SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2308.00261 · ICCV · 2023

Improving Pixel-based MIM by Reducing Wasted Modeling Capability

Yuan Liu, Dahua Lin, Kai Chen, Jiacheng Chen, Songyang Zhang, Zhaohui Yu

arXiv · PDF · Open in the Atlas

Code that ran

2 of 3 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; 2 failures of our instrument; 1 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/2308.00261.json · badge: /api/ran/2308.00261.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
open-mmlab/mmpretrain canonical (classifier label)
classifier basis: census
2 ran · 1 no recorded run
FunctionStatusWhere it lives
filter_transforms Ran
ran; our contract draft was wrong, not the code
open-mmlab/mmpretrain/mmpretrain/apis/multimodal_retrieval.py
code served (permissive licence) · get_code("0df7d8afd46175e7")
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.

from copy import deepcopy


class Transform1:
    """Transform that accesses 'image' key - will succeed"""
    def __call__(self, data_info):
        # This transform requires 'image' key
        _ = data_info['image']
        data_info['processed_image'] = True
        return data_info


class Transform2:
    """Transform that accesses 'missing_key' - will raise KeyError"""
    def __call__(self, data_info):
        # This transform requires a key that doesn't exist
        _ = data_info['missing_key']
        data_info['result'] = True
        return data_info


class Transform3:
    """Transform that accesses 'label' key - will succeed after Transform1"""
    def __call__(self, data_info):
        # This transform requires 'label' key which exists
        _ = data_info['label']
        data_info['processed_label'] = True
        return data_info


transforms = [
    Transform1(),
    Transform2(),
    Transform3(),
]

data_info = {
    'image': [1, 2, 3, 4],
    'label': 'cat',
}

ARGS = {
    'transforms': transforms,
    'data_info': data_info,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
init_empty_weights Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
open-mmlab/mmpretrain/mmpretrain/apis/utils.py
code served (permissive licence) · get_code("0a9c732c8627db37")
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 _GeneratorContextManager. 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)

# The function init_empty_weights is a context manager that takes one parameter
ARGS = {
    "include_buffers": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
mmengine, torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _GeneratorContextManager — a _GeneratorContextManager output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _GeneratorContextManager 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
compute_module_sizes Not yet run
unverified (harvested, no recorded run)
open-mmlab/mmpretrain/mmpretrain/apis/utils.py
code served (permissive licence) · get_code("67ea1de226fa0b01")
How it ran

no invocation recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

There has been significant progress in Masked Image Modeling (MIM). Existing MIM methods can be broadly categorized into two groups based on the reconstruction target: pixel-based and tokenizer-based approaches. The former offers a simpler pipeline and lower computational cost, but it is known to be biased toward high-frequency details. In this paper, we provide a set of empirical studies to confirm this limitation of pixel-based MIM and propose a new method that explicitly utilizes low-level features from shallow layers to aid pixel reconstruction. By incorporating this design into our base method, MAE, we reduce the wasted modeling capability of pixel-based MIM, improving its convergence and achieving non-trivial improvements across various downstream tasks. To the best of our knowledge, we are the first to systematically investigate multilevel feature fusion for isotropic architectures like the standard Vision Transformer (ViT). Notably, when applied to a smaller model (e.g., ViT-S), our method yields significant performance gains, such as 1.2% on fine-tuning, 2.8% on linear probing, and 2.6% on semantic segmentation. Code and models are available in MMPretrain 1 .

For agents

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

get_harvested_code_for_paper("2308.00261")
get_code_for_paper("2308.00261")
have("2308.00261")

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

curl https://syntology.ai/api/ran/2308.00261.json

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

[![Syntology run record](https://syntology.ai/api/ran/2308.00261.svg)](https://syntology.ai/paper/2308.00261)

Connect an agent — have() is free.