SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2401.17868 · ICLR · 2024

Convolution Meets LoRA: Parameter Efficient Finetuning for Segment Anything Model

Chun Yuan, Tong He, Zhiqiang Tang, Zihan Zhong

arXiv · PDF · Open in the Atlas

Code that ran

8 of 8 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; 8 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/2401.17868.json · badge: /api/ran/2401.17868.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
autogluon/autogluon canonical (classifier label)
classifier basis: census
8 ran · 0 no recorded run
FunctionStatusWhere it lives
bin_column Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autogluon/autogluon/features/src/autogluon/features/binning.py
code served (permissive licence) · get_code("2607590bf0e7883b")
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 ndarray, ndim 1, dtype family int, shape (6,) 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 numpy as np
from pandas import Series

# Create a simple Series with numeric values
series_data = Series([1.5, 2.3, 3.7, 4.1, 5.9, 6.2])

# Define bins for digitize
bins = np.array([2.0, 4.0, 6.0])

# Define dtype for the output
dtype = np.int32

ARGS = {
    "series": series_data,
    "bins": bins,
    "dtype": dtype,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (6,) — shape (6,) 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
clip_and_astype Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autogluon/autogluon/features/src/autogluon/features/utils.py
code served (permissive licence) · get_code("0cce82f5a2c83f58")
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 DataFrame. 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 numpy as np
from pandas import DataFrame
import pandas as pd

# Create a small test DataFrame with numeric values
df = DataFrame({
    'col1': [10, 150, 300, 50],
    'col2': [5, 200, 100, 255],
    'col3': [0, 128, 256, 64]
})

ARGS = {
    'df': df,
    'columns': ['col1', 'col2'],
    'clip_min': 0,
    'clip_max': 255,
    'dtype': 'uint8'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 and dtype family int 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
generate_bins Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autogluon/autogluon/features/src/autogluon/features/binning.py
code served (permissive licence) · get_code("e5c9afc2fef8368c")
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 dict. 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 numpy as np
import pandas as pd
from pandas import DataFrame

# Create a small DataFrame with numeric columns to bin
np.random.seed(0)
X_features = DataFrame({
    'feature_1': np.random.uniform(0, 100, 50),
    'feature_2': np.random.uniform(0, 50, 50),
    'feature_3': np.random.uniform(10, 200, 50),
})

# Define arguments
ARGS = {
    'X_features': X_features,
    'features_to_bin': ['feature_1', 'feature_2'],
    'ideal_bins': 10,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
get_bins Ran
ran; our contract draft was wrong, not the code
autogluon/autogluon/features/src/autogluon/features/binning.py
code served (permissive licence) · get_code("9ab1251987476bd3")
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 IntervalIndex. 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 numpy as np
from pandas import DataFrame, IntervalIndex, Series

# Create a simple Series with numeric values
bins_data = Series([0.0, 1.5, 2.3, 3.7, 5.0, 5.0])

# Create bin_index as a list of indices to select from the Series
bin_index = [0, 1, 3, 4, 5]

# bin_epsilon as a small float value
bin_epsilon = 0.01

ARGS = {
    "bins": bins_data,
    "bin_index": bin_index,
    "bin_epsilon": bin_epsilon,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type IntervalIndex — a IntervalIndex output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a IntervalIndex output the recorded ndim 1 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
get_ngram_freq Ran
ran; our contract draft was wrong, not the code
autogluon/autogluon/features/src/autogluon/features/vectorizers.py
code served (permissive licence) · get_code("9614ea507fa59bf0")
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 dict. 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 numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer

# Create a simple vectorizer
vectorizer = TfidfVectorizer(max_features=4, ngram_range=(1, 2))

# Create some sample text data and fit the vectorizer
sample_texts = ["hello world", "hello there", "world of python"]
vectorizer.fit(sample_texts)

# Transform the texts to get a sparse matrix
transform_matrix = vectorizer.transform(sample_texts)

# Convert to dense matrix for the function
# The function expects transform_matrix.sum(axis=0) to work and return a 2D array-like
# with indexing [0] to get a 1D list-like result
transform_matrix_dense = transform_matrix.toarray()
# Convert to a matrix-like object that has sum and indexing similar to sparse matrix output
transform_matrix_result = np.matrix(transform_matrix_dense)

ARGS = {
    "vectorizer": vectorizer,
    "transform_matrix": transform_matrix_result
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
get_smallest_valid_dtype_int Ran
ran; our contract draft was wrong, not the code
autogluon/autogluon/features/src/autogluon/features/utils.py
code served (permissive licence) · get_code("7043d0d583fd5160")
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 type. 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 numpy as np

# Test case 1: small positive integers (should use uint8)
min_val_1 = 0
max_val_1 = 100

# Test case 2: negative integers (should use int16 or smaller)
min_val_2 = -1000
max_val_2 = 1000

# Test case 3: large positive integers (should use uint32)
min_val_3 = 0
max_val_3 = 100000

# Create ARGS dict with parameters for the function
ARGS = {
    "min_val": min_val_1,
    "max_val": max_val_1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type type — a type output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a type 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
is_useless_feature Ran
ran; our contract draft was wrong, not the code
autogluon/autogluon/features/src/autogluon/features/utils.py
code served (permissive licence) · get_code("b38bf754376eaa21")
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 bool, ndim 0, dtype family float, 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 numpy as np
from pandas import DataFrame, Series

# Create a useless feature (all same values)
useless_feature = Series([5, 5, 5, 5], name='useless')

ARGS = {
    'X': useless_feature
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, 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
update_version Ran
ran; our contract draft was wrong, not the code
autogluon/autogluon/core/src/autogluon/core/_setup_utils.py
code served (permissive licence) · get_code("da258afb34969824")
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 str. 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 os

# Mock the environment for testing
os.environ.pop('RELEASE', None)
os.environ.pop('AUTOGLUON_VERSION_SUFFIX', None)

# Test case 1: Normal dev version (no RELEASE, no suffix env var)
ARGS = {
    'version': '1.5.1'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 8 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 0 the same month, 0 later) and 0 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
bin_column 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
clip_and_astype 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
generate_bins 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
get_bins 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
get_ngram_freq 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
get_smallest_valid_dtype_int 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
is_useless_feature 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications
update_version 1 other paper
1 dated earlier
2311.02971 (2023-11)
TabRepo: A Large Scale Repository of Tabular Model Evaluations and its AutoML Applications

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 Segment Anything Model (SAM) stands as a foundational framework for image segmentation. While it exhibits remarkable zero-shot generalization in typical scenarios, its advantage diminishes when applied to specialized domains like medical imagery and remote sensing. To address this limitation, this paper introduces Conv-LoRA, a simple yet effective parameter-efficient fine-tuning approach. By integrating ultra-lightweight convolutional parameters into Low-Rank Adaptation (LoRA), Conv-LoRA can inject image-related inductive biases into the plain ViT encoder, further reinforcing SAM's local prior assumption. Notably, Conv-LoRA not only preserves SAM's extensive segmentation knowledge but also revives its capacity of learning high-level image semantics, which is constrained by SAM's foreground-background segmentation pretraining. Comprehensive experimentation across diverse benchmarks spanning multiple domains underscores Conv-LoRA's superiority in adapting SAM to real-world semantic segmentation tasks. 1

For agents

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

get_harvested_code_for_paper("2401.17868")
get_code_for_paper("2401.17868")
have("2401.17868")

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

curl https://syntology.ai/api/ran/2401.17868.json

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

[![Syntology run record](https://syntology.ai/api/ran/2401.17868.svg)](https://syntology.ai/paper/2401.17868)

Connect an agent — have() is free.