SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_16433 · AAAI · 2021

CAKES: Channel-wise Automatic KErnel Shrinking for Efficient 3D Networks

Alan Yuille, Yingwei Li, Qihang Yu, Jieru Mei, Yuyin Zhou

Open in the Atlas

Code that ran

3 of 4 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 2 have no contract verdict; 1 failure of our instrument; 1 no recorded run), measured 2026-11-22; 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/aaai_16433.json · badge: /api/ran/aaai_16433.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
yucornetto/CAKES canonical (classifier label)
classifier basis: census
3 ran · 1 no recorded run
FunctionStatusWhere it lives
return_dataset Ran
ran; our contract draft was wrong, not the code
yucornetto/CAKES/datasets_video.py
code served (permissive licence) · get_code("ad16fb9c63011eff")
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 os
import tempfile

# Create temporary directory structure with minimal files
temp_root = tempfile.mkdtemp()
ROOT_DATASET = temp_root + '/'

# Create directory structure
os.makedirs(os.path.join(temp_root, 'something-something-v1'), exist_ok=True)
os.makedirs(os.path.join(temp_root, '20bn-something-something-v2-frames'), exist_ok=True)

# Create category.txt file with some sample categories
category_file_v1 = os.path.join(temp_root, 'something-something-v1', 'category.txt')
with open(category_file_v1, 'w') as f:
    f.write('action1\n')
    f.write('action2\n')
    f.write('action3\n')

category_file_v2 = os.path.join(temp_root, 'category.txt')
with open(category_file_v2, 'w') as f:
    f.write('action1\n')
    f.write('action2\n')

# Create train/val video folder files for v1
train_file_v1 = os.path.join(temp_root, 'something-something-v1', 'train_videofolder.txt')
with open(train_file_v1, 'w') as f:
    f.write('video1 1\n')
    f.write('video2 2\n')

val_file_v1 = os.path.join(temp_root, 'something-something-v1', 'val_videofolder.txt')
with open(val_file_v1, 'w') as f:
    f.write('video3 1\n')

# Create train/val video folder files for v2
train_file_v2 = os.path.join(temp_root, 'train_video_folder.txt')
with open(train_file_v2, 'w') as f:
    f.write('video1 1\n')
    f.write('video2 2\n')

val_file_v2 = os.path.join(temp_root, 'val_video_folder.txt')
with open(val_file_v2, 'w') as f:
    f.write('video3 1\n')

ARGS = {
    'dataset': 'somethingv2',
    'ROOT_DATASET': ROOT_DATASET
}
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 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
return_somethingv1 Ran
ran on a synthesized input
yucornetto/CAKES/datasets_video.py
code served (permissive licence) · get_code("f83857326ee082b5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# The function takes ROOT_DATASET as a parameter
# It constructs file paths and returns them
# No actual file I/O is performed in the function body

ARGS = {
    'ROOT_DATASET': '/fake/dataset/root/'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
return_somethingv2 Ran
ran on a synthesized input
yucornetto/CAKES/datasets_video.py
code served (permissive licence) · get_code("d23f73a8468f2212")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from pathlib import Path
import tempfile
import os

torch.manual_seed(0)

# Create a temporary directory structure to simulate ROOT_DATASET
temp_dir = tempfile.mkdtemp()
ROOT_DATASET = temp_dir + '/'

# Create the expected subdirectory
frames_dir = os.path.join(ROOT_DATASET, '20bn-something-something-v2-frames')
os.makedirs(frames_dir, exist_ok=True)

# Create the category.txt file
category_file = os.path.join(ROOT_DATASET, 'category.txt')
with open(category_file, 'w') as f:
    f.write("class1\nclass2\nclass3\n")

# Create train_video_folder.txt
train_file = os.path.join(ROOT_DATASET, 'train_video_folder.txt')
with open(train_file, 'w') as f:
    f.write("video1 1\nvideo2 2\n")

# Create val_video_folder.txt
val_file = os.path.join(ROOT_DATASET, 'val_video_folder.txt')
with open(val_file, 'w') as f:
    f.write("video3 1\nvideo4 2\n")

ARGS = {
    'ROOT_DATASET': ROOT_DATASET
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_state_dict_supernet Not yet run
unverified (harvested, no recorded run)
yucornetto/CAKES/I3D.py
code served (permissive licence) · get_code("1617bf6b8c63d836")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 3 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
return_dataset 1 other paper
1 dated earlier
1909.13130 (2019-09)
Grouped Spatial-Temporal Aggregation for Efficient Action Recognition

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

3D Convolution Neural Networks (CNNs) have been widely applied to 3D scene understanding, such as video analysis and volumetric image recognition. However, 3D networks can easily lead to over-parameterization which incurs expensive computation cost. In this paper, we propose Channelwise Automatic KErnel Shrinking (CAKES), to enable efficient 3D learning by shrinking standard 3D convolutions into a set of economic operations (e.g., 1D, 2D convolutions). Unlike previous methods, CAKES performs channel-wise kernel shrinkage, which enjoys the following benefits: 1) enabling operations deployed in every layer to be heterogeneous, so that they can extract diverse and complementary information to benefit the learning process; and 2) allowing for an efficient and flexible replacement design, which can be generalized to both spatial-temporal and volumetric data. Further, we propose a new search space based on CAKES, so that the replacement configuration can be determined automatically for simplifying 3D networks. CAKES shows superior performance to other methods with similar model size, and it also achieves comparable performance to state-of-theart with much fewer parameters and computational costs on tasks including 3D medical imaging segmentation and video action recognition. Codes and models are available at https: //github.com/yucornetto/CAKES.

For agents

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

get_harvested_code_for_paper("aaai_16433")
get_code_for_paper("aaai_16433")
have("aaai_16433")

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

curl https://syntology.ai/api/ran/aaai_16433.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_16433.svg)](https://syntology.ai/paper/aaai_16433)

Connect an agent — have() is free.