SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.09476 · NeurIPS · 2022

SparCL: Sparse Continual Learning on the Edge

Yanzhi Wang, Zifeng Wang, Tong Jian, Stratis Ioannidis, Jennifer Dy, Geng Yuan, Wei Niu, Yifan Gong, Zheng Zhan, Bin Ren

arXiv · PDF · Open in the Atlas

Code that ran

4 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 0 have no contract verdict; 4 failures of our instrument; 2 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/2209.09476.json · badge: /api/ran/2209.09476.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
neu-spiral/SparCL canonical (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
copy not recorded community
not in the paper's declared repositories
2 ran · 1 no recorded run
FunctionStatusWhere it lives
mixup_criterion Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from boone891214/MEST
pointer only; the stored copy get_code serves, from boone891214/MEST, is pointer only (licence: NONE) · get_code("12031ff7668a8502")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

mixup_data Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from boone891214/MEST
pointer only; the stored copy get_code serves, from boone891214/MEST, is pointer only (licence: NONE) · get_code("bd19a75b8bed0114")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

weight_growing Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
neu-spiral/SparCL/prune_utils/retrain.py
code served (permissive licence) · get_code("bd354b97067b64c3")
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 4, 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 numpy as np
import random

torch.manual_seed(0)
np.random.seed(0)
random.seed(0)

class Args:
    def __init__(self):
        self.sp_admm_sparsity_type = 'irregular'
        self.sp_global_magnitude = False

args = Args()
name = 'test_layer'
pruned_weight_np = np.random.randn(4, 3, 3, 3).astype(np.float32)
pruned_weight_np[pruned_weight_np < 0.5] = 0
lower_bound_value = 0.3
upper_bound_value = 0.7
update_init_method = 'zero'
mask_fixed_params = torch.ones(4, 3, 3, 3)

ARGS = {
    'args': args,
    'name': name,
    'pruned_weight_np': pruned_weight_np,
    'lower_bound_value': lower_bound_value,
    'upper_bound_value': upper_bound_value,
    'update_init_method': update_init_method,
    'mask_fixed_params': mask_fixed_params
}
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 Tensor, ndim 4, dtype family float — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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
weight_pruning Ran
ran; our contract draft was wrong, not the code
neu-spiral/SparCL/prune_utils/retrain.py
code served (permissive licence) · get_code("15e574af576cf346")
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 5, 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 numpy as np
import copy
import random
from numpy import linalg as LA
import operator

torch.manual_seed(0)
np.random.seed(0)

class Args:
    def __init__(self):
        self.sp_admm_sparsity_type = 'irregular'
        self.sp_lmd = 0.1

class Configs:
    def __init__(self):
        self.model_name = 'test_model'

# Create a simple weight tensor (4D for conv layer: out_channels, in_channels, kernel_h, kernel_w)
w = torch.randn(4, 3, 3, 3, requires_grad=True)
w.grad = torch.randn(4, 3, 3, 3)

# Create a mask for fixed params (same shape as w)
mask_fixed_params = torch.ones(4, 3, 3, 3)

args = Args()
configs = Configs()
name = 'conv1.weight'
prune_ratio = 0.5

ARGS = {
    'args': args,
    'configs': configs,
    'name': name,
    'w': w,
    'prune_ratio': prune_ratio,
    'mask_fixed_params': mask_fixed_params
}
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 5, 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
SparseTraining Not yet run
unverified (harvested, no recorded run)
neu-spiral/SparCL/prune_utils/retrain.py
code served (permissive licence) · get_code("2074d8ed03574cdf")
How it ran

no invocation recorded.

accuracy Not yet run
unverified (harvested, no recorded run)
this paper's copy was not recorded; identical code first harvested from boone891214/MEST
pointer only; the stored copy get_code serves, from boone891214/MEST, is pointer only (licence: NONE) · get_code("913d82af53065418")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

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, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 1 later) and 1 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
65 dated earlier, 14 later, 1 same year, order not determinable
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
mixup_data 3 other papers
3 dated earlier
2107.00166 (2021-07)
Sanity Checks for Lottery Tickets: Does Your Winning Ticket Really Win the Jackpot?
2110.14032 (2021-10)
MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge
2207.13378 (2022-07)
Identifying Hard Noise in Long-Tailed Sample Distribution
mixup_criterion 1 other paper
1 dated earlier
2110.14032 (2021-10)
MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge
weight_growing
brick of this paper's assembly: update_rule
1 other paper
1 dated earlier · 1 only as an assembly brick
2110.14032 (2021-10)
MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge
weight_pruning
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2110.14032 (2021-10)
MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge

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

Existing work in continual learning (CL) focuses on mitigating catastrophic forgetting, i.e., model performance deterioration on past tasks when learning a new task. However, the training efficiency of a CL system is under-investigated, which limits the real-world application of CL systems under resource-limited scenarios. In this work, we propose a novel framework called Sparse Continual Learning (SparCL), which is the first study that leverages sparsity to enable cost-effective continual learning on edge devices. SparCL achieves both training acceleration and accuracy preservation through the synergy of three aspects: weight sparsity, data efficiency, and gradient sparsity. Specifically, we propose task-aware dynamic masking (TDM) to learn a sparse network throughout the entire CL process, dynamic data removal (DDR) to remove less informative training data, and dynamic gradient masking (DGM) to sparsify the gradient updates. Each of them not only improves efficiency, but also further mitigates catastrophic forgetting. SparCL consistently improves the training efficiency of existing state-of-the-art (SOTA) CL methods by at most 23× less training FLOPs, and, surprisingly, further improves the SOTA accuracy by at most 1.7%. SparCL also outperforms competitive baselines obtained from adapting SOTA sparse training methods to the CL setting in both efficiency and accuracy. We also evaluate the effectiveness of SparCL on a real mobile phone, further indicating the practical potential of our method. Source code will be released.

For agents

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

get_harvested_code_for_paper("2209.09476")
get_code_for_paper("2209.09476")
have("2209.09476")

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

curl https://syntology.ai/api/ran/2209.09476.json

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

[![Syntology run record](https://syntology.ai/api/ran/2209.09476.svg)](https://syntology.ai/paper/2209.09476)

Connect an agent — have() is free.