SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.03282 · 2024

AMES: Asymmetric and Memory-Efficient Similarity Estimation for Instance-level Retrieval

arXiv · PDF · Open in the Atlas

Code that ran

13 of 13 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; 4 honoured the contract we drafted, 1 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/2408.03282.json · badge: /api/ran/2408.03282.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
pavelsuma/ames canonical (classifier label)
classifier basis: census
13 ran · 0 no recorded run
FunctionStatusWhere it lives
calculate_receptive_boxes Ran
ran, honoured the contract we drafted
pavelsuma/ames/extract/extract_cvnet.py
code served (permissive licence) · get_code("aee9e3a7b1b42f70")
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 (6, 4) 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 = {
    "height": 2,
    "width": 3,
    "rf": 5,
    "stride": 2,
    "padding": 1
}
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 (6, 4) — shape (6, 4) 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
calculate_receptive_boxes Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
pavelsuma/ames/extract/extract_dino.py
code served (permissive licence) · get_code("e0ea19cb7ca9d5a7")
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 3, dtype family float, shape (1, 12, 2) 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)

# Analyze the function:
# - imsize: converted to tensor, used with // operator and to divide boxes
#   Should be a sequence of 2 elements (image dimensions)
# - ps: patch size, used for division and multiplication operations
#   Should be a scalar integer or tensor

ARGS = {
    "imsize": [64, 48],  # Image size (height, width) - distinct dimensions
    "ps": 16,            # Patch size - scalar integer
}
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 3, dtype family float, shape (1, 12, 2) — shape (1, 12, 2) 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
color_norm Ran
ran, violated the contract we drafted
pavelsuma/ames/extract/transforms.py
code served (permissive licence) · get_code("d160a205c24a70f9")
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 3, dtype family float, shape (3, 4, 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)

# The function performs per-channel normalization on a CHW format image
# It iterates over the first dimension (channels) and normalizes each channel
# by subtracting mean and dividing by std

# Create a small image tensor in CHW format (Channels, Height, Width)
im = torch.randn(3, 4, 5, dtype=torch.float32)

# Create mean and std tensors with one value per channel
mean = torch.tensor([0.5, 0.4, 0.3], dtype=torch.float32)
std = torch.tensor([0.2, 0.15, 0.1], dtype=torch.float32)

ARGS = {
    "im": im,
    "mean": mean,
    "std": std
}
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 3, dtype family float, shape (3, 4, 5) — shape (3, 4, 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
combine Ran
ran; our contract draft was wrong, not the code
pavelsuma/ames/extract/merge_hdf5.py
code served (permissive licence) · get_code("50ad0cb6c181e867")
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 tuple. 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 tempfile
import os
import h5py
import numpy as np

# Create a temporary directory for the test
temp_dir = tempfile.mkdtemp()

# Create split HDF5 files to simulate the scenario
# Create xa0_local.hdf5
split_file_1 = os.path.join(temp_dir, 'test_desc_xa0_local.hdf5')
with h5py.File(split_file_1, 'w') as f:
    # Create a features dataset with shape (2, 4, 3) - 2 images, 4x3 feature maps
    f.create_dataset('features', data=np.random.randn(2, 4, 3).astype(np.float32))

# Create xa1_local.hdf5
split_file_2 = os.path.join(temp_dir, 'test_desc_xa1_local.hdf5')
with h5py.File(split_file_2, 'w') as f:
    # Create a features dataset with shape (3, 4, 3) - 3 images, 4x3 feature maps
    f.create_dataset('features', data=np.random.randn(3, 4, 3).astype(np.float32))

ARGS = {
    'feat_dir': temp_dir,
    'desc_name': 'test_desc',
    'file_type': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
h5py, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
find_divisors Ran
ran, honoured the contract we drafted
pavelsuma/ames/extract/extract_dino.py
code served (permissive licence) · get_code("04a2af9d93b09d8c")
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 (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 numpy as np

def find_divisors(number):
    divisors = np.arange(1, int(np.sqrt(number)) + 1)
    divisors = divisors[number % divisors == 0]
    return divisors

ARGS = {
    "number": 36
}
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 ndarray, ndim 1, dtype family int, shape (5,) — shape (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
generate_coordinates Ran
ran, honoured the contract we drafted
pavelsuma/ames/extract/extract_cvnet.py
code served (permissive licence) · get_code("f75959fc6c795680")
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 (6, 2) 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 = {
    "h": 2,
    "w": 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 Tensor, ndim 2, dtype family float, shape (6, 2) — shape (6, 2) 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
horizontal_flip Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
pavelsuma/ames/extract/transforms.py
code served (permissive licence) · get_code("6f9305ce535ed9d3")
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 3, dtype family float, shape (3, 4, 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create a small test image in CHW format (3 channels, 4 height, 5 width)
im_chw = np.random.rand(3, 4, 5).astype(np.float32)

# Create a small test image in HWC format (4 height, 5 width, 3 channels)
im_hwc = np.random.rand(4, 5, 3).astype(np.float32)

# Test with CHW format
ARGS = {
    'im': im_chw,
    'p': 0.5,
    'order': 'CHW'
}
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 ndarray, ndim 3, dtype family float, shape (3, 4, 5) — shape (3, 4, 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
load_or_combine Ran
ran, honoured the contract we drafted
pavelsuma/ames/extract/prepare_topk_global.py
code served (permissive licence) · get_code("0412a457791220ef")
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 2, dtype family float, shape (8, 128) 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 os
import tempfile
import h5py
import numpy as np

# Create a temporary directory for testing
temp_dir = tempfile.mkdtemp()

# Create sample HDF5 files to test the function
# Create split files that will be combined
split1_data = np.random.rand(5, 128).astype(np.float32)
split2_data = np.random.rand(3, 128).astype(np.float32)

with h5py.File(os.path.join(temp_dir, 'test_desc_xa0_global.hdf5'), 'w') as f:
    f.create_dataset('features', data=split1_data, dtype=np.float32)

with h5py.File(os.path.join(temp_dir, 'test_desc_xa1_global.hdf5'), 'w') as f:
    f.create_dataset('features', data=split2_data, dtype=np.float32)

ARGS = {
    'feat_dir': temp_dir,
    'desc_name': 'test_desc',
    'file_type': None,
    'global_type': 'global'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
h5py, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (8, 128) — shape (8, 128) 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
load_paths Ran
ran; our contract draft was wrong, not the code
pavelsuma/ames/extract/prepare_topk_global.py
code served (permissive licence) · get_code("1f32bbb8f8d33681")
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 tuple. 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
import tempfile
import os.path as osp

# Create a temporary directory with the required test files
temp_dir = tempfile.mkdtemp()

# Create test_query.txt with sample query lines
query_content = "query_sample_1\nquery_sample_2\nquery_sample_3\n"
with open(osp.join(temp_dir, 'test_query.txt'), 'w') as f:
    f.write(query_content)

# Create test_gallery.txt with sample gallery lines
gallery_content = "gallery_sample_1\ngallery_sample_2\ngallery_sample_3\ngallery_sample_4\n"
with open(osp.join(temp_dir, 'test_gallery.txt'), 'w') as f:
    f.write(gallery_content)

ARGS = {
    'feat_dir': temp_dir
}
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 tuple — a tuple output has no array shape of its own; none is 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
non_maxima_suppression_2d Ran
ran; our contract draft was wrong, not the code
pavelsuma/ames/extract/extract_cvnet.py
code served (permissive licence) · get_code("79817d036d710336")
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 4, dtype family float, shape (2, 1, 4, 3) 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.functional as F

torch.manual_seed(0)

# Create a small heatmap tensor: (batch=2, channels=1, height=4, width=3)
# This is a 4D tensor as expected by F.max_pool2d
heatmap = torch.randn(2, 1, 4, 3)

ARGS = {
    "heatmap": heatmap
}
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 4, dtype family float, shape (2, 1, 4, 3) — shape (2, 1, 4, 3) 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
non_maxima_suppression_2d Ran
ran; our contract draft was wrong, not the code
pavelsuma/ames/extract/extract_dino.py
code served (permissive licence) · get_code("bf31e950a3af0841")
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 4, dtype family float, shape (2, 1, 4, 3) 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.functional as F

torch.manual_seed(0)

# Create a small heatmap tensor
# Shape: (batch=2, channels=1, height=4, width=3)
heatmap = torch.rand(2, 1, 4, 3)

ARGS = {
    "heatmap": heatmap
}
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 4, dtype family float, shape (2, 1, 4, 3) — shape (2, 1, 4, 3) 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
read_imlist Ran
ran; our contract draft was wrong, not the code
pavelsuma/ames/extract/image_dataset.py
code served (permissive licence) · get_code("a93a2142c18e2e88")
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.

import tempfile
import os

# Create a temporary file with some image list content
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("image1.jpg\n")
temp_file.write("image2.jpg\n")
temp_file.write("image3.jpg\n")
temp_file.close()

ARGS = {
    'imlist_fn': temp_file.name
}
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
zero_pad Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
pavelsuma/ames/extract/transforms.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from blackfeather-wang/AdaFocus, is pointer only (licence: NONE) · get_code("8e712c383c677282")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from blackfeather-wang/AdaFocus, 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 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 0 later) and 11 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
zero_pad 3 other papers
3 dated earlier
2105.03245 (2021-05)
Adaptive Focus for Efficient Video Recognition
2209.12892 (2022-09)
Learning to Learn with Generative Models of Neural Network Checkpoints
2308.04008 (2023-08)
Coarse-to-Fine: Learning Compact Discriminative Representation for Single-Stage Image Retrieval
color_norm 1 other paper
1 dated earlier
2209.12892 (2022-09)
Learning to Learn with Generative Models of Neural Network Checkpoints

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

This work investigates the problem of instance-level image retrieval re-ranking with the constraint of memory efficiency, ultimately aiming to limit memory usage to 1KB per image. Departing from the prevalent focus on performance enhancements, this work prioritizes the crucial trade-off between performance and memory requirements. The proposed model uses a transformer-based architecture designed to estimate image-to-image similarity by capturing interactions within and across images based on their local descriptors. A distinctive property of the model is the capability for asymmetric similarity estimation. Database images are represented with a smaller number of descriptors compared to query images, enabling performance improvements without increasing memory consumption. To ensure adaptability across different applications, a universal model is introduced that adjusts to a varying number of local descriptors during the testing phase. Results on standard benchmarks demonstrate the superiority of our approach over both hand-crafted and learned models. In particular, compared with current state-of-the-art methods that overlook their memory footprint, our approach not only attains superior performance but does so with a significantly reduced memory footprint. The code and pretrained models are publicly available at: https://github.com/pavelsuma/ames

For agents

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

get_harvested_code_for_paper("2408.03282")
get_code_for_paper("2408.03282")
have("2408.03282")

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

curl https://syntology.ai/api/ran/2408.03282.json

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

[![Syntology run record](https://syntology.ai/api/ran/2408.03282.svg)](https://syntology.ai/paper/2408.03282)

Connect an agent — have() is free.