SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2111.09451 · 2021

Benchmarking and scaling of deep learning models for land cover image classification

arXiv · PDF · Open in the Atlas

Code that ran

3 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 3 have no contract verdict; 0 failures of our instrument; 2 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/2111.09451.json · badge: /api/ran/2111.09451.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
orion-ai-lab/efficientbigearthnet canonical (classifier label)
classifier basis not recorded on the link
3 ran · 2 no recorded run
FunctionStatusWhere it lives
MlpMixerModel Ran
ran on a synthesized input
orion-ai-lab/efficientbigearthnet/models.py
code served (permissive licence) · get_code("a90db59f5000c624")
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 Functional. 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 tensorflow as tf
import numpy as np

# Set random seeds for reproducibility
np.random.seed(0)
tf.random.set_seed(0)

ARGS = {
    'input_shape': (4, 4, 3),  # height=4, width=4, channels=3
    'num_blocks': 2,  # Number of mixer blocks
    'patch_size': 2,  # Patch size for Conv2D projection
    'hidden_dim': 8,  # Hidden dimension for channel mixing
    'tokens_mlp_dim': 16,  # Token mixer hidden dimension
    'channels_mlp_dim': 16,  # Channel mixer hidden dimension (optional, will default to tokens_mlp_dim if None)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Functional — a Functional output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Functional 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
create_batched_dataset Ran
ran on a synthesized input
orion-ai-lab/efficientbigearthnet/inputs.py
code served (permissive licence) · get_code("305e2195acf9efc5")
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 _PrefetchDataset. 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 tensorflow as tf
import tempfile
import os

# Create temporary TFRecord files for testing
def create_dummy_tfrecord():
    """Create a minimal valid TFRecord file."""
    output_path = tempfile.NamedTemporaryFile(delete=False, suffix='.tfrecord').name
    
    writer = tf.io.TFRecordWriter(output_path)
    
    # Create a single example
    feature = {
        'B02': tf.train.Feature(int64_list=tf.train.Int64List(value=[429] * (120 * 120))),
        'B03': tf.train.Feature(int64_list=tf.train.Int64List(value=[614] * (120 * 120))),
        'B04': tf.train.Feature(int64_list=tf.train.Int64List(value=[590] * (120 * 120))),
        'B05': tf.train.Feature(int64_list=tf.train.Int64List(value=[950] * (60 * 60))),
        'B06': tf.train.Feature(int64_list=tf.train.Int64List(value=[1792] * (60 * 60))),
        'B07': tf.train.Feature(int64_list=tf.train.Int64List(value=[2075] * (60 * 60))),
        'B08': tf.train.Feature(int64_list=tf.train.Int64List(value=[2218] * (120 * 120))),
        'B8A': tf.train.Feature(int64_list=tf.train.Int64List(value=[2266] * (60 * 60))),
        'B11': tf.train.Feature(int64_list=tf.train.Int64List(value=[1594] * (60 * 60))),
        'B12': tf.train.Feature(int64_list=tf.train.Int64List(value=[1009] * (60 * 60))),
        'VV': tf.train.Feature(float_list=tf.train.FloatList(value=[-12.6] * (120 * 120))),
        'VH': tf.train.Feature(float_list=tf.train.FloatList(value=[-19.3] * (120 * 120))),
        'patch_name_s1': tf.train.Feature(bytes_list=tf.train.BytesList(value=[b'patch_s1'])),
        'patch_name_s2': tf.train.Feature(bytes_list=tf.train.BytesList(value=[b'patch_s2'])),
        'original_labels': tf.train.Feature(bytes_list=tf.train.BytesList(value=[b'label1', b'label2'])),
        'original_labels_multi_hot': tf.train.Feature(int64_list=tf.train.Int64List(value=[0] * 43)),
    }
    
    example = tf.train.Example(features=tf.train.Features(feature=feature))
    writer.write(example.SerializeToString())
    writer.close()
    
    return output_path

# Create temporary TFRecord file
tfrecord_file = create_dummy_tfrecord()

ARGS = {
    'TFRecord_paths': [tfrecord_file],
    'batch_size': 2,
    'shuffle_buffer_size': 10,
    'label_type': 'original',
    'num_workers': 1,
    'worker_index': 0,
    'num_parallel_calls': tf.data.experimental.AUTOTUNE,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _PrefetchDataset — a _PrefetchDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _PrefetchDataset 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
eca_module Ran
ran on a synthesized input
orion-ai-lab/efficientbigearthnet/modules.py
code served (permissive licence) · get_code("6bad926586ca04c0")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 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 tensorflow as tf
import numpy as np
import math

# Create a small 4D tensor with shape (batch=2, height=4, width=3, channels=5)
# This represents a typical CNN feature map
inputs_tensor = tf.random.normal(shape=(2, 4, 3, 5), dtype=tf.float32)

ARGS = {
    'inputs_tensor': inputs_tensor,
    'num': None,
    'gamma': 2,
    'b': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 5) — shape (2, 4, 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
coord_module Not yet run
unverified (harvested, no recorded run)
orion-ai-lab/efficientbigearthnet/modules.py
code served (permissive licence) · get_code("a54c0fcde85f305b")
How it ran

no invocation recorded.

ghost_module Not yet run
unverified (harvested, no recorded run)
orion-ai-lab/efficientbigearthnet/modules.py
code served (permissive licence) · get_code("90e46f4b8d8dd1b9")
How it ran

no invocation 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, 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 5 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

The availability of the sheer volume of Copernicus Sentinel-2 imagery has created new opportunities for exploiting deep learning (DL) methods for land use land cover (LULC) image classification. However, an extensive set of benchmark experiments is currently lacking, i.e. DL models tested on the same dataset, with a common and consistent set of metrics, and in the same hardware. In this work, we use the BigEarthNet Sentinel-2 dataset to benchmark for the first time different state-of-the-art DL models for the multi-label, multi-class LULC image classification problem, contributing with an exhaustive zoo of 60 trained models. Our benchmark includes standard CNNs, as well as non-convolutional methods. We put to the test EfficientNets and Wide Residual Networks (WRN) architectures, and leverage classification accuracy, training time and inference rate. Furthermore, we propose to use the EfficientNet framework for the compound scaling of a lightweight WRN. Enhanced with an Efficient Channel Attention mechanism, our scaled lightweight model emerged as the new state-of-the-art. It achieves 4.5% higher averaged F-Score classification accuracy for all 19 LULC classes compared to a standard ResNet50 baseline model, with an order of magnitude less trainable parameters. We provide access to all trained models, along with our code for distributed training on multiple GPU nodes. This model zoo of pre-trained encoders can be used for transfer learning and rapid prototyping in different remote sensing tasks that use Sentinel-2 data, instead of exploiting backbone models trained with data from a different domain, e.g., from ImageNet. We validate their suitability for transfer learning in different datasets of diverse volumes. Our top-performing WRN achieves state-of-the-art performance (71.1% F-Score) on the SEN12MS dataset while being exposed to only a small fraction of the training dataset.

For agents

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

get_harvested_code_for_paper("2111.09451")
get_code_for_paper("2111.09451")
have("2111.09451")

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

curl https://syntology.ai/api/ran/2111.09451.json

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

[![Syntology run record](https://syntology.ai/api/ran/2111.09451.svg)](https://syntology.ai/paper/2111.09451)

Connect an agent — have() is free.