SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.06110 · 2022

Uplift and Upsample: Efficient 3D Human Pose Estimation with Uplifting Transformers

arXiv · PDF · Open in the Atlas

Code that ran

15 of 17 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 15 have no contract verdict; 0 failures of our instrument; 2 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/2210.06110.json · badge: /api/ran/2210.06110.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
goldbricklemon/uplift-upsample-3dhpe canonical (classifier label)
classifier basis not recorded on the link
7 ran · 2 no recorded run
kaulquappe23/a2b_human_mesh community (classifier label)
classifier basis not recorded on the link
8 ran · 0 no recorded run
FunctionStatusWhere it lives
create_image_paths Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/dataset/h36m_splits.py
code served (permissive licence) · get_code("05ac12c362717f7a")
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 list, 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 os
import tempfile

ARGS = {
    'base_path': tempfile.gettempdir(),
    'subject': 'subject_001',
    'action': 'walk',
    'cam_id': '01',
    'frame_nums': [0, 1, 5, 10, 100]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list 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
drop_path Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/net/vision_transformer.py
code served (permissive licence) · get_code("6aafcfcb574cb205")
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

# Create a small TensorFlow tensor for x
# Using shape (2, 4) - batch size 2, feature dimension 4
x = tf.constant(np.random.randn(2, 4).astype(np.float32))

# Create a drop_path rate (probability of dropping)
rate = 0.5

ARGS = {
    "x": x,
    "rate": rate
}
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
interpolate_between_keyframes Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/dataset/action_wise_eval.py
code served (permissive licence) · get_code("b723e9f4c150f22f")
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

# Construct arguments for interpolate_between_keyframes
# The function interpolates 3D predictions between keyframes

# pred3d: numpy array of 3D predictions with shape (num_frames, num_points, 3)
# frame_indices: numpy array of frame indices
# keyframe_stride: integer specifying the stride between keyframes

pred3d = np.random.randn(8, 5, 3).astype(np.float32)

frame_indices = np.array([0, 1, 2, 3, 4, 5, 6, 7])

keyframe_stride = 2

ARGS = {
    "pred3d": pred3d,
    "frame_indices": frame_indices,
    "keyframe_stride": keyframe_stride,
}
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
mpjpe Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/dataset/metrics.py
code served (permissive licence) · get_code("3b759fd80717758c")
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

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

# Define batch size and number of keypoints
batch_size = 2
num_keypoints = 4
root_index = 0

# Create pred: (B, K, 3) - 3D pose predictions in (x, y, z) format
pred = np.random.randn(batch_size, num_keypoints, 3).astype(np.float32)

# Create gt: (B, K, 4) - 3D pose ground truth in (x, y, z, v) format
# v is the valid flag (> 0 means valid)
gt = np.random.randn(batch_size, num_keypoints, 4).astype(np.float32)
# Ensure valid flags are positive (valid=1, invalid=0)
gt[:, :, 3] = np.random.choice([0, 1], size=(batch_size, num_keypoints))

ARGS = {
    "pred": pred,
    "gt": gt,
    "root_index": root_index,
    "normalize": True,
}
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
mpjpe Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/metrics.py
code served (permissive licence) · get_code("2f8afe6d7dc3c1cc")
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

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

# Create small dummy data
batch_size = 2
num_keypoints = 4
root_index = 0

# pred: (B, K, 3) - 3D pose predictions
pred = np.random.randn(batch_size, num_keypoints, 3).astype(np.float32)

# gt: (B, K, 4) - 3D pose ground truth with validity flag
# Last dimension: (x, y, z, v) where v is validity
gt = np.random.randn(batch_size, num_keypoints, 4).astype(np.float32)
# Set validity flags (v > 0 means valid)
gt[:, :, 3] = np.random.uniform(0, 2, size=(batch_size, num_keypoints))

ARGS = {
    "pred": pred,
    "gt": gt,
    "root_index": root_index,
    "normalize": True,
}
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
nmpjpe Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/dataset/metrics.py
code served (permissive licence) · get_code("9e168cb44cb7d139")
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

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

# Create small test data
batch_size = 2
num_keypoints = 4
root_index = 0

# pred: (B, K, 3) - 3D pose predictions
pred = np.random.randn(batch_size, num_keypoints, 3).astype(np.float32)

# gt: (B, K, 4) - 3D pose ground truth with validity flag
# Last dimension is (x, y, z, valid_flag)
gt = np.random.randn(batch_size, num_keypoints, 4).astype(np.float32)
# Set validity flags to 1.0 (all keypoints valid)
gt[:, :, 3] = 1.0

ARGS = {
    'pred': pred,
    'gt': gt,
    'root_index': root_index,
    'alignment': 'root',
    'normalize': True,
}
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
nmpjpe Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/metrics.py
code served (permissive licence) · get_code("ff8248d869523ab9")
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

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

# Create small test data
batch_size = 2
num_keypoints = 4
root_index = 0

# pred: (B, K, 3) - 3D pose predictions
# Shape: (2, 4, 3)
pred = np.random.randn(batch_size, num_keypoints, 3).astype(np.float32)

# gt: (B, K, 4) - 3D pose ground truth with validity flag (x, y, z, v)
# Shape: (2, 4, 4)
gt = np.random.randn(batch_size, num_keypoints, 4).astype(np.float32)
# Set validity flags: last column should be > 0 for valid keypoints
gt[:, :, 3] = np.array([[1.0, 1.0, 1.0, 0.0],
                        [1.0, 1.0, 0.0, 1.0]], dtype=np.float32)

# Create ARGS dictionary with all parameters
ARGS = {
    'pred': pred,
    'gt': gt,
    'root_index': root_index,
    'alignment': 'root',
    'normalize': True
}
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
np_qinverse Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/quaternion.py
code served (permissive licence) · get_code("3058f77b597a257f")
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, 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

# Create a small quaternion array for testing
# Shape: (2, 4) - batch of 2 quaternions, each with 4 components (w, x, y, z)
q = np.array([
    [1.0, 0.5, 0.3, 0.2],
    [0.9, 0.1, 0.4, 0.6]
], dtype=np.float32)

ARGS = {
    "q": q,
    "inplace": False
}
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, 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
np_qrot Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/quaternion.py
code served (permissive licence) · get_code("8242cae54a825bc8")
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

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

# Create quaternion q with shape (..., 4)
# Using batch size 2 for testing
q = np.random.randn(2, 4).astype(np.float32)
# Normalize quaternions to be valid
q = q / np.linalg.norm(q, axis=-1, keepdims=True)

# Create vector v with shape (..., 3)
# Must have same leading dimensions as q
v = np.random.randn(2, 3).astype(np.float32)

ARGS = {
    "q": q,
    "v": v,
}
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
pmpjpe Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/dataset/metrics.py
code served (permissive licence) · get_code("a921292bf596fdc7")
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

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

# Create small batch of 3D pose predictions
# Shape: (B=2, K=4, 3) - 2 examples, 4 keypoints, 3D coordinates
pred = np.random.randn(2, 4, 3).astype(np.float32)

# Create small batch of ground truth 3D poses with validity flags
# Shape: (B=2, K=4, 4) - 2 examples, 4 keypoints, (x, y, z, validity)
gt = np.random.randn(2, 4, 4).astype(np.float32)
# Ensure validity flags are > 0 for all keypoints (set to 1)
gt[:, :, 3] = 1.0

ARGS = {
    "pred": pred,
    "gt": gt,
    "normalize": True,
}
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
pmpjpe Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/metrics.py
code served (permissive licence) · get_code("58f06c045c7a3705")
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

np.random.seed(0)

# Create small test data
# pred: (B, K, 3) - batch of 2, 4 keypoints, 3D coordinates
batch_size = 2
num_keypoints = 4
pred = np.random.randn(batch_size, num_keypoints, 3).astype(np.float32)

# gt: (B, K, 4) - batch of 2, 4 keypoints, 3D coordinates + validity flag
gt = np.random.randn(batch_size, num_keypoints, 4).astype(np.float32)
# Set validity flags (last column) to positive values
gt[:, :, 3] = np.random.uniform(0.5, 1.5, size=(batch_size, num_keypoints))

# Test with normalize=True (default)
normalize = True

ARGS = {
    "pred": pred,
    "gt": gt,
    "normalize": normalize,
}
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
pt_qinverse Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/uplifiting_dataset.py
code served (permissive licence) · get_code("6a6a59635b7602ec")
How it ran

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

import torch

torch.manual_seed(0)

# q is a tensor with shape (..., 4) representing quaternions
# The function extracts the first element (w) and negates the last 3 elements (xyz)
q = torch.randn(2, 4)

ARGS = {
    "q": q
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 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
pt_qrot Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/uplifiting_dataset.py
code served (permissive licence) · get_code("95c89076de9f6be8")
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, 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)

# Create a quaternion tensor with shape (2, 4)
# Last dimension is 4 for quaternion (w, x, y, z)
q = torch.randn(2, 4)

# Create a vector tensor with shape (2, 3)
# Last dimension is 3 for 3D vector (x, y, z)
v = torch.randn(2, 3)

ARGS = {
    "q": q,
    "v": v,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 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
tf_mpjpe Ran
ran on a synthesized input
goldbricklemon/uplift-upsample-3dhpe/common/utils/losses_3d.py
code served (permissive licence) · get_code("eec340f4b62ea073")
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

# Create small tensors for pred and gt
# tf_mpjpe computes tf.norm(gt - pred, axis=-1)
# So both pred and gt should have the same shape with last dimension being coordinates

# Create batch of 2 samples with 3 joints, 4 coordinates each
pred = tf.constant(np.random.randn(2, 3, 4).astype(np.float32))
gt = tf.constant(np.random.randn(2, 3, 4).astype(np.float32))

ARGS = {
    "pred": pred,
    "gt": gt,
}
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
world_to_cam_and_2d Ran
ran on a synthesized input
kaulquappe23/a2b_human_mesh/dataset/uplifiting_dataset.py
code served (permissive licence) · get_code("ac0a2e613ce83f5c")
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 torch

torch.manual_seed(0)

# sequence_3d: (N, K, 3) - 3D points in world coordinates
# N=2 sequences, K=4 keypoints, 3 coordinates
sequence_3d = torch.randn(2, 4, 3)

# cam: (4+3+11=18) - quaternion rotation (4) + translation (3) + intrinsics (11)
# Quaternion: [w, x, y, z]
# Translation: [tx, ty, tz]
# Intrinsics: [?, ?, fx, fy, cx, cy, k1, k2, k3, p1, p2]
cam = torch.randn(18)

ARGS = {
    "sequence_3d": sequence_3d,
    "cam": cam,
}
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 — 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
image_coordinates Not yet run
unverified (harvested, no recorded run)
goldbricklemon/uplift-upsample-3dhpe/common/dataset/camera.py
pointer only (licence: MIT); the stored copy get_code serves, from JihuaPeng/KTPFormer, is pointer only (licence: NONE) · get_code("7669d6b1ce1e7096")
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 JihuaPeng/KTPFormer, 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.

normalize_screen_coordinates Not yet run
unverified (harvested, no recorded run)
goldbricklemon/uplift-upsample-3dhpe/common/dataset/camera.py
pointer only (licence: MIT); the stored copy get_code serves, from JihuaPeng/KTPFormer, is pointer only (licence: NONE) · get_code("82baf6aa4fb040a1")
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 JihuaPeng/KTPFormer, 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 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 15 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_screen_coordinates 10 other papers
5 dated earlier, 5 later
1904.03345 (2019-04)
Semantic Graph Convolutional Networks for 3D Human Pose Regression
2004.13985 (2020-04)
Motion Guided 3D Pose Estimation from Videos
2103.14304 (2021-03)
Exploiting Temporal Contexts with Strided Transformer for 3D Human Pose Estimation
and 7 more not named here
image_coordinates 4 other papers
1 dated earlier, 3 later
1904.03345 (2019-04)
Semantic Graph Convolutional Networks for 3D Human Pose Regression
2211.16940 (2022-11)
DiffPose: Toward More Reliable 3D Pose Estimation
2403.11310 (2024-03)
A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation
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

The state-of-the-art for monocular 3D human pose estimation in videos is dominated by the paradigm of 2D-to-3D pose uplifting. While the uplifting methods themselves are rather efficient, the true computational complexity depends on the per-frame 2D pose estimation. In this paper, we present a Transformer-based pose uplifting scheme that can operate on temporally sparse 2D pose sequences but still produce temporally dense 3D pose estimates. We show how masked token modeling can be utilized for temporal upsampling within Transformer blocks. This allows to decouple the sampling rate of input 2D poses and the target frame rate of the video and drastically decreases the total computational complexity. Additionally, we explore the option of pre-training on large motion capture archives, which has been largely neglected so far. We evaluate our method on two popular benchmark datasets: Human3.6M and MPI-INF-3DHP. With an MPJPE of 45.0 mm and 46.9 mm, respectively, our proposed method can compete with the state-of-the-art while reducing inference time by a factor of 12. This enables real-time throughput with variable consumer hardware in stationary and mobile applications. We release our code and models at https://github.com/goldbricklemon/uplift-upsample-3dhpe

For agents

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

get_harvested_code_for_paper("2210.06110")
get_code_for_paper("2210.06110")
have("2210.06110")

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

curl https://syntology.ai/api/ran/2210.06110.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.06110.svg)](https://syntology.ai/paper/2210.06110)

Connect an agent — have() is free.