SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.08935 · 2020

Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 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 4 have no contract verdict; 2 failures of our instrument; 0 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/2003.08935.json · badge: /api/ran/2003.08935.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
ofsoundof/group_sparsity canonical (classifier label)
classifier basis: census
6 ran · 0 no recorded run
FunctionStatusWhere it lives
default_conv Ran
ran on a synthesized input
ofsoundof/group_sparsity/model/common.py
code served (permissive licence) · get_code("d80e2c9a3e33d17a")
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 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_channels': 3,
    'out_channels': 64,
    'kernel_size': 3,
    'stride': 1,
    'bias': True,
    'groups': 1,
    'args': 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 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
default_linear Ran
ran; our contract draft was wrong, not the code
ofsoundof/group_sparsity/model/common.py
pointer only (licence: MIT); the stored copy get_code serves, from rtao499/SFedHIFI, is pointer only (licence: NONE) · get_code("95b1fbf20859038f")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from rtao499/SFedHIFI, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

gen Ran
ran on a synthesized input
ofsoundof/group_sparsity/model/densenet.py
code served (permissive licence) · get_code("3dd7f9208f8f454c")
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 generator. 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)

# Build a model with Conv2d layers that satisfy the criterion
# Criterion: in_channels >= 24 and kernel_size[0] != 1
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        # Conv2d that satisfies criterion: in_channels=24, kernel_size=(3,3)
        self.conv1 = nn.Conv2d(24, 16, kernel_size=(3, 3))
        # Conv2d that satisfies criterion: in_channels=32, kernel_size=(5,5)
        self.conv2 = nn.Conv2d(32, 16, kernel_size=(5, 5))
        # Conv2d that does NOT satisfy: in_channels=16 < 24
        self.conv3 = nn.Conv2d(16, 8, kernel_size=(3, 3))
        # Conv2d that does NOT satisfy: kernel_size=(1,1)
        self.conv4 = nn.Conv2d(24, 8, kernel_size=(1, 1))

model = SimpleModel()

ARGS = {
    "target": model
}
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 generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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_compress_idx Ran
ran on a synthesized input
ofsoundof/group_sparsity/model_hinge/hinge_densenet.py
code served (permissive licence) · get_code("c1b843d23be6880d")
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 numpy as np
from easydict import EasyDict as edict

torch.manual_seed(0)

# Create small tensors for network_norm and network_pindex
# network_norm is indexed by layer index l, each element is a tensor of channel norms
# network_pindex is indexed by layer index l, each element is indices for remaining channels

# Layer 0: 6 total channels, keep indices [0, 2, 4]
norm_l0 = torch.tensor([1.0, 0.5, 2.0, 0.3, 1.5, 0.8], dtype=torch.float32)
pindex_l0 = torch.tensor([0, 2, 4], dtype=torch.long)

# Layer 1: 5 total channels, keep indices [1, 3]
norm_l1 = torch.tensor([0.9, 1.2, 0.4, 1.8, 0.6], dtype=torch.float32)
pindex_l1 = torch.tensor([1, 3], dtype=torch.long)

# Create network_norm and network_pindex as lists indexed by layer
network_norm = [norm_l0, norm_l1]
network_pindex = [pindex_l0, pindex_l1]

ARGS = {
    'l': 0,
    'network_pindex': network_pindex,
    'network_norm': network_norm,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
easydict, numpy, 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
make_model Ran
ran on a synthesized input
ofsoundof/group_sparsity/model/resnext.py
code served (permissive licence) · get_code("9c81eda4e64437b7")
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 ResNeXt. 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 Args:
    def __init__(self):
        self.depth = 29  # (depth - 2) // 9 = 3, so n = 3 blocks per layer
        self.cardinality = 32
        self.bottleneck_width = 4
        self.data_train = 'CIFAR10'

args_instance = Args()

ARGS = {
    'args': [args_instance],
    'parent': False
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
nopad_conv Ran
ran; our contract draft was wrong, not the code
ofsoundof/group_sparsity/model/common.py
pointer only (licence: MIT); the stored copy get_code serves, from rtao499/SFedHIFI, is pointer only (licence: NONE) · get_code("d5ed39e5a9464bed")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from rtao499/SFedHIFI, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 2 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
default_linear 2 other papers
1 dated earlier, 1 later
1908.08932 (2019-08)
Learning Filter Basis for Convolutional Neural Network Compression
2603.14956 (2026-03)
SFedHIFI: Fire Rate-Based Heterogeneous Information Fusion for Spiking Federated Learning
nopad_conv 2 other papers
1 dated earlier, 1 later
1908.08932 (2019-08)
Learning Filter Basis for Convolutional Neural Network Compression
2603.14956 (2026-03)
SFedHIFI: Fire Rate-Based Heterogeneous Information Fusion for Spiking Federated Learning
gen 1 other paper
1 dated earlier
1908.08932 (2019-08)
Learning Filter Basis for Convolutional Neural Network Compression

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

In this paper, we analyze two popular network compression techniques, i.e. filter pruning and low-rank decomposition, in a unified sense. By simply changing the way the sparsity regularization is enforced, filter pruning and low-rank decomposition can be derived accordingly. This provides another flexible choice for network compression because the techniques complement each other. For example, in popular network architectures with shortcut connections (e.g. ResNet), filter pruning cannot deal with the last convolutional layer in a ResBlock while the low-rank decomposition methods can. In addition, we propose to compress the whole network jointly instead of in a layer-wise manner. Our approach proves its potential as it compares favorably to the state-of-the-art on several benchmarks.

For agents

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

get_harvested_code_for_paper("2003.08935")
get_code_for_paper("2003.08935")
have("2003.08935")

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

curl https://syntology.ai/api/ran/2003.08935.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.08935.svg)](https://syntology.ai/paper/2003.08935)

Connect an agent — have() is free.