SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.04968 · 2020

Would Mega-scale Datasets Further Enhance Spatiotemporal 3D CNNs?

arXiv · PDF · Open in the Atlas

Code that ran

16 of 22 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 14 have no contract verdict; 2 failures of our instrument; 6 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/2004.04968.json · badge: /api/ran/2004.04968.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
kenshohara/3D-ResNets-PyTorch canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
ziqi-zhang/3D-ResNets-PyTorch canonical (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
KevinPang180/3D-ResNets-PaddlePaddle community (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
Oichii/resnet3D_pulse community (classifier label)
classifier basis not recorded on the link
0 ran · 3 no recorded run
FunctionStatusWhere it lives
conv1x1x1 Ran
ran; our contract draft was wrong, not the code
ziqi-zhang/3D-ResNets-PyTorch/models/resnet.py
code served (permissive licence) · get_code("213ccc8fb87ee8d9")
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 Conv3d. 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 conv1x1x1 takes in_planes, out_planes, and stride parameters
# in_planes: number of input channels
# out_planes: number of output channels
# stride: stride for the convolution (default=1)

in_planes = 3
out_planes = 5
stride = 2

ARGS = {
    'in_planes': in_planes,
    'out_planes': out_planes,
    'stride': stride
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d 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
conv1x3x3 Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/models/resnet2p1d.py
code served (permissive licence) · get_code("b471bd70ae7033b7")
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 Conv3d. 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,
    "mid_planes": 16,
    "stride": 1
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d 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
conv3x1x1 Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/models/resnet2p1d.py
code served (permissive licence) · get_code("69cf8d9e032b3251")
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 Conv3d. 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 = {
    'mid_planes': 16,
    'planes': 32,
    'stride': 1
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d 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
conv3x3x3 Ran
ran; our contract draft was wrong, not the code
ziqi-zhang/3D-ResNets-PyTorch/models/resnet.py
code served (permissive licence) · get_code("c3b215555357b9d1")
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 Conv3d. 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 conv3x3x3 creates a 3D convolution layer
# Parameters:
# - in_planes: number of input channels
# - out_planes: number of output channels  
# - stride: stride for the convolution (default=1)

# We need to provide distinct values for in_planes and out_planes
in_planes = 3
out_planes = 5
stride = 2

ARGS = {
    'in_planes': in_planes,
    'out_planes': out_planes,
    'stride': stride
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d 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
generate_model Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/models/densenet.py
code served (permissive licence) · get_code("f1faf2b3a11cd3c6")
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 DenseNet. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from collections import OrderedDict

torch.manual_seed(0)

# Build a minimal input tensor: batch_size=2, channels=3, depth=4, height=4, width=4
input_tensor = torch.randn(2, 3, 4, 4, 4)

# The function generate_model is incomplete in the provided code,
# but based on the pattern, it should accept model_depth and **kwargs.
# We'll construct arguments that would work with a complete DenseNet implementation.

ARGS = {
    "model_depth": 121,
}
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 DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
get_class_labels Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/datasets/activitynet.py
code served (permissive licence) · get_code("0b3b02ceef079aae")
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 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

torch.manual_seed(0)

# The function expects a dict with a 'taxonomy' key
# The taxonomy is a list of dicts, each with 'nodeId', 'parentId', and 'nodeName'
# We need to construct a valid taxonomy where some nodes are leaves (have no children)

data = {
    'taxonomy': [
        {
            'nodeId': 1,
            'parentId': None,
            'nodeName': 'root'
        },
        {
            'nodeId': 2,
            'parentId': 1,
            'nodeName': 'parent_node'
        },
        {
            'nodeId': 3,
            'parentId': 2,
            'nodeName': 'leaf_class_1'
        },
        {
            'nodeId': 4,
            'parentId': 2,
            'nodeName': 'leaf_class_2'
        },
        {
            'nodeId': 5,
            'parentId': None,
            'nodeName': 'leaf_class_3'
        }
    ]
}

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 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
get_fine_tuning_parameters Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/model.py
code served (permissive licence) · get_code("c59254c1a5526d41")
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 torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.features = nn.Sequential(
            nn.Linear(4, 3),
            nn.ReLU()
        )
        self.layer1 = nn.Linear(3, 2)
        self.layer2 = nn.Linear(2, 1)
    
model = SimpleModel()

ft_begin_module = "layer1"

ARGS = {
    "model": model,
    "ft_begin_module": ft_begin_module
}
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
get_mean_std Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/mean.py
code served (permissive licence) · get_code("d35f7d3b70ef9b70")
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 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

ARGS = {
    'value_scale': 255.0,
    'dataset': 'activitynet'
}
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 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_module_name Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/model.py
code served (permissive licence) · get_code("c347904eafd0d22e")
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 str. 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

def get_module_name(name):
    name = name.split('.')
    if name[0] == 'module':
        i = 1
    else:
        i = 0
    if name[i] == 'features':
        i += 1
    return name[i]

# The function takes a string parameter 'name' and splits it by '.'
# It then indexes into the resulting list based on whether 'module' and 'features' are present
# We need to construct a valid string that will work with this logic

# Test case 1: Simple name without 'module' or 'features'
# "layer1" -> name = ["layer1"], i starts at 0, returns "layer1"

# Test case 2: Name with 'module' prefix
# "module.layer1" -> name = ["module", "layer1"], i becomes 1, returns "layer1"

# Test case 3: Name with 'module' and 'features'
# "module.features.layer1" -> name = ["module", "features", "layer1"], i becomes 2, returns "layer1"

# Test case 4: Name with just 'features'
# "features.layer1" -> name = ["features", "layer1"], i becomes 1, returns "layer1"

# Any of these would be valid. Let's use a comprehensive one:
name_arg = "module.features.layer1"

ARGS = {
    "name": name_arg
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
get_n_frames Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/datasets/activitynet.py
code served (permissive licence) · get_code("92f569d7dc7c5f17")
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 int, ndim 0, dtype family int, 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 tempfile
from pathlib import Path

# Create a temporary directory structure with image files
temp_dir = tempfile.TemporaryDirectory()
video_path = Path(temp_dir.name)

# Create some image files
(video_path / "image_001.png").touch()
(video_path / "image_002.png").touch()
(video_path / "image_003.png").touch()
# Add a hidden file that should be ignored
(video_path / ".hidden_image.png").touch()
# Add a non-image file that should be ignored
(video_path / "metadata.txt").touch()

ARGS = {
    "video_path": video_path
}
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 int, ndim 0, dtype family int, 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
get_video_ids_annotations_and_fps Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/datasets/activitynet.py
code served (permissive licence) · get_code("8dd3e55b4ebf8cc6")
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)

ARGS = {
    'data': {
        'database': {
            'video_001': {
                'subset': 'train',
                'annotations': [
                    {'segment': [0, 10], 'label': 'action1'},
                    {'segment': [15, 25], 'label': 'action2'}
                ],
                'fps': 30
            },
            'video_002': {
                'subset': 'val',
                'annotations': [
                    {'segment': [5, 15], 'label': 'action3'}
                ],
                'fps': 24
            },
            'video_003': {
                'subset': 'train',
                'annotations': [
                    {'segment': [0, 20], 'label': 'action1'}
                ],
                'fps': 30
            },
            'video_004': {
                'subset': 'test',
                'annotations': [],
                'fps': 25
            }
        }
    },
    'subset': 'train'
}
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
group_multi_scale_crop Ran
ran on a synthesized input
KevinPang180/3D-ResNets-PaddlePaddle/reader.py
code served (permissive licence) · get_code("1f00a5b96d91ec0d")
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 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 random
from PIL import Image
import numpy as np

random.seed(0)

# Create a small test image group
# The function expects img_group to be a list of PIL Images
# Each image needs .size attribute and .crop() and .resize() methods
test_image_1 = Image.new('RGB', (640, 480), color='red')
test_image_2 = Image.new('RGB', (640, 480), color='blue')
img_group = [test_image_1, test_image_2]

ARGS = {
    'img_group': img_group,
    'target_size': 224,
    'scales': [1, 0.875, 0.75, 0.66],
    'max_distort': 1,
    'fix_crop': True,
    'more_fix_crop': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 4, dtype family int — 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
group_random_crop Ran
ran on a synthesized input
KevinPang180/3D-ResNets-PaddlePaddle/reader.py
code served (permissive licence) · get_code("865572e58d23a992")
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 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.

from PIL import Image
import random

# Set seed for reproducibility
random.seed(42)

# Create a group of test images
# The function expects img_group[0].size to work, so we need PIL Image objects
img1 = Image.new('RGB', (100, 80))  # width=100, height=80
img2 = Image.new('RGB', (100, 80))  # same size
img_group = [img1, img2]

# Define target_size
target_size = 64

ARGS = {
    'img_group': img_group,
    'target_size': target_size,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 4, dtype family int — 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
group_random_flip Ran
ran on a synthesized input
KevinPang180/3D-ResNets-PaddlePaddle/reader.py
code served (permissive licence) · get_code("e5bd557fa83521ca")
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 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 random
from PIL import Image
import numpy as np

# Create a small group of PIL Images
img1 = Image.new('RGB', (4, 3), color='red')
img2 = Image.new('RGB', (4, 3), color='blue')
img_group = [img1, img2]

ARGS = {
    'img_group': img_group
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 4, dtype family int — 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
image_name_formatter Ran
ran on a synthesized input
kenshohara/3D-ResNets-PyTorch/dataset.py
code served (permissive licence) · get_code("138fe18f637e2e35")
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 str. 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

def image_name_formatter(x):
    return f'image_{x:05d}.jpg'

ARGS = {
    'x': 42
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
merge_configs Ran
ran on a synthesized input
KevinPang180/3D-ResNets-PaddlePaddle/config.py
code served (permissive licence) · get_code("0e03ada3bc54e4fb")
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 Config. 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)

class Config:
    def __init__(self):
        self.TRAIN = ConfigSection()
        self.VALID = ConfigSection()
        self.TEST = ConfigSection()
        self.INFER = ConfigSection()

class ConfigSection:
    def __init__(self):
        self.learning_rate = 0.001
        self.batch_size = 32
        self.epochs = 10
        self.dropout = 0.5

cfg = Config()
sec = 'train'
args_dict = {
    'learning_rate': 0.01,
    'batch_size': 64,
    'dropout': None,
    'unknown_param': 0.2
}

ARGS = {
    'cfg': cfg,
    'sec': sec,
    'args_dict': args_dict
}
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 Config — a Config output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Config 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
accuracy Not yet run
unverified (harvested, no recorded run)
Oichii/resnet3D_pulse/ResNet_train.py
code served (permissive licence) · get_code("f0c9a29156911331")
How it ran

no invocation recorded.

generate_model Not yet run
unverified (harvested, no recorded run)
ziqi-zhang/3D-ResNets-PyTorch/models/resnet.py
code served (permissive licence) · get_code("392a1ed5b8624559")
How it ran

no invocation recorded.

generate_model Not yet run
unverified (harvested, no recorded run)
kenshohara/3D-ResNets-PyTorch/models/pre_act_resnet.py
code served (permissive licence) · get_code("45343fb6310c6c49")
How it ran

no invocation recorded.

generate_model Not yet run
unverified (harvested, no recorded run)
Oichii/resnet3D_pulse/ResNet_model.py
code served (permissive licence) · get_code("c03ab7046d986487")
How it ran

no invocation recorded.

get_video_results Not yet run
unverified (harvested, no recorded run)
kenshohara/3D-ResNets-PyTorch/inference.py
code served (permissive licence) · get_code("c0e7431aff3ed281")
How it ran

no invocation recorded.

validate Not yet run
unverified (harvested, no recorded run)
Oichii/resnet3D_pulse/ResNet_train.py
code served (permissive licence) · get_code("9436376f4d71d64e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 2 the same month, 4 later) and 17 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
accuracy 71 other papers
26 dated earlier, 1 the same month, 44 later · 2 only as an assembly brick
1505.05424 (2015-05)
Weight Uncertainty in Neural Networks
1705.07204 (2017-05)
Ensemble Adversarial Training: Attacks and Defenses
1802.10399 (2018-02)
Compressing Neural Networks using the Variational Information Bottleneck
and 68 more not named here
conv3x3x3 19 other papers
7 dated earlier, 1 the same month, 11 later
1711.08580 (2017-11)
3D Anisotropic Hybrid Network: Transferring Convolutional Features from 2D Images to 3D Anisotropic Volumes
1711.09577 (2017-11)
Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
1908.01373 (2019-08)
Unsupervised Microvascular Image Segmentation Using an Active Contours Mimicking Neural Network
and 16 more not named here
conv1x1x1 6 other papers
2 dated earlier, 4 later · 1 only as an assembly brick
1711.09577 (2017-11)
Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
2003.10432 (2020-03)
Atlas: End-to-End 3D Scene Reconstruction from Posed Images
2107.07695 (2021-07)
Self-supervised Representation Learning Framework for Remote Physiological Measurement Using Spatiotemporal Augmentation Loss
and 3 more not named here
generate_model 3 other papers
1 dated earlier, 2 later
1711.09577 (2017-11)
Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
2107.07695 (2021-07)
Self-supervised Representation Learning Framework for Remote Physiological Measurement Using Spatiotemporal Augmentation Loss
2108.02183 (2021-08)
Enhancing Self-supervised Video Representation Learning via Multi-level Feature Optimization

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

How can we collect and use a video dataset to further improve spatiotemporal 3D Convolutional Neural Networks (3D CNNs)? In order to positively answer this open question in video recognition, we have conducted an exploration study using a couple of large-scale video datasets and 3D CNNs. In the early era of deep neural networks, 2D CNNs have been better than 3D CNNs in the context of video recognition. Recent studies revealed that 3D CNNs can outperform 2D CNNs trained on a large-scale video dataset. However, we heavily rely on architecture exploration instead of dataset consideration. Therefore, in the present paper, we conduct exploration study in order to improve spatiotemporal 3D CNNs as follows: (i) Recently proposed large-scale video datasets help improve spatiotemporal 3D CNNs in terms of video classification accuracy. We reveal that a carefully annotated dataset (e.g., Kinetics-700) effectively pre-trains a video representation for a video classification task. (ii) We confirm the relationships between #category/#instance and video classification accuracy. The results show that #category should initially be fixed, and then #instance is increased on a video dataset in case of dataset construction. (iii) In order to practically extend a video dataset, we simply concatenate publicly available datasets, such as Kinetics-700 and Moments in Time (MiT) datasets. Compared with Kinetics-700 pre-training, we further enhance spatiotemporal 3D CNNs with the merged dataset, e.g., +0.9, +3.4, and +1.1 on UCF-101, HMDB-51, and ActivityNet datasets, respectively, in terms of fine-tuning. (iv) In terms of recognition architecture, the Kinetics-700 and merged dataset pre-trained models increase the recognition performance to 200 layers with the Residual Network (ResNet), while the Kinetics-400 pre-trained model cannot successfully optimize the 200-layer architecture.

For agents

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

get_harvested_code_for_paper("2004.04968")
get_code_for_paper("2004.04968")
have("2004.04968")

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

curl https://syntology.ai/api/ran/2004.04968.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.04968.svg)](https://syntology.ai/paper/2004.04968)

Connect an agent — have() is free.