SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.05960 · 2026

Omni-Masked Gradient Descent: Memory-Efficient Optimization via Mask Traversal with Improved Convergence

Tao Ren, Jinyang Jiang, Yijie Peng, Hui Yang, Wan Tian

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 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 1 has no contract verdict; 0 failures of our instrument; 0 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/2603.05960.json · badge: /api/ran/2603.05960.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
yh6-coder/OMGD community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
SIFT Ran
ran on a synthesized input
yh6-coder/OMGD/src/SIFT/sift/sift.py
code served (permissive licence) · get_code("4709330a98ff3d13")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned NoneType. 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

# Create a tiny model to wrap
class TinyModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.layer1 = nn.Linear(4, 6)
        self.layer2 = nn.Linear(6, 3)
        self.output = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.layer1(x)
        x = torch.relu(x)
        x = self.layer2(x)
        x = torch.relu(x)
        x = self.output(x)
        return x

model = TinyModel()

INIT_ARGS = {
    'model': model,
    'sparse_rate': 0.1,
    'sparse_module': ['layer1', 'layer2'],
    'exception': ['output'],
    'grad_acc': 1,
    'gradient_checkpointing': False
}

# SIFT doesn't have a forward or __call__ method that does computation
# It's a wrapper that modifies the model's training behavior
# We'll call a method that exercises the sparse gradient mechanism
# The main computation happens through the model's forward pass and backward hooks

# Create dummy input and target for a forward-backward pass
dummy_input = torch.randn(2, 4)
dummy_target = torch.randn(2, 2)

# The class is used by:
# 1. Constructing it (which registers hooks)
# 2. Running forward pass on the model
# 3. Running backward pass (which triggers the hooks)

# Since there's no single method to call, we define a callable wrapper
METHOD = 'print_trainable_parameters'
ARGS = {}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type NoneType — a NoneType output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a NoneType 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)

Also present in other papers' code

Of 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 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 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.

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

Memory-efficient optimization methods have recently gained increasing attention for scaling full-parameter training of large language models under the GPU-memory bottleneck. Existing approaches either lack clear convergence guarantees, or only achieve the standard O(ϵ -4 ) iteration complexity in the nonconvex settings. We propose Omni-Masked Gradient Descent (OMGD), an optimization method based on mask traversal for memory efficient training, and provide a nonconvex convergence analysis that establishes a strictly improved iteration complexity of Õ(ϵ -3 ) for finding an ϵ-approximate stationary point. Empirically, OMGD is a lightweight, plug-and-play approach that integrates seamlessly into most mainstream optimizers, yielding consistent improvements over competitive baselines in both fine-tuning and pre-training tasks. Our code is available at https://github.com/yh6-coder/OMGD.

For agents

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

get_harvested_code_for_paper("2603.05960")
get_code_for_paper("2603.05960")
have("2603.05960")

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

curl https://syntology.ai/api/ran/2603.05960.json

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

[![Syntology run record](https://syntology.ai/api/ran/2603.05960.svg)](https://syntology.ai/paper/2603.05960)

Connect an agent — have() is free.