SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.05419 · CVPR · 2024

Rethinking Transformers Pre-training for Multi-Spectral Satellite Imagery

Muzammal Naseer, Salman Khan, Shahbaz Fahad, Khan, Hisham Cholakkal, Mubashir Noman, Rao Anwar

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 4 have 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/2403.05419.json · badge: /api/ran/2403.05419.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
techmn/satmae_pp marked official in Papers with Code
Papers with Code archive
4 ran · 0 no recorded run
FunctionStatusWhere it lives
LayerNorm Ran
ran on a synthesized input · constructed an object; no result was computed
techmn/satmae_pp/models_mae_group_channels.py
code served (permissive licence) · get_code("f7cde5f401e45872")
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 an instance of its own class, LayerNorm: the run constructed the object and did not call it, so no result was computed. 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 LayerNorm class takes the following parameters in __init__:
# - normalized_shape: the shape of the normalized dimensions
# - eps: epsilon for numerical stability (default 1e-06)
# - data_format: either 'channels_last' or 'channels_first' (default 'channels_last')

# We'll create arguments for both initialization and forward pass

# normalized_shape should be an integer or tuple representing the last dimension(s) to normalize
normalized_shape = 4

# eps is a small float for numerical stability
eps = 1e-6

# data_format can be 'channels_last' or 'channels_first'
data_format = 'channels_last'

ARGS = {
    'normalized_shape': normalized_shape,
    'eps': eps,
    'data_format': data_format
}
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 LayerNorm — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a LayerNorm 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
MaskedAutoencoderGroupChannelViT Ran
ran on a synthesized input · constructed an object; no result was computed
techmn/satmae_pp/models_mae_group_channels.py
code served (permissive licence) · get_code("2af6dbba2ec2b3de")
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 an instance of its own class, MaskedAutoencoderGroupChannelViT: the run constructed the object and did not call it, so no result was computed. 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
import numpy as np

torch.manual_seed(0)

ARGS = {
    'img_size': 96,
    'patch_size': 8,
    'in_chans': 10,
    'spatial_mask': False,
    'channel_groups': ((0, 1, 2, 6), (3, 4, 5, 7), (8, 9)),
    'channel_embed': 256,
    'embed_dim': 1024,
    'depth': 24,
    'num_heads': 16,
    'decoder_channel_embed': 128,
    'decoder_embed_dim': 512,
    'decoder_depth': 8,
    'decoder_num_heads': 16,
    'mlp_ratio': 4.0,
    'norm_layer': nn.LayerNorm,
    'norm_pix_loss': False,
    'proj_ratio': 4,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, timm, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MaskedAutoencoderGroupChannelViT — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MaskedAutoencoderGroupChannelViT 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
ResidualBlock Ran
ran on a synthesized input · constructed an object; no result was computed
techmn/satmae_pp/models_mae_group_channels.py
code served (permissive licence) · get_code("df04f82fc84f53c2")
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 an instance of its own class, ResidualBlock: the run constructed the object and did not call it, so no result was computed. 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)

class ResidualBlock(nn.Module):

    def __init__(self, channels):
        super(ResidualBlock, self).__init__()
        self.conv1 = nn.Conv2d(channels, channels, kernel_size=3, stride=1, padding=1)
        self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, stride=1, padding=1)
        self.relu = nn.ReLU()

    def forward(self, x):
        residual = x
        out = self.relu(self.conv1(x))
        out = self.conv2(out) * 0.5
        out = out + residual
        return out

# Create a test input tensor: batch size 2, 4 channels, 3x3 spatial dimensions
x = torch.randn(2, 4, 3, 3)

# Create an instance of ResidualBlock with 4 channels
block = ResidualBlock(channels=4)

# Test forward pass
output = block(x)

ARGS = {
    "channels": 4
}
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 ResidualBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ResidualBlock 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
UpsampleBlock Ran
ran on a synthesized input · constructed an object; no result was computed
techmn/satmae_pp/models_mae_group_channels.py
code served (permissive licence) · get_code("79d85f1a4a3e06f9")
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 an instance of its own class, UpsampleBlock: the run constructed the object and did not call it, so no result was computed. 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)

ARGS = {
    'in_channels': 4,
    'out_channels': 3,
}
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 UpsampleBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a UpsampleBlock 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

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 1 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.

FunctionAlso present inEarliest-datedOthers named
LayerNorm
brick of this paper's assembly: helper
14 other papers
7 dated earlier, 2 the same month, 5 later · 9 only as an assembly brick
2110.02178 (2021-10)
MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer
2201.03545 (2022-01)
A ConvNet for the 2020s
2301.00808 (2023-01)
ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders
and 11 more not named here

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

Recent advances in unsupervised learning have demonstrated the ability of large vision models to achieve promising results on downstream tasks by pre-training on large amount of unlabelled data. Such pre-training techniques have also been explored recently in the remote sensing domain due to the availability of large amount of unlabelled data. Different from standard natural image datasets, remote sensing data is acquired from various sensor technologies and exhibit diverse range of scale variations as well as modalities. Existing satellite image pre-training methods either ignore the scale information present in the remote sensing imagery or restrict themselves to use only a single type of data modality. In this paper, we re-visit transformers pre-training and leverage multi-scale information that is effectively utilized with multiple modalities. Our proposed approach, named SatMAE++, performs multiscale pre-training and utilizes convolution based upsampling blocks to reconstruct the image at higher scales making it extensible to include more scales. Compared to existing works, the proposed SatMAE++ with multi-scale pre-training is equally effective for both optical as well as multi-spectral imagery. Extensive experiments on six datasets reveal the merits of proposed contributions, leading to state-of-the-art performance on all datasets. Sat-MAE++ achieves mean average precision (mAP) gain of 2.5% for multi-label classification task on BigEarthNet dataset. Our code and pre-trained models are available at https://github.com/techmn/satmae_pp.

For agents

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

get_harvested_code_for_paper("2403.05419")
get_code_for_paper("2403.05419")
have("2403.05419")

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

curl https://syntology.ai/api/ran/2403.05419.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.05419.svg)](https://syntology.ai/paper/2403.05419)

Connect an agent — have() is free.