SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.01946 · 2020

Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild

arXiv · PDF · Open in the Atlas

Code that ran

18 of 20 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, 1 violated it and 10 have no contract verdict; 7 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/2004.01946.json · badge: /api/ran/2004.01946.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
EAST-J/Youtubehand community (classifier label)
classifier basis not recorded on the link
18 ran · 2 no recorded run
FunctionStatusWhere it lives
all_gather Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/utils/comm.py
code served (permissive licence) · get_code("d741eb912b24880a")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 pickle

torch.manual_seed(0)

# The function all_gather expects a single argument 'data' which can be any picklable object
# Looking at the function body, it pickles the data, converts to tensor, and gathers across ranks
# Since world_size will be 1 (no distributed setup), it will just return [data]
# We can pass any picklable object: dict, list, tensor, string, etc.

# Create a simple picklable data object
data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": torch.tensor([1.0, 2.0, 3.0, 4.0]),
    "nested": {"inner": "data"}
}

ARGS = {
    "data": data
}
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 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
conv1x1 Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/model/resnet.py
code served (permissive licence) · get_code("d9def42110729a85")
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 Conv2d. 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)

# The function conv1x1 takes in_planes, out_planes, and stride as parameters
# It returns an nn.Conv2d layer configured for 1x1 convolution
# We need to provide valid integer arguments for these parameters

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "stride": 1
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
conv3x3 Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/model/resnet.py
code served (permissive licence) · get_code("160bb14bd76201b4")
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 Conv2d. 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)

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "stride": 1,
    "groups": 1,
    "dilation": 1,
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
gather_on_master Ran
ran on a synthesized input
EAST-J/Youtubehand/utils/comm.py
code served (permissive licence) · get_code("d1a2abc62228483c")
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 pickle
import torch
import torch.distributed as dist

torch.manual_seed(0)

# Since the function doesn't require distributed setup to work
# (it handles the case where dist is not initialized),
# we can pass simple data that will be pickled and gathered.

# The function takes a single parameter 'data' which gets:
# 1. Pickled with pickle.dumps()
# 2. Converted to a torch.ByteTensor
# 3. Eventually unpickled back

# We'll create a small, picklable data structure
test_data = {
    "tensor": torch.randn(2, 4),
    "list": [1, 2, 3],
    "string": "test",
    "number": 42
}

ARGS = {
    "data": test_data
}
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 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
img_from_base64 Ran
ran on a synthesized input
EAST-J/Youtubehand/utils/image_ops.py
code served (permissive licence) · get_code("dc1e6e3b36faae62")
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 (100, 100, 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
import base64
import cv2

# Create a simple test image
test_image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8)

# Encode the image to base64
_, buffer = cv2.imencode('.jpg', test_image)
imagestring = base64.b64encode(buffer).decode('utf-8')

ARGS = {
    "imagestring": imagestring
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (100, 100, 3) — shape (100, 100, 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
keypoint_2d_loss Ran
ran on a synthesized input
EAST-J/Youtubehand/model/loss.py
code served (permissive licence) · get_code("a6aaf1479b7e0b9e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a simple loss criterion that takes predictions and targets
# and returns per-sample losses
class SimpleKeypointLoss(nn.Module):
    def forward(self, pred, target):
        # pred: (batch, num_keypoints, 2)
        # target: (batch, num_keypoints, 2)
        # Returns: (batch, num_keypoints, 1) losses
        diff = pred - target
        loss = torch.sqrt((diff ** 2).sum(dim=-1, keepdim=True) + 1e-8)
        return loss

# Create batch size 2, 4 keypoints, 3 coordinates (x, y, confidence)
batch_size = 2
num_keypoints = 4

# Ground truth keypoints: (batch, num_keypoints, 3) where last dim is (x, y, conf)
gt_keypoints_2d = torch.randn(batch_size, num_keypoints, 3)
gt_keypoints_2d[:, :, -1] = torch.rand(batch_size, num_keypoints)  # confidence in [0, 1]

# Predicted keypoints: (batch, num_keypoints, 2) for (x, y)
pred_keypoints_2d = torch.randn(batch_size, num_keypoints, 2)

# Loss criterion
criterion_keypoints = SimpleKeypointLoss()

# has_pose_2d is not used in the function body, but we include it
has_pose_2d = torch.ones(batch_size, dtype=torch.bool)

ARGS = {
    "criterion_keypoints": criterion_keypoints,
    "pred_keypoints_2d": pred_keypoints_2d,
    "gt_keypoints_2d": gt_keypoints_2d,
    "has_pose_2d": has_pose_2d,
}
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
keypoint_3d_loss Ran
ran on a synthesized input
EAST-J/Youtubehand/model/loss.py
code served (permissive licence) · get_code("e894cb2858907ff6")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 loss criterion that accepts two tensors and returns a tensor
class KeypointLoss(nn.Module):
    def forward(self, pred, gt):
        # Returns element-wise squared difference
        return (pred - gt) ** 2

criterion_keypoints = KeypointLoss()

# Create tensors with distinct dimensions
# Batch size: 2
# Number of keypoints: 4
# Coordinates: 3 (x, y, z)
# With confidence: 4 (x, y, z, conf)

batch_size = 2
num_keypoints = 4
coord_dim = 3

# gt_keypoints_3d: shape (batch_size, num_keypoints, coord_dim + 1)
# Last dimension includes confidence score
gt_keypoints_3d = torch.randn(batch_size, num_keypoints, coord_dim + 1)

# pred_keypoints_3d: shape (batch_size, num_keypoints, coord_dim)
pred_keypoints_3d = torch.randn(batch_size, num_keypoints, coord_dim)

# has_pose_3d: shape (batch_size,) with binary values indicating which samples have 3D pose
# Use distinct value (2) for batch dimension to avoid confusion
has_pose_3d = torch.tensor([1, 1], dtype=torch.long)

ARGS = {
    "criterion_keypoints": criterion_keypoints,
    "pred_keypoints_3d": pred_keypoints_3d,
    "gt_keypoints_3d": gt_keypoints_3d,
    "has_pose_3d": has_pose_3d,
}
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
make_batch_data_sampler Ran
ran on a synthesized input
EAST-J/Youtubehand/datasets/build.py
code served (permissive licence) · get_code("5d148f7214500fac")
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 IterationBasedBatchSampler. 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.utils.data

torch.manual_seed(0)

# Create a simple dataset
class SimpleDataset(torch.utils.data.Dataset):
    def __init__(self, size=10):
        self.size = size
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        return torch.randn(4), idx

# Create dataset and sampler
dataset = SimpleDataset(size=10)
sampler = torch.utils.data.SequentialSampler(dataset)

ARGS = {
    'sampler': sampler,
    'images_per_gpu': 2,
    'num_iters': 5,
    'start_iter': 0,
}
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 IterationBasedBatchSampler — a IterationBasedBatchSampler output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a IterationBasedBatchSampler 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
make_data_sampler Ran
ran on a synthesized input
EAST-J/Youtubehand/datasets/build.py
code served (permissive licence) · get_code("e17f1049b8148f71")
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 RandomSampler. 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
from torch.utils.data import Dataset

torch.manual_seed(0)

class DummyDataset(Dataset):
    def __init__(self, size=10):
        self.size = size
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        return torch.randn(4), torch.tensor(idx % 2)

dataset = DummyDataset(size=10)

ARGS = {
    "dataset": dataset,
    "shuffle": True,
    "distributed": False
}
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 RandomSampler — a RandomSampler output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a RandomSampler 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
myimresize Ran
ran on a synthesized input
EAST-J/Youtubehand/utils/image_ops.py
code served (permissive licence) · get_code("9b7d8b83f47e1b5f")
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 (5, 6, 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 cv2
import numpy as np

# Create a small test image (4x3 BGR image)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define target size
size = (6, 5)

ARGS = {
    'img': img,
    'size': size,
    'return_scale': False,
    'interpolation': 'bilinear'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (5, 6, 3) — shape (5, 6, 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
myimrotate Ran
ran on a synthesized input
EAST-J/Youtubehand/utils/image_ops.py
code served (permissive licence) · get_code("75b7a02e2231a13c")
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, 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
import cv2

# Create a small test image
# Using a 4x3 image (height=4, width=3) with 3 channels (RGB)
img = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Define arguments
ARGS = {
    'img': img,
    'angle': 45.0,
    'center': None,
    'scale': 1.0,
    'border_value': 0,
    'auto_bound': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
orthographic_projection Ran
ran on a synthesized input
EAST-J/Youtubehand/utils/geometric_layers.py
code served (permissive licence) · get_code("e6704260d6ef70f5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 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

torch.manual_seed(0)

# X: size = [B, N, 3] where B is batch size, N is number of points
# camera: size = [B, 3] where B is batch size, 3 contains [scale, tx, ty]

B = 2  # batch size
N = 4  # number of points
X = torch.randn(B, N, 3)
camera = torch.randn(B, 3)

ARGS = {
    "X": X,
    "camera": camera,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 2) — shape (2, 4, 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
quat2mat Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/utils/geometric_layers.py
code served (permissive licence) · get_code("baa1c1737c3796df")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 3, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function quat2mat expects a quaternion tensor of shape [B, 4]
# where B is batch size and 4 represents (w, x, y, z) components
# Create a small batch of 2 quaternions
quat = torch.randn(2, 4)

ARGS = {
    "quat": quat
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
reduce_dict Ran
ran, violated the contract we drafted
EAST-J/Youtubehand/utils/comm.py
code served (permissive licence) · get_code("2a3280757ce5a198")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned dict. 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.distributed as dist

torch.manual_seed(0)

# Create a simple input dictionary with tensor values
# The function expects all values to be tensors that can be stacked
input_dict = {
    'loss': torch.tensor(2.5),
    'accuracy': torch.tensor(0.85),
    'metric_a': torch.tensor(1.2),
}

# The average parameter is a boolean
average = True

ARGS = {
    'input_dict': input_dict,
    'average': average
}
Setup block drafted by
a language model, Claude Sonnet 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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
resnet18 Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/model/resnet.py
code served (permissive licence) · get_code("23b518f1014a72f5")
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 ResNet. 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)

# The function signature for resnet18 is not shown, but based on the ResNet class
# and common PyTorch patterns, resnet18 likely takes parameters like:
# pretrained, progress, num_classes, etc.

# Since we can't see the full resnet18 function, we'll construct arguments
# based on what the ResNet.__init__ method accepts and what a typical
# resnet18 wrapper function would accept

ARGS = {
    'pretrained': False,
    'progress': True,
    'num_classes': 1000,
    'zero_init_residual': False,
    'groups': 1,
    'width_per_group': 64,
    'replace_stride_with_dilation': None,
    'norm_layer': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResNet — a ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet 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
rodrigues Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/utils/geometric_layers.py
code served (permissive licence) · get_code("c15f1cb769ff504f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 3, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function rodrigues takes theta as input
# theta: size = [B, 3] - batch of axis-angle representations
# The function uses:
# - torch.norm(theta + 1e-08, p=2, dim=1)
# - torch.unsqueeze(l1norm, -1)
# - torch.div(theta, angle)
# - torch.cos(angle)
# - torch.sin(angle)
# - torch.cat([v_cos, v_sin * normalized], dim=1)
# - quat2mat(quat) where quat has size [B, 4]

# Create a batch of axis-angle representations
# Batch size: 2, each axis-angle vector has 3 components
theta = torch.randn(2, 3)

ARGS = {
    "theta": theta
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
setup_logger Ran
ran; our contract draft was wrong, not the code
EAST-J/Youtubehand/utils/logger.py
code served (permissive licence) · get_code("553f5a3047de32df")
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 Logger. 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 logging
import os
import sys
import tempfile

# Create a temporary directory for logging
temp_dir = tempfile.mkdtemp()

ARGS = {
    'name': 'test_logger',
    'save_dir': temp_dir,
    'distributed_rank': 0,
    'filename': 'log.txt'
}
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 Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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
vertices_loss Ran
ran on a synthesized input
EAST-J/Youtubehand/model/loss.py
code served (permissive licence) · get_code("087b738a254d4697")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 criterion that computes L2 loss between predictions and targets
criterion_vertices = nn.MSELoss()

# Create small tensors for vertices
# Batch size: 2, Vertex features: 3 (x, y, z coordinates)
batch_size = 2
num_vertices = 4
vertex_dim = 3

pred_vertices = torch.randn(batch_size, num_vertices, vertex_dim)
gt_vertices = torch.randn(batch_size, num_vertices, vertex_dim)

# Create has_smpl mask indicating which samples have SMPL annotations
# Use values 0 and 1 to filter samples
has_smpl = torch.tensor([1, 1])  # Both samples have SMPL annotations

ARGS = {
    "criterion_vertices": criterion_vertices,
    "pred_vertices": pred_vertices,
    "gt_vertices": gt_vertices,
    "has_smpl": has_smpl,
}
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
build_dataset Not yet run
unverified (harvested, no recorded run)
EAST-J/Youtubehand/datasets/build.py
code served (permissive licence) · get_code("3050f2388d1a578f")
How it ran

no invocation recorded.

get_cls_net Not yet run
unverified (harvested, no recorded run)
EAST-J/Youtubehand/model/hrnet/hrnet_cls_net.py
code served (permissive licence) · get_code("61d1082bf827cd4e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 15 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 3 the same month, 15 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
conv1x1 431 other papers
72 dated earlier, 5 the same month, 353 later, 1 undated · 38 only as an assembly brick
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1506.01186 (2015-06)
Cyclical Learning Rates for Training Neural Networks
1507.05717 (2015-07)
An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition
and 428 more not named here
conv3x3 312 other papers
39 dated earlier, 2 the same month, 270 later, 1 undated · 31 only as an assembly brick
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1507.05717 (2015-07)
An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
and 309 more not named here
resnet18 95 other papers
10 dated earlier, 84 later, 1 undated
1805.00907 (2018-05)
Glow: Graph Lowering Compiler Techniques for Neural Networks
1902.08153 (2019-02)
Learned Step Size Quantization
1904.01355 (2019-04)
FCOS: Fully Convolutional One-Stage Object Detection
and 92 more not named here
reduce_dict 15 other papers
1 dated earlier, 14 later
2003.04030 (2020-03)
Learning Delicate Local Representations for Multi-Person Pose Estimation
2108.02833 (2021-08)
Elaborative Rehearsal for Zero-shot Action Recognition
136730727 (2022 (year only))
Self-Supervised Sparse Representation for Video Anomaly Detection
and 12 more not named here
all_gather 7 other papers
7 later
2007.10985 (2020-07)
PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding
2104.10011 (2021-04)
Perceptual Loss for Robust Unsupervised Homography Estimation
2203.07111 (2022-03)
Disentangled Representation Learning for Text-Video Retrieval
and 4 more not named here
quat2mat 7 other papers
1 dated earlier, 6 later
1905.03244 (2019-05)
Convolutional Mesh Regression for Single-Image Human Shape Reconstruction
2006.05734 (2020-06)
3D Human Mesh Regression with Dense Correspondence
2108.06819 (2021-08)
EventHPE: Event-based 3D Human Pose and Shape Estimation
and 4 more not named here
orthographic_projection 6 other papers
1 dated earlier, 5 later
1905.03244 (2019-05)
Convolutional Mesh Regression for Single-Image Human Shape Reconstruction
2006.05734 (2020-06)
3D Human Mesh Regression with Dense Correspondence
2010.16143 (2020-10)
HyperText: Endowing FastText with Hyperbolic Geometry
and 3 more not named here
make_data_sampler 4 other papers
1 the same month, 3 later
2004.06165 (2020-04)
Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks
2101.00529 (2021-01)
VinVL: Revisiting Visual Representations in Vision-Language Models
2211.09371 (2022-11)
CapEnrich: Enriching Caption Semantics for Web Images via Cross-modal Pre-trained Knowledge
and 1 more not named here
rodrigues 4 other papers
1 dated earlier, 3 later
1905.03244 (2019-05)
Convolutional Mesh Regression for Single-Image Human Shape Reconstruction
2006.05734 (2020-06)
3D Human Mesh Regression with Dense Correspondence
and 1 more not named here
gather_on_master 1 other paper
1 later
Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper (2023 (year only))
Deformable Mesh Transformer for 3D Human Mesh Recovery
img_from_base64 1 other paper
1 later
Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper (2023 (year only))
Deformable Mesh Transformer for 3D Human Mesh Recovery
make_batch_data_sampler 1 other paper
1 later
Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper (2023 (year only))
Deformable Mesh Transformer for 3D Human Mesh Recovery
myimresize 1 other paper
1 later
Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper (2023 (year only))
Deformable Mesh Transformer for 3D Human Mesh Recovery
myimrotate 1 other paper
1 later
Yoshiyasu_Deformable_Mesh_Transformer_for_3D_Human_Mesh_Recovery_CVPR_2023_paper (2023 (year only))
Deformable Mesh Transformer for 3D Human Mesh Recovery
setup_logger 1 other paper
1 later
2108.04024 (2021-08)
Image Retrieval on Real-life Images with Pre-trained Vision-and-Language Models

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 introduce a simple and effective network architecture for monocular 3D hand pose estimation consisting of an image encoder followed by a mesh convolutional decoder that is trained through a direct 3D hand mesh reconstruction loss. We train our network by gathering a large-scale dataset of hand action in YouTube videos and use it as a source of weak supervision. Our weakly-supervised mesh convolutions-based system largely outperforms state-of-the-art methods, even halving the errors on the in the wild benchmark. The dataset and additional resources are available at https://arielai.com/mesh_hands.

For agents

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

get_harvested_code_for_paper("2004.01946")
get_code_for_paper("2004.01946")
have("2004.01946")

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

curl https://syntology.ai/api/ran/2004.01946.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.01946.svg)](https://syntology.ai/paper/2004.01946)

Connect an agent — have() is free.