SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.09081 · 2020

Progressive Skeletonization: Trimming more fat from a network at initialization

arXiv · PDF · Open in the Atlas

Code that ran

5 of 14 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 3 have no contract verdict; 2 failures of our instrument; 9 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/2006.09081.json · badge: /api/ran/2006.09081.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
naver/force canonical (classifier label)
classifier basis: census
5 ran · 9 no recorded run
FunctionStatusWhere it lives
apply_prune_mask Ran
ran on a synthesized input
naver/force/pruning/mask_networks.py
code served (permissive licence) · get_code("edda8bc5fb1dfa11")
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)

# Create a simple network with Conv2d and Linear layers
class SimpleNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 4, kernel_size=3)
        self.linear1 = nn.Linear(4, 2)
    
    def forward(self, x):
        return x

net = SimpleNet()

# Create keep_masks for each prunable layer (Conv2d and Linear)
# Conv2d weight shape: (out_channels, in_channels, kernel_h, kernel_w) = (4, 3, 3, 3)
conv_mask = torch.ones(4, 3, 3, 3)
conv_mask[0, 0, 0, 0] = 0.0  # Mark one parameter for pruning

# Linear weight shape: (out_features, in_features) = (2, 4)
linear_mask = torch.ones(2, 4)
linear_mask[0, 0] = 0.0  # Mark one parameter for pruning

keep_masks = [conv_mask, linear_mask]

ARGS = {
    "net": net,
    "keep_masks": keep_masks,
    "apply_hooks": True
}
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_mask Ran
ran on a synthesized input
naver/force/pruning/pruning_algos.py
code served (permissive licence) · get_code("3f216712b9e01cbb")
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. 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)

# Construct saliency: a list of tensors with different shapes
# The function flattens and concatenates them, so we need multiple tensors
saliency = [
    torch.randn(2, 3),      # 6 parameters
    torch.randn(4),         # 4 parameters
    torch.randn(2, 2),      # 4 parameters
]
# Total: 14 parameters

# pruning_factor: float between 0 and 1 representing sparsity
# If pruning_factor = 0.5, we keep 50% of parameters
pruning_factor = 0.5

ARGS = {
    "saliency": saliency,
    "pruning_factor": pruning_factor,
}
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 — a list 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
get_tiny_imagenet_train_valid_loader Ran
ran on a synthesized input
naver/force/experiments/datasets.py
code served (permissive licence) · get_code("e5ca87502d56226e")
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
import torchvision
import torchvision.transforms as transforms
from torch.utils.data import DataLoader, Dataset
import tempfile
import os
from PIL import Image
import numpy as np

torch.manual_seed(0)

# Create a minimal stand-in dataset since we can't access the real tiny-imagenet
class FakeTinyImageNetDataset(Dataset):
    def __init__(self, size=10, transform=None):
        self.size = size
        self.transform = transform
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        # Create a fake 64x64 RGB image
        img_array = np.random.randint(0, 256, (64, 64, 3), dtype=np.uint8)
        img = Image.fromarray(img_array)
        label = idx % 200  # Tiny ImageNet has 200 classes
        
        if self.transform:
            img = self.transform(img)
        
        return img, label

# Patch torchvision.datasets.ImageFolder to use our fake dataset
original_imagefolder = torchvision.datasets.ImageFolder

def patched_imagefolder(root, transform=None):
    return FakeTinyImageNetDataset(size=10, transform=transform)

torchvision.datasets.ImageFolder = patched_imagefolder

# Define arguments for the function
ARGS = {
    'batch_size': 2,
    'augment': True,
    'shuffle': True,
    'num_workers': 0,  # Use 0 for single-process to avoid issues
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch, torchvision — 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
resnet18 Ran
ran; our contract draft was wrong, not the code
naver/force/experiments/models.py
code served (permissive licence) · get_code("0327078528100688")
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)

# Define ARGS dict with all parameters for resnet18
ARGS = {
    'temp': 1.0,
    'num_classes': 10,
    'in_planes': 64,
    'stable_resnet': 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 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
resnet34 Ran
ran; our contract draft was wrong, not the code
naver/force/experiments/models.py
code served (permissive licence) · get_code("67cc5d9a1cd2e972")
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)

ARGS = {
    'temp': 1.0
}
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 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
accuracy Not yet run
unverified (harvested, no recorded run)
naver/force/train_imagenet.py
pointer only (licence: MIT); the stored copy get_code serves, from zbh2047/clipping-algorithms, is pointer only (licence: NONE) · get_code("9b8289076669fe4f")
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 zbh2047/clipping-algorithms, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

get_average_gradients Not yet run
unverified (harvested, no recorded run)
naver/force/pruning/pruning_algos.py
code served (permissive licence) · get_code("c2fe2e0433a97c2a")
How it ran

no invocation recorded.

get_average_saliencies Not yet run
unverified (harvested, no recorded run)
naver/force/pruning/pruning_algos.py
code served (permissive licence) · get_code("a3aae03238435730")
How it ran

no invocation recorded.

get_cifar_test_loader Not yet run
unverified (harvested, no recorded run)
naver/force/experiments/datasets.py
code served (permissive licence) · get_code("b6cbcb7ccb3e0a51")
How it ran

no invocation recorded.

get_cifar_train_valid_loader Not yet run
unverified (harvested, no recorded run)
naver/force/experiments/datasets.py
code served (permissive licence) · get_code("24854d716b08e3a5")
How it ran

no invocation recorded.

resnet50 Not yet run
unverified (harvested, no recorded run)
naver/force/experiments/models.py
code served (permissive licence) · get_code("881b8281d25db1b8")
How it ran

no invocation recorded.

validate Not yet run
unverified (harvested, no recorded run)
naver/force/train_imagenet.py
code served (permissive licence) · get_code("62eead0439dcfbc8")
How it ran

no invocation recorded.

vgg_cifar_experiment Not yet run
unverified (harvested, no recorded run)
naver/force/experiments/experiments.py
code served (permissive licence) · get_code("1cbecd8ff404af70")
How it ran

no invocation recorded.

vgg_tiny_imagenet_experiment Not yet run
unverified (harvested, no recorded run)
naver/force/experiments/experiments.py
code served (permissive licence) · get_code("472a885200c16eb7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 1 the same month, 4 later) and 10 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 80 other papers
39 dated earlier, 2 the same month, 39 later
1307.0414 (2013-07)
Challenges in Representation Learning: A report on three machine learning contests
1412.6553 (2014-12)
Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decomposition
1503.02531 (2015-03)
Distilling the Knowledge in a Neural Network
and 77 more not named here
resnet34 3 other papers
1 dated earlier, 2 later
2002.09437 (2020-02)
Calibrating Deep Neural Networks using Focal Loss
Liu_The_Devil_Is_in_the_Margin_Margin-Based_Label_Smoothing_for_CVPR_2022_paper (2022 (year only))
The Devil is in the Margin: Margin-based Label Smoothing for Network Calibration
resnet18 2 other papers
1 dated earlier, 1 later
2002.09437 (2020-02)
Calibrating Deep Neural Networks using Focal Loss
resnet50 2 other papers
1 dated earlier, 1 later
2002.09437 (2020-02)
Calibrating Deep Neural Networks using Focal Loss
Liu_The_Devil_Is_in_the_Margin_Margin-Based_Label_Smoothing_for_CVPR_2022_paper (2022 (year only))
The Devil is in the Margin: Margin-based Label Smoothing for Network Calibration

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

Recent studies have shown that skeletonization (pruning parameters) of networks \textit{at initialization} provides all the practical benefits of sparsity both at inference and training time, while only marginally degrading their performance. However, we observe that beyond a certain level of sparsity (approx $95\%$), these approaches fail to preserve the network performance, and to our surprise, in many cases perform even worse than trivial random pruning. To this end, we propose an objective to find a skeletonized network with maximum {\em foresight connection sensitivity} (FORCE) whereby the trainability, in terms of connection sensitivity, of a pruned network is taken into consideration. We then propose two approximate procedures to maximize our objective (1) Iterative SNIP: allows parameters that were unimportant at earlier stages of skeletonization to become important at later stages; and (2) FORCE: iterative process that allows exploration by allowing already pruned parameters to resurrect at later stages of skeletonization. Empirical analyses on a large suite of experiments show that our approach, while providing at least as good a performance as other recent approaches on moderate pruning levels, provides remarkably improved performance on higher pruning levels (could remove up to $99.5\%$ parameters while keeping the networks trainable). Code can be found in https://github.com/naver/force.

For agents

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

get_harvested_code_for_paper("2006.09081")
get_code_for_paper("2006.09081")
have("2006.09081")

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

curl https://syntology.ai/api/ran/2006.09081.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.09081.svg)](https://syntology.ai/paper/2006.09081)

Connect an agent — have() is free.