1) - compute norm along last dimension
# - w.device - get device
# So w should be a tensor with shape (..., 3) since SO(3) is 3D

# Create a small batch of 3D vectors (angular velocities)
# Shape: (2, 3) - batch of 2, dimension 3
w = torch.randn(2, 3, dtype=torch.float32)

ARGS = {
    "w": w
}
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 3, dtype family float, shape (2, 3, 3) — shape (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
to_hom Ran
ran; our contract draft was wrong, not the code
IntelLabs/baa-ngp/baangp/pose_utils.py
code served (permissive licence) · get_code("d8f116f87e3af2fa")
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 (8, 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)

# X should be a tensor of shape [H*W, 2]
# Using H*W = 8 (e.g., 4*2), so shape [8, 2]
X = torch.randn(8, 2)

ARGS = {
    "X": X
}
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 (8, 3) — shape (8, 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
SO3_to_so3 Not yet run
unverified (harvested, no recorded run)
IntelLabs/baa-ngp/baangp/lie_utils.py
code served (permissive licence) · get_code("8d103cd888cb0319")
How it ran

no invocation recorded.

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, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 9 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
to_hom 4 other papers
3 dated earlier, 1 later
2010.10505 (2020-10)
SDF-SRN: Learning Signed Distance 3D Object Reconstruction from Static Images
2104.06405 (2021-04)
BARF : Bundle-Adjusting Neural Radiance Fields
2304.12294 (2023-04)
Explicit Correspondence Matching for Generalizable Neural Radiance Fields
and 1 more not named here

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

Implicit neural representations have become pivotal in robotic perception, enabling robots to comprehend 3D environments from 2D images. Given a set of camera poses and associated images, the models can be trained to synthesize novel, unseen views. To successfully navigate and interact in dynamic settings, robots require the understanding of their spatial surroundings driven by unassisted reconstruction of 3D scenes and camera poses from real-time video footage. Existing approaches like COLMAP and bundle-adjusting neural radiance field methods take hours to days to process due to the high computational demands of feature matching, dense point sampling, and training of a multi-layer perceptron structure with a large number of parameters. To address these challenges, we propose a framework called bundle-adjusting accelerated neural graphics primitives (BAA-NGP) which leverages accelerated sampling and hash encoding to expedite automatic pose refinement/estimation and 3D scene reconstruction. Experimental results demonstrate 10 to 20 x speed improvement compared to other bundle-adjusting neural radiance field methods without sacrificing the quality of pose estimation. The github repository can be found here https://github.com/IntelLabs/baa-ngp.

For agents

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

get_harvested_code_for_paper("2306.04166")
get_code_for_paper("2306.04166")
have("2306.04166")

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

curl https://syntology.ai/api/ran/2306.04166.json

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

[![Syntology run record](https://syntology.ai/api/ran/2306.04166.svg)](https://syntology.ai/paper/2306.04166)

Connect an agent — have() is free.