SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2008.05742 · 2020

SkeletonNet: A Topology-Preserving Solution for Learning Mesh Reconstruction of Object Surfaces from RGB Images

arXiv · PDF · Open in the Atlas

Code that ran

16 of 23 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 14 have no contract verdict; 1 failure of our instrument; 7 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/2008.05742.json · badge: /api/ran/2008.05742.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
tangjiapeng/SkeletonNet canonical (classifier label)
classifier basis: census
16 ran · 7 no recorded run
FunctionStatusWhere it lives
compute_geodesic_distance_from_two_matrices Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/models/posenet_old.py
code served (permissive licence) · get_code("263a4c266b7dd7be")
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 EagerTensor, 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 tensorflow as tf

# The function expects two matrices m1 and m2 that will be multiplied via matmul
# m1 shape: [batch, 3, 3] (rotation matrices)
# m2 shape: [batch, 3, 3] (rotation matrices)
# The function computes matmul(m1, transpose(m2)) and extracts trace-based geodesic distance

batch_size = 2

# Create small valid rotation-like matrices (batch_size x 3 x 3)
# Using identity-like matrices for simplicity
m1_data = tf.constant([
    [[1.0, 0.0, 0.0],
     [0.0, 1.0, 0.0],
     [0.0, 0.0, 1.0]],
    [[0.9, 0.1, 0.0],
     [-0.1, 0.9, 0.0],
     [0.0, 0.0, 1.0]]
], dtype=tf.float32)

m2_data = tf.constant([
    [[1.0, 0.0, 0.0],
     [0.0, 1.0, 0.0],
     [0.0, 0.0, 1.0]],
    [[0.8, 0.2, 0.0],
     [-0.2, 0.8, 0.0],
     [0.0, 0.0, 1.0]]
], dtype=tf.float32)

ARGS = {
    "m1": m1_data,
    "m2": m2_data,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
compute_rotation_matrix_from_ortho6d Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/models/posenet.py
code served (permissive licence) · get_code("ff62c3bc33efc141")
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 EagerTensor, 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 tensorflow as tf

# Create a small batch of ortho6d representations
# poses should have shape [batch_size, 6]
# where first 3 columns are x_raw and next 3 columns are y_raw
batch_size = 2
poses = tf.constant([
    [1.0, 0.0, 0.0, 0.0, 1.0, 0.0],
    [0.9, 0.1, 0.0, -0.1, 0.95, 0.0]
], dtype=tf.float32)

ARGS = {
    "poses": poses
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
decode_shapelayer Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/sharedata/voxel2layer_torch.py
code served (permissive licence) · get_code("bef7ed3fa186b3c6")
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 (4, 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.functional as tf
import numpy as np

torch.manual_seed(0)

# Create a small shape_layer tensor
# The function expects shape_layer to have shape (side, side, 6)
# where the 6 channels represent [x0, x1, y0, y1, z0, z1]
side = 4
shape_layer = torch.randint(0, 3, (side, side, 6), dtype=torch.float32)

# Pre-generate indices for memory efficiency (optional, but shown in docstring)
id1, id2, id3 = None, None, None

ARGS = {
    'shape_layer': shape_layer,
    'id1': id1,
    'id2': id2,
    'id3': id3
}
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 3, dtype family float, shape (4, 4, 4) — shape (4, 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
dot Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
tangjiapeng/SkeletonNet/SkeGCNN/deformation/layers.py
pointer only (licence: MIT); the stored copy get_code serves, from thumlp/tensorgcn, is pointer only (licence: NONE) · get_code("bc3a2072a2a9cca3")
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 thumlp/tensorgcn, 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.

encode_shapelayer Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/sharedata/voxel2layer.py
code served (permissive licence) · get_code("e93bdf1174ceef8c")
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 3, dtype family int, shape (4, 4, 6) 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 a small cubic voxel grid (4x4x4)
# The voxel should contain some non-zero values to test the encoding
voxel = np.zeros((4, 4, 4), dtype=np.uint8)
# Add some voxel data - a simple pattern
voxel[1:3, 1:3, 1:3] = 1

# Pre-generate indices for memory efficiency
id1, id2, id3 = np.meshgrid(np.arange(0, 6), np.arange(0, 6), np.arange(0, 6), indexing='ij')

ARGS = {
    'voxel': voxel,
    'id1': id1,
    'id2': id2,
    'id3': id3
}
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 3, dtype family int, shape (4, 4, 6) — shape (4, 4, 6) 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
encode_shapelayer Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/sharedata/voxel2layer_torch.py
code served (permissive licence) · get_code("78eedf8cde9fb04e")
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 int, shape (4, 4, 6) 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)

# Create a small cubic voxel grid (4x4x4)
# The function expects a cube, so all dimensions must be equal
voxel = torch.randint(0, 2, (4, 4, 4), dtype=torch.uint8)

# Pre-generate indices to pass as optional parameters
id1, id2, id3 = None, None, None

ARGS = {
    'voxel': voxel,
    'id1': id1,
    'id2': id2,
    'id3': id3,
}
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 3, dtype family int, shape (4, 4, 6) — shape (4, 4, 6) 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
generate_indices Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/sharedata/voxel2layer.py
code served (permissive licence) · get_code("0b816a02ad9db4d5")
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 int; 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

ARGS = {
    "side": 5
}
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 int — 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
generate_indices Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/sharedata/voxel2layer_torch.py
code served (permissive licence) · get_code("c9b4e8f09bbdf20d")
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 int; 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)

ARGS = {
    "side": 3,
    "device": torch.device('cpu')
}
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 int — 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_fixed Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/models/posenet.py
code served (permissive licence) · get_code("6b1dc46a65c7b542")
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 tensorflow as tf
import numpy as np

# The function takes a single parameter: batch_size
# We'll set it to a small value for testing
batch_size = 2

ARGS = {
    "batch_size": batch_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
get_img_points Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/cam_est/model_cam.py
code served (permissive licence) · get_code("6d9d55677e711ac7")
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 tensorflow as tf
import numpy as np

# Create a small stand-in FLAGS class with required attributes
class FLAGS:
    def __init__(self):
        self.shift = True
        self.img_h = 256
        self.num_points = 5

flags = FLAGS()

# sample_pc: shape (batch_size, num_points, 3) - 3D point cloud
sample_pc = tf.constant(np.random.randn(2, 5, 3).astype(np.float32), dtype=tf.float32)

# trans_mat_right: transformation matrix (4x4) for homogeneous coordinates
# Shape: (4, 4) since homo_pc is (batch, num_points, 4) and we matmul with (4, 4)
trans_mat_right = tf.constant(np.eye(4, dtype=np.float32), dtype=tf.float32)

# pred_xyshift: predicted xy shift, shape (batch_size, 2)
pred_xyshift = tf.constant(np.random.randn(2, 2).astype(np.float32), dtype=tf.float32)

ARGS = {
    "sample_pc": sample_pc,
    "trans_mat_right": trans_mat_right,
    "pred_xyshift": pred_xyshift,
    "FLAGS": flags
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
get_layer_uid Ran
ran, honoured the contract we drafted
tangjiapeng/SkeletonNet/SkeGCNN/deformation/layers.py
pointer only (licence: MIT); the stored copy get_code serves, from thumlp/tensorgcn, is pointer only (licence: NONE) · get_code("b82968db452628fd")
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 thumlp/tensorgcn, 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.

laplace_coord Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeGCNN/deformation/losses.py
code served (permissive licence) · get_code("7d9c627fe8e7dcca")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)

# Create pred: shape [2, 3] - 2 vertices with 3D coordinates
pred = tf.constant([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]], dtype=tf.float32)

# Create placeholders dict with 'lape_idx'
# The function uses indices[:, :19] and weights[:, -1]
# So we need shape [num_vertices, 20] where last column is weights
# Let's create indices for 2 vertices, each with 19 neighbor indices + 1 weight
lape_idx = np.random.randint(0, 2, size=(2, 19)).astype(np.float32)
weights = np.array([[0.5], [0.3]], dtype=np.float32)
lape_idx_full = np.hstack([lape_idx, weights])

placeholders = {
    'lape_idx': tf.constant(lape_idx_full, dtype=tf.float32)
}

ARGS = {
    'pred': pred,
    'placeholders': placeholders
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
laplace_loss Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeGCNN/deformation/losses.py
code served (permissive licence) · get_code("8d5159a268b4c09f")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Create pred1 and pred2 tensors
# Shape: [num_vertices, 3] - representing vertex coordinates
# Using batch size 2 vertices for testing
pred1 = tf.constant([
    [0.1, 0.2, 0.3],
    [0.4, 0.5, 0.6]
], dtype=tf.float32)

pred2 = tf.constant([
    [0.11, 0.21, 0.31],
    [0.41, 0.51, 0.61]
], dtype=tf.float32)

# Create placeholders dict with 'lape_idx'
# lape_idx shape: [num_vertices, 20] where first 19 columns are indices and last column is weight
# For 2 vertices, we need indices that reference vertices (0 or 1) plus the padding vertex (2)
lape_idx = np.array([
    [0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0.5],
    [1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0.5]
], dtype=np.float32)

placeholders = {
    'lape_idx': tf.constant(lape_idx, dtype=tf.float32)
}

ARGS = {
    'pred1': pred1,
    'pred2': pred2,
    'placeholders': placeholders
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
normalize_vector Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/models/posenet.py
code served (permissive licence) · get_code("7d038625a3f89ebf")
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 SymbolicTensor. 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 tensorflow as tf

# Create a small batch of vectors with shape (batch_size, vector_dim)
v = tf.constant([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=tf.float32)

ARGS = {
    "v": v
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
repeat_tensor Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeDISN/models/model_add_skevox.py
code served (permissive licence) · get_code("5d1958e4bc7feb1c")
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 SymbolicTensor. 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 tensorflow as tf
import numpy as np

def tf_static_shape(T):
    return T.get_shape().as_list()

def repeat_tensor(T, nrep, rep_dim=1):
    repT = tf.expand_dims(T, rep_dim)
    tile_dim = [1] * len(tf_static_shape(repT))
    tile_dim[rep_dim] = nrep
    repT = tf.tile(repT, tile_dim)
    return repT

# Create a small test tensor with shape (2, 4)
T = tf.constant(np.random.randn(2, 4).astype(np.float32))

# Define arguments
ARGS = {
    'T': T,
    'nrep': 3,
    'rep_dim': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
unit Ran
ran on a synthesized input
tangjiapeng/SkeletonNet/SkeGCNN/deformation/losses.py
code served (permissive licence) · get_code("ecfd96810a9fa29d")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

def unit(tensor):
    return tf.nn.l2_normalize(tensor, dim=1)

# Create a small tensor with shape (2, 4) - batch size 2, dimension 4
tensor_np = np.array([[1.0, 2.0, 3.0, 4.0],
                       [5.0, 6.0, 7.0, 8.0]], dtype=np.float32)
tensor = tf.constant(tensor_np)

ARGS = {
    "tensor": tensor
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
decode_shapelayer Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/sharedata/voxel2layer.py
code served (permissive licence) · get_code("db0c1bf18185d93f")
How it ran

no invocation recorded.

placeholder_features Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeDISN/models/model_add_skevox.py
code served (permissive licence) · get_code("c71ac9a6926c461a")
How it ran

no invocation recorded.

placeholder_inputs Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeDISN/cam_est/model_cam.py
code served (permissive licence) · get_code("9a91699ee2fe9841")
How it ran

no invocation recorded.

placeholder_inputs Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeDISN/models/model_add_skevox.py
code served (permissive licence) · get_code("7cacff7bd55a7cc6")
How it ran

no invocation recorded.

placeholder_inputs Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeDISN/models/model_add_skevox_multiview.py
code served (permissive licence) · get_code("e8a74ed2dd506d62")
How it ran

no invocation recorded.

placeholder_inputs Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeDISN/models/model_normalization.py
code served (permissive licence) · get_code("bac3c27707f4b8e1")
How it ran

no invocation recorded.

sparse_dropout Not yet run
unverified (harvested, no recorded run)
tangjiapeng/SkeletonNet/SkeGCNN/deformation/layers.py
pointer only (licence: MIT); the stored copy get_code serves, from thumlp/tensorgcn, is pointer only (licence: NONE) · get_code("81be8a5f4732fe52")
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 thumlp/tensorgcn, 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. No invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 3 later) and 19 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
get_layer_uid 33 other papers
26 dated earlier, 7 later · 5 only as an assembly brick
1603.08861 (2016-03)
Revisiting Semi-Supervised Learning with Graph Embeddings
1606.09375 (2016-06)
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
1609.02907 (2016-09)
Semi-Supervised Classification with Graph Convolutional Networks
and 30 more not named here
sparse_dropout 18 other papers
15 dated earlier, 3 later · 2 only as an assembly brick
1603.08861 (2016-03)
Revisiting Semi-Supervised Learning with Graph Embeddings
1606.09375 (2016-06)
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
1609.02907 (2016-09)
Semi-Supervised Classification with Graph Convolutional Networks
and 15 more not named here
dot 17 other papers
14 dated earlier, 3 later · 3 only as an assembly brick
1603.08861 (2016-03)
Revisiting Semi-Supervised Learning with Graph Embeddings
1606.09375 (2016-06)
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
1609.02907 (2016-09)
Semi-Supervised Classification with Graph Convolutional Networks
and 14 more not named here
unit 1 other paper
1 dated earlier
1804.01654 (2018-04)
Pixel2Mesh: Generating 3D Mesh Models from Single RGB Images

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

This paper focuses on the challenging task of learning 3D object surface reconstructions from RGB images. Existingmethods achieve varying degrees of success by using different surface representations. However, they all have their own drawbacks,and cannot properly reconstruct the surface shapes of complex topologies, arguably due to a lack of constraints on the topologicalstructures in their learning frameworks. To this end, we propose to learn and use the topology-preserved, skeletal shape representationto assist the downstream task of object surface reconstruction from RGB images. Technically, we propose the novelSkeletonNetdesign that learns a volumetric representation of a skeleton via a bridged learning of a skeletal point set, where we use paralleldecoders each responsible for the learning of points on 1D skeletal curves and 2D skeletal sheets, as well as an efficient module ofglobally guided subvolume synthesis for a refined, high-resolution skeletal volume; we present a differentiablePoint2Voxellayer tomake SkeletonNet end-to-end and trainable. With the learned skeletal volumes, we propose two models, the Skeleton-Based GraphConvolutional Neural Network (SkeGCNN) and the Skeleton-Regularized Deep Implicit Surface Network (SkeDISN), which respectivelybuild upon and improve over the existing frameworks of explicit mesh deformation and implicit field learning for the downstream surfacereconstruction task. We conduct thorough experiments that verify the efficacy of our proposed SkeletonNet. SkeGCNN and SkeDISNoutperform existing methods as well, and they have their own merits when measured by different metrics. Additional results ingeneralized task settings further demonstrate the usefulness of our proposed methods. We have made both our implementation codeand the ShapeNet-Skeleton dataset publicly available at ble at https://github.com/tangjiapeng/SkeletonNet.

For agents

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

get_harvested_code_for_paper("2008.05742")
get_code_for_paper("2008.05742")
have("2008.05742")

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

curl https://syntology.ai/api/ran/2008.05742.json

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

[![Syntology run record](https://syntology.ai/api/ran/2008.05742.svg)](https://syntology.ai/paper/2008.05742)

Connect an agent — have() is free.