SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.04073 · 2021

Direct-PoseNet: Absolute Pose Regression with Photometric Consistency

arXiv · PDF · Open in the Atlas

Code that ran

20 of 22 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, 2 violated it and 0 have no contract verdict; 16 failures of our instrument; 2 no recorded run), measured 2026-11-21; 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/2104.04073.json · badge: /api/ran/2104.04073.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
ActiveVisionLab/direct-posenet canonical (classifier label)
classifier basis: census
20 ran · 2 no recorded run
FunctionStatusWhere it lives
PoseLoss Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/direct_pose_model.py
code served (permissive licence) · get_code("019452c6c4769152")
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)

class ArgsConfig:
    def __init__(self):
        self.batch_size = 2

args = ArgsConfig()

device = torch.device('cpu')

pose_ = torch.randn(2, 12, device=device)

pose = torch.randn(2, 12, device=device)

ARGS = {
    'args': args,
    'pose_': pose_,
    'pose': pose,
    'device': device
}
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
batchify Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/run_nerf.py
code served (permissive licence) · get_code("fd57e365830d09de")
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 function. 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

torch.manual_seed(0)

# Define a simple function that processes batches
def simple_fn(inputs):
    """A simple function that takes a tensor and returns a processed tensor."""
    return inputs * 2

# Create a small input tensor (batch_size=4, features=3)
input_tensor = torch.randn(4, 3)

# Define chunk size for batching
chunk_size = 2

ARGS = {
    "fn": simple_fn,
    "chunk": chunk_size
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
camera_frustum_initializer Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/dataset_loaders/load_7Scenes.py
code served (permissive licence) · get_code("c9625d568955ffac")
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. 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 numpy as np

class Args:
    def __init__(self, near_far):
        self.near_far = near_far

ARGS = {
    "args": Args(near_far=[0.1, 5.0])
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
filter_hook Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/pose_model.py
code served (permissive licence) · get_code("c7bea2e25b534df3")
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 3, 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)

# The function filter_hook is a backward hook that processes input gradients
# m: the module (not directly used in the function body)
# g_in: tuple of input gradients to the module
# g_out: tuple of output gradients from the module (not used in the function body)

# Create a simple module
m = nn.Linear(4, 3)

# Create gradient tensors for g_in
# The function iterates over g_in, clones each tensor, replaces NaN values with 0
# We need to create tensors that could contain NaNs
g_tensor_1 = torch.randn(2, 4)
g_tensor_2 = torch.randn(2, 4)

# Introduce some NaN values to make the filtering meaningful
g_tensor_1[0, 0] = float('nan')
g_tensor_2[1, 2] = float('nan')

# g_in should be a tuple of gradients
g_in = (g_tensor_1, g_tensor_2)

# g_out is not used in the function body, but it's a parameter
# Typically it's a tuple of output gradients
g_out = (torch.randn(2, 3),)

ARGS = {
    "m": m,
    "g_in": g_in,
    "g_out": g_out
}
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 3, 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
get_embedder Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/run_nerf_helpers.py
code served (permissive licence) · get_code("737c403b9968484c")
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 numpy as np

torch.manual_seed(0)

ARGS = {
    'multires': 8,
    'i': 0,
    'reduce_mode': -1,
    'epochToMaxFreq': 100
}
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 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
get_rays Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/run_nerf_helpers.py
code served (permissive licence) · get_code("c03007b6cafd9690")
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 4, 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 numpy as np

torch.manual_seed(0)

# Parameters for get_rays function
H = 2  # image height
W = 4  # image width
focal = 500.0  # focal length

# camera-to-world transformation matrix (4x4)
c2w = torch.tensor([
    [1.0, 0.0, 0.0, 0.0],
    [0.0, 1.0, 0.0, 0.0],
    [0.0, 0.0, 1.0, 5.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=torch.float32)

ARGS = {
    "H": H,
    "W": W,
    "focal": focal,
    "c2w": c2w
}
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 tuple, ndim 4, 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
get_rays Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/dataset_loaders/frustum/frustum_util.py
code served (permissive licence) · get_code("3627b1373cfa4e33")
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 4, 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 numpy as np

torch.manual_seed(0)

# Parameters for get_rays function
H = 2  # image height
W = 4  # image width
focal = 500.0  # focal length

# Camera-to-world transformation matrix
# Shape: (4, 4) homogeneous transformation matrix
c2w = torch.tensor([
    [1.0, 0.0, 0.0, 0.0],
    [0.0, 1.0, 0.0, 0.0],
    [0.0, 0.0, 1.0, -3.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=torch.float32)

ARGS = {
    'H': H,
    'W': W,
    'focal': focal,
    'c2w': c2w
}
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 tuple, ndim 4, 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
get_rays_np Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/script/run_nerf_helpers.py
code served (permissive licence) · get_code("929e526ea6fc4975")
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 4, 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 numpy as np

# Camera intrinsics
H = 64  # image height
W = 128  # image width
focal = 50.0  # focal length

# Camera-to-world transformation matrix (4x4)
# Top 3 rows are used by the function
c2w = np.array([
    [1.0, 0.0, 0.0, 0.0],
    [0.0, 1.0, 0.0, 0.0],
    [0.0, 0.0, 1.0, 5.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=np.float32)

ARGS = {
    'H': H,
    'W': W,
    'focal': focal,
    'c2w': c2w
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 4, 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
initK Ran
ran, honoured the contract we drafted
ActiveVisionLab/direct-posenet/dataset_loaders/load_7Scenes.py
code served (permissive licence) · get_code("13be307ef0c24c0e")
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 (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 numpy as np

ARGS = {
    "f": 500.0,
    "cx": 320.0,
    "cy": 240.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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 (3, 3) — shape (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
is_inside_frustum Ran
ran, violated the contract we drafted
ActiveVisionLab/direct-posenet/dataset_loaders/load_7Scenes.py
code served (permissive licence) · get_code("f869aad3a90d4156")
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 1, dtype family float, shape (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
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - p: appears to be indexable with [0] and [1], likely a tensor or array
# - x_res: a scalar representing x resolution (compared with p[0])
# - y_res: a scalar representing y resolution (compared with p[1])
# The function checks if point p is within bounds (0, x_res) and (0, y_res)

# Create a point tensor with 2 coordinates
p = torch.tensor([[2.5, 3.5], [0.5, 4.5]], dtype=torch.float32)  # shape (2, 2)

# Resolution values
x_res = 10
y_res = 8

ARGS = {
    "p": p,
    "x_res": x_res,
    "y_res": y_res
}
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 1, dtype family float, shape (2,) — shape (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
ndc_rays Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/dataset_loaders/frustum/frustum_util.py
pointer only (licence: MIT); the stored copy get_code serves, from sunset1995/directvoxgo, is pointer only (licence: NOASSERTION) · get_code("4366d225ac2fe16d")
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 sunset1995/directvoxgo, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

normalize Ran
ran, violated the contract we drafted
ActiveVisionLab/direct-posenet/dataset_loaders/load_llff.py
code served (permissive licence) · get_code("1d6f8390d799f055")
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 1, dtype family float, shape (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 numpy as np

# The normalize function takes a single parameter 'x' which is expected to be
# a numpy array (or array-like). It divides x by its L2 norm.
# We need to provide a small numpy array as input.

x = np.array([3.0, 4.0])  # Simple 2D vector with norm=5, will normalize to [0.6, 0.8]

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (2,) — shape (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
prepare_data Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/script/prepare_data.py
code served (permissive licence) · get_code("6e2a649fb0ac265e")
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 numpy as np
from torch.utils.data import TensorDataset, DataLoader

torch.manual_seed(0)

# Create a small args object with batch_size attribute
class Args:
    def __init__(self):
        self.batch_size = 2

args = Args()

# Create small dummy data
# images: shape (N, H, W, C) - we'll use 6 images, 4x3 spatial, 3 channels
images = np.random.randn(6, 4, 3, 3).astype(np.float32)

# poses_train: shape (N, 7) - typical for pose (e.g., 3D position + 4D quaternion)
poses_train = np.random.randn(6, 7).astype(np.float32)

# i_split: tuple of three index arrays for train/val/test split
i_train = np.array([0, 1, 2])
i_val = np.array([3, 4])
i_test = np.array([5])
i_split = (i_train, i_val, i_test)

ARGS = {
    "args": args,
    "images": images,
    "poses_train": poses_train,
    "i_split": i_split
}
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 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
preprocess_data Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/pose_model.py
code served (permissive licence) · get_code("1d4f466dd0b12111")
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 4, dtype family float, shape (2, 3, 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create a small input tensor: batch size 2, 3 channels, 4x4 spatial dimensions
inputs = torch.randn(2, 3, 4, 4)

# Set device to CPU
device = torch.device('cpu')

ARGS = {
    'inputs': inputs,
    'device': device
}
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 4, dtype family float, shape (2, 3, 4, 4) — shape (2, 3, 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
process_poses_rotmat Ran
ran, honoured the contract we drafted
ActiveVisionLab/direct-posenet/dataset_loaders/seven_scenes.py
code served (permissive licence) · get_code("ecb54915a44f88b3")
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, 12) 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

torch.manual_seed(0)

# Construct arguments for process_poses_rotmat
# poses_in: N x 12 tensor of raw poses
poses_in = torch.randn(2, 12)

# mean_t: mean translation vector (typically 3D)
mean_t = torch.randn(3)

# std_t: standard deviation for translation (typically 3D)
std_t = torch.randn(3).abs() + 0.1  # ensure positive

# align_R: alignment rotation matrix (typically 3x3)
align_R = torch.eye(3)

# align_t: alignment translation vector (typically 3D)
align_t = torch.randn(3)

# align_s: alignment scale (scalar)
align_s = torch.tensor(1.0)

ARGS = {
    "poses_in": poses_in,
    "mean_t": mean_t,
    "std_t": std_t,
    "align_R": align_R,
    "align_t": align_t,
    "align_s": align_s,
}
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, 12) — shape (2, 12) 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
ptstocam Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/dataset_loaders/load_llff.py
code served (permissive licence) · get_code("36f5de7a0aae4a9b")
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 (2, 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 numpy as np

# Construct valid arguments for ptstocam function
# The function expects:
# - pts: points array that will be used in (pts - c2w[:3, 3])
# - c2w: camera-to-world transformation matrix with shape (4, 4) or similar
#        where c2w[:3, :3] is the rotation part and c2w[:3, 3] is the translation

# Create a c2w matrix (4x4 camera-to-world transformation)
c2w = np.array([
    [1.0, 0.0, 0.0, 0.5],
    [0.0, 1.0, 0.0, 1.0],
    [0.0, 0.0, 1.0, 2.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=np.float32)

# Create points array with shape (2, 3) - 2 points with 3 coordinates each
pts = np.array([
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0]
], dtype=np.float32)

ARGS = {
    "pts": pts,
    "c2w": c2w
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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 (2, 3) — shape (2, 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
qlog Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/dataset_loaders/seven_scenes.py
code served (permissive licence) · get_code("667f7d75349dad79")
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 1, dtype family float, shape (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 numpy as np

# The function qlog takes a quaternion q of shape (4,) and applies a logarithm map
# to return a vector of shape (3,)
# 
# Looking at the function body:
# - q is indexed with q[0], q[1:], and np.linalg.norm(q[1:])
# - It's a 1D numpy array of length 4

# Create a valid quaternion input
# A quaternion is typically [w, x, y, z] where w is the scalar part
# We'll create one with a non-zero imaginary part to avoid the zero-check path

q_value = np.array([0.7071, 0.7071, 0.0, 0.0], dtype=np.float64)

ARGS = {
    "q": q_value
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (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
quaternion_angular_error Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/script/pose_model.py
code served (permissive licence) · get_code("d93282ce062190e1")
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 float32, 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 numpy as np

# Create two sample quaternions (normalized 4D vectors)
# Quaternions are typically normalized, so we'll create unit-norm vectors
q1 = np.array([0.0, 0.0, 0.0, 1.0], dtype=np.float32)
q2 = np.array([0.7071068, 0.0, 0.0, 0.7071068], dtype=np.float32)

# Normalize to ensure they're proper unit quaternions
q1 = q1 / np.linalg.norm(q1)
q2 = q2 / np.linalg.norm(q2)

ARGS = {
    "q1": q1,
    "q2": q2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float32, 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
run_network Ran
ran; our contract draft was wrong, not the code
ActiveVisionLab/direct-posenet/script/run_nerf.py
code served (permissive licence) · get_code("8f235fb889b93174")
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 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
import torch.nn as nn

torch.manual_seed(0)

# Create a simple embedding function
def embed_fn(x):
    # x shape: [N, 3] -> output [N, 8]
    return torch.randn(x.shape[0], 8)

# Create a simple view direction embedding function
def embeddirs_fn(x):
    # x shape: [N, 3] -> output [N, 4]
    return torch.randn(x.shape[0], 4)

# Create a simple network that takes embedded inputs and outputs values
class SimpleNetwork(nn.Module):
    def __init__(self):
        super().__init__()
        # Input: 8 (from embed_fn) + 4 (from embeddirs_fn) = 12
        self.linear = nn.Linear(12, 3)
    
    def forward(self, x):
        return self.linear(x)

fn = SimpleNetwork()

# Create inputs tensor: [batch=2, spatial=3, feature=3]
inputs = torch.randn(2, 3, 3)

# Create viewdirs tensor: [batch=2, feature=3]
viewdirs = torch.randn(2, 3)

ARGS = {
    "inputs": inputs,
    "viewdirs": viewdirs,
    "fn": fn,
    "embed_fn": embed_fn,
    "embeddirs_fn": embeddirs_fn,
    "netchunk": 1024 * 64
}
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 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
viewmatrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ActiveVisionLab/direct-posenet/dataset_loaders/load_llff.py
code served (permissive licence) · get_code("aa52acee18f54d38")
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 (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 numpy as np

# Construct arguments for viewmatrix function
# The function expects:
# - z: a vector (used in normalize and cross product)
# - up: a vector (used in cross product)
# - pos: a vector (stacked into result)

# Create small 3D vectors as numpy arrays
z = np.array([0.0, 0.0, 1.0])
up = np.array([0.0, 1.0, 0.0])
pos = np.array([1.0, 2.0, 3.0])

ARGS = {
    "z": z,
    "up": up,
    "pos": pos
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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 (3, 4) — shape (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
rgb_loss Not yet run
unverified (harvested, no recorded run)
ActiveVisionLab/direct-posenet/script/direct_pose_model.py
code served (permissive licence) · get_code("26d5443e570cf251")
How it ran

no invocation recorded.

run_network_DNeRF Not yet run
unverified (harvested, no recorded run)
ActiveVisionLab/direct-posenet/script/run_nerf.py
code served (permissive licence) · get_code("5bd6e1cb35d8420a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 16 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 1 the same month, 16 later) and 5 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
normalize 19 other papers
2 dated earlier, 17 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2007.02442 (2020-07)
GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 16 more not named here
batchify 17 other papers
2 dated earlier, 1 the same month, 14 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2102.07064 (2021-02)
NeRF--: Neural Radiance Fields Without Known Camera Parameters
2104.04532 (2021-04)
Neural RGB-D Surface Reconstruction
and 14 more not named here
viewmatrix 17 other papers
2 dated earlier, 15 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2007.02442 (2020-07)
GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 14 more not named here
run_network 11 other papers
2 dated earlier, 9 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2102.07064 (2021-02)
NeRF--: Neural Radiance Fields Without Known Camera Parameters
2107.02791 (2021-07)
Depth-supervised NeRF: Fewer Views and Faster Training for Free
and 8 more not named here
ndc_rays 10 other papers
10 later
2111.11215 (2021-11)
Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction
2202.05263 (2022-02)
Block-NeRF: Scalable Large Scene Neural View Synthesis
2206.05085 (2022-06)
Improved Direct Voxel Grid Optimization for Radiance Fields Reconstruction
and 7 more not named here
ptstocam 10 other papers
2 dated earlier, 8 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2007.02442 (2020-07)
GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 7 more not named here
camera_frustum_initializer 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
filter_hook 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
initK 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
is_inside_frustum 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
preprocess_data 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
process_poses_rotmat 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
qlog 2 other papers
2 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching
2303.10087 (2023-03)
Neural Refinement for Absolute Pose Regression with Feature Synthesis
get_rays 1 other paper
1 later
2209.09050 (2022-09)
Loc-NeRF: Monte Carlo Localization using Neural Radiance Fields
get_rays_np 1 other paper
1 later
2209.09050 (2022-09)
Loc-NeRF: Monte Carlo Localization using Neural Radiance Fields
run_network_DNeRF 1 other paper
1 later
2204.00559 (2022-04)
DFNet: Enhance Absolute Pose Regression with Direct Feature Matching

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 2026-11-21.

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 a relocalization pipeline, which combines an absolute pose regression (APR) network with a novel view synthesis based direct matching module, offering superior accuracy while maintaining low inference time. Our contribution is twofold: i) we design a direct matching module that supplies a photometric supervision signal to refine the pose regression network via differentiable rendering; ii) we modify the rotation representation from the classical quaternion to SO(3) in pose regression, removing the need for balancing rotation and translation loss terms. As a result, our network Direct-PoseNet achieves state-of-the-art performance among all other single-image APR methods on the 7-Scenes benchmark and the LLFF dataset.

For agents

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

get_harvested_code_for_paper("2104.04073")
get_code_for_paper("2104.04073")
have("2104.04073")

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

curl https://syntology.ai/api/ran/2104.04073.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.04073.svg)](https://syntology.ai/paper/2104.04073)

Connect an agent — have() is free.