SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.01230 · 2025

Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descent

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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 2 have no contract verdict; 3 failures 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/2501.01230.json · badge: /api/ran/2501.01230.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
walkerworldpeace/doge canonical (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
get_attr Ran
ran on a synthesized input
walkerworldpeace/doge/fusion_bench/models/wrappers/layer_wise_fusion_doge_ta.py
code served (permissive licence) · get_code("7076dab2ff2fb2cd")
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 int, ndim 0, dtype family int, shape () 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)

# Create a nested object structure with attributes
class InnerModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.value = 42

class MiddleModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.inner = InnerModule()

class OuterModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.middle = MiddleModule()

# Create an object with nested attributes
obj = OuterModule()

# Test case 1: Single level attribute access
obj_single = type('obj', (), {'attr1': 100})()

# Test case 2: Multi-level attribute access
obj_nested = type('obj', (), {
    'level1': type('obj', (), {
        'level2': type('obj', (), {
            'level3': 'target_value'
        })()
    })()
})()

# Arguments for single-level access
ARGS = {
    "obj": obj_single,
    "names": ["attr1"]
}
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 int, ndim 0, dtype family int, shape () — shape () 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
get_layer_wise_weights Ran
ran on a synthesized input
walkerworldpeace/doge/fusion_bench/models/wrappers/layer_wise_fusion_doge_ta.py
code served (permissive licence) · get_code("694c0fef8290f250")
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 2, dtype family float, shape (3, 5) 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)

ARGS = {
    "num_models": 3,
    "num_layers": 5,
    "init_values": 0.5,
    "dtype": torch.float32,
}
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 Tensor, ndim 2, dtype family float, shape (3, 5) — shape (3, 5) 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
recombine_modeldict Ran
ran; our contract draft was wrong, not the code
walkerworldpeace/doge/fusion_bench/method/model_recombination.py
pointer only (licence: MIT); the stored copy get_code serves, from yangjinluan/dam, is pointer only (licence: NONE) · get_code("6866d2a6d69a1ecd")
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 yangjinluan/dam, 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.

recombine_modellist Ran
ran; our contract draft was wrong, not the code
walkerworldpeace/doge/fusion_bench/method/model_recombination.py
pointer only (licence: MIT); the stored copy get_code serves, from yangjinluan/dam, is pointer only (licence: NONE) · get_code("b5d44b089a2cdd0b")
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 yangjinluan/dam, 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.

recombine_state_dict Ran
ran; our contract draft was wrong, not the code
walkerworldpeace/doge/fusion_bench/method/model_recombination.py
pointer only (licence: MIT); the stored copy get_code serves, from yangjinluan/dam, is pointer only (licence: NONE) · get_code("3e6d09646a3092dc")
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 yangjinluan/dam, 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 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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, 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.

FunctionAlso present inEarliest-datedOthers named
recombine_modeldict 2 other papers
2 dated earlier
2406.03280 (2024-06)
FusionBench: A Unified Library and Comprehensive Benchmark for Deep Model Fusion
2410.13910 (2024-10)
Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace
recombine_modellist 2 other papers
2 dated earlier
2406.03280 (2024-06)
FusionBench: A Unified Library and Comprehensive Benchmark for Deep Model Fusion
2410.13910 (2024-10)
Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace
recombine_state_dict 2 other papers
2 dated earlier
2406.03280 (2024-06)
FusionBench: A Unified Library and Comprehensive Benchmark for Deep Model Fusion
2410.13910 (2024-10)
Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace

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

Merging multiple expert models offers a promising approach for performing multi-task learning without accessing their original data. Existing methods attempt to alleviate task conflicts by sparsifying task vectors or promoting orthogonality among them. However, they overlook the fundamental target of model merging: the merged model performs as closely as possible to task-specific models on respective tasks. We find these methods inevitably discard task-specific information that, while causing conflicts, is crucial for performance. Based on our findings, we frame model merging as a constrained optimization problem ($\textit{i.e.}$, minimizing the gap between the merged model and individual models, subject to the constraint of retaining shared knowledge) and solve it via adaptive projective gradient descent. Specifically, we align the merged model with individual models by decomposing and reconstituting the loss function, alleviating conflicts through $\textit{data-free}$ optimization of task vectors. To retain shared knowledge, we optimize this objective by projecting gradients within a $\textit{shared subspace}$ spanning all tasks. Moreover, we view merging coefficients as adaptive learning rates and propose a task-aware, training-free strategy. Experiments show that our plug-and-play approach consistently outperforms previous methods, achieving state-of-the-art results across diverse architectures and tasks in both vision and NLP domains.

For agents

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

get_harvested_code_for_paper("2501.01230")
get_code_for_paper("2501.01230")
have("2501.01230")

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

curl https://syntology.ai/api/ran/2501.01230.json

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

[![Syntology run record](https://syntology.ai/api/ran/2501.01230.svg)](https://syntology.ai/paper/2501.01230)

Connect an agent — have() is free.