SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2105.14636 · 2021

LEAP: Learnable Pruning for Transformer-based Models

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 0 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/2105.14636.json · badge: /api/ran/2105.14636.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
yaozhewei/mlpruning canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
regularization Ran
ran, honoured the contract we drafted
yaozhewei/mlpruning/training/masked_run_glue.py
code served (permissive licence) · get_code("729d19d073f5c736")
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) and returned Tensor, ndim 1, dtype family float, shape (1,) 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 torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        # Create threshold parameters for 4 thresholds per layer × 12 layers = 48 parameters
        # Each layer has: attn, o_matrix, fc1, fc2
        self.threshold_attn = nn.ParameterList([nn.Parameter(torch.randn(1)) for _ in range(12)])
        self.threshold_o_matrix = nn.ParameterList([nn.Parameter(torch.randn(1)) for _ in range(12)])
        self.threshold_fc1 = nn.ParameterList([nn.Parameter(torch.randn(1)) for _ in range(12)])
        self.threshold_fc2 = nn.ParameterList([nn.Parameter(torch.randn(1)) for _ in range(12)])
    
    def forward(self, x):
        return x

model = SimpleModel()

# Verify the model has the right structure
threshold_list = []
for name, param in model.named_parameters():
    if 'threshold' in name:
        threshold_list.append(torch.sigmoid(param))

# Should have 48 threshold parameters (4 per layer × 12 layers)
assert len(threshold_list) == 48, f"Expected 48 threshold parameters, got {len(threshold_list)}"

threshold_value = 0.5

ARGS = {
    "model": model,
    "threshold": threshold_value
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 1, dtype family float, shape (1,) — shape (1,) 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
remain_param_compute Ran
ran; our contract draft was wrong, not the code
yaozhewei/mlpruning/training/masked_run_glue.py
code served (permissive licence) · get_code("93f427a13e700b5d")
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) and returned Tensor, ndim 1, dtype family float, shape (4,) 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 torch

torch.manual_seed(0)

# Analyze the function:
# - threshold_list is unpacked into 4 tensors: attn, o_matrix, fc1, fc2
# - Each tensor is used with torch.max() and .type() method
# - .type() is called on fc1 to get its dtype
# - The tensors are compared with scalar tensors and multiplied together

# Create small tensors with distinct shapes to allow probing
attn = torch.tensor([0.1, 0.2, 0.15, 0.05], dtype=torch.float32)
o_matrix = torch.tensor([0.001, 0.002, 0.0015, 0.0005], dtype=torch.float32)
fc1 = torch.tensor([0.0003, 0.0004, 0.00035, 0.00025], dtype=torch.float32)
fc2 = torch.tensor([0.001, 0.0015, 0.0012, 0.0008], dtype=torch.float32)

threshold_list = [attn, o_matrix, fc1, fc2]

ARGS = {
    "threshold_list": threshold_list
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 1, dtype family float, shape (4,) — shape (4,) 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

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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.

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

Pruning is an effective method to reduce the memory footprint and computational cost associated with large natural language processing models. However, current pruning algorithms either only focus on one pruning category, e.g., structured pruning and unstructured, or need extensive hyperparameter tuning in order to get reasonable accuracy performance. To address these challenges, we propose LEArnable Pruning (LEAP), an effective method to gradually prune the model based on thresholds learned by gradient descent. Different than previous learnable pruning methods, which utilize $L_0$ or $L_1$ penalty to indirectly affect the final pruning ratio, LEAP introduces a novel regularization function, that directly interacts with the preset target pruning ratio. Moreover, in order to reduce hyperparameter tuning, a novel adaptive regularization coefficient is deployed to control the regularization penalty adaptively. With the new regularization term and its associated adaptive regularization coefficient, LEAP is able to be applied for different pruning granularity, including unstructured pruning, structured pruning, and hybrid pruning, with minimal hyperparameter tuning. We apply LEAP for BERT models on QQP/MNLI/SQuAD for different pruning settings. Our result shows that for all datasets, pruning granularity, and pruning ratios, LEAP achieves on-par or better results as compared to previous heavily hand-tuned methods.

For agents

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

get_harvested_code_for_paper("2105.14636")
get_code_for_paper("2105.14636")
have("2105.14636")

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

curl https://syntology.ai/api/ran/2105.14636.json

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

[![Syntology run record](https://syntology.ai/api/ran/2105.14636.svg)](https://syntology.ai/paper/2105.14636)

Connect an agent — have() is free.