SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.03078 · 2024

ATAT: Astronomical Transformer for time series And Tabular data

arXiv · PDF · Open in the Atlas

Code that ran

12 of 15 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, 2 violated it and 0 have no contract verdict; 10 failures of our instrument; 3 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/2405.03078.json · badge: /api/ran/2405.03078.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
alercebroker/atat canonical (classifier label)
classifier basis: census
12 ran · 3 no recorded run
FunctionStatusWhere it lives
add_name_config Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/layers/encoder_ellastic.py
code served (permissive licence) · get_code("0beb91278b381f19")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# The function add_name_config takes a config dict and reads:
# - config['which_encoder']
# - config['which_decoder']
# - config['head_dim']
# - config['num_heads']

# Build a minimal config dict with all required keys
config = {
    'which_encoder': 'mha',
    'which_decoder': 'mha',
    'head_dim': 64,
    'num_heads': 8
}

ARGS = {
    'config': config
}
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 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
add_name_config Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/layers/time_modulator.py
code served (permissive licence) · get_code("26385af78390d8dc")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Analyze what the function needs:
# - config: a dict with keys 'which_encoder', 'which_decoder', and 'M'
# - The function checks if 'which_encoder' or 'which_decoder' equals 'mha'
# - If true, it accesses config['M'] and formats it with %d

ARGS = {
    'config': {
        'which_encoder': 'mha',
        'which_decoder': 'transformer',
        'M': 8
    }
}
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 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
add_sample_parser Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/layers/encoder_ellastic.py
code served (permissive licence) · get_code("af12e0efd0ca5c63")
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 MockParser. 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 argparse
import torch
import torch.nn as nn

torch.manual_seed(0)

class MockParser:
    def __init__(self):
        self.arguments = {}
    
    def add_argument(self, name, **kwargs):
        self.arguments[name] = kwargs

parser = MockParser()

ARGS = {
    'parser': parser
}
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 MockParser — a MockParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockParser 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
add_sample_parser Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/layers/mha.py
code served (permissive licence) · get_code("a398f59f3bc6f12f")
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 ArgumentParser. 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 argparse
import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal argument parser
parser = argparse.ArgumentParser()

# The function add_sample_parser adds arguments to an existing parser
# It takes a parser object and returns it after adding arguments
# No other parameters are used in the function body

ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
add_sample_parser Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/layers/time_modulator.py
code served (permissive licence) · get_code("88a5c065620d59c9")
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 ArgumentParser. 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 argparse
import torch
import torch.nn as nn

torch.manual_seed(0)

# The function add_sample_parser takes a parser object and adds arguments to it
# It doesn't actually use the parser beyond calling add_argument on it
# So we need to construct a valid ArgumentParser instance

parser = argparse.ArgumentParser(description='Test parser')

ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
best_model_dir_and_step Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/loading.py
code served (permissive licence) · get_code("3dc3c18308b9955e")
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, 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 os
import tempfile
import shutil

# Create a temporary directory structure with a checkpoint directory
temp_base_dir = tempfile.mkdtemp()
checkpoint_dir = os.path.join(temp_base_dir, 'checkpoints')
os.makedirs(checkpoint_dir, exist_ok=True)

# Create a checkpoint directory with the CHECKPOINT_NAME and some digits
checkpoint_subdir = os.path.join(checkpoint_dir, 'my_best_checkpoint_step_42')
os.makedirs(checkpoint_subdir, exist_ok=True)

# Create the ARGS dictionary with the required parameter
ARGS = {
    'some_dir': checkpoint_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, ndim 1, dtype family float — a tuple 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
find_classes Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/datasets.py
code served (permissive licence) · get_code("52d10a10fb72fb9f")
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 shutil

# Create a temporary directory with some subdirectories to simulate a class directory structure
temp_dir = tempfile.mkdtemp()
os.makedirs(os.path.join(temp_dir, "class_a"), exist_ok=True)
os.makedirs(os.path.join(temp_dir, "class_b"), exist_ok=True)
os.makedirs(os.path.join(temp_dir, "class_c"), exist_ok=True)

ARGS = {
    "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
mask_errors Ran
ran, violated the contract we drafted
alercebroker/atat/datasets.py
code served (permissive licence) · get_code("89cd0c4b1573f937")
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 (2, 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 numpy as np

# Create error array with shape (batch_size, sequence_length)
# Using distinct sizes: 2 for batch, 4 for sequence dimension
err = np.random.randn(2, 4).astype(np.float32)

# Create mask array with same shape as err
# Initialize as ones so we can see filtering effect
mask = np.ones((2, 4), dtype=np.float32)

ARGS = {
    "err": err,
    "mask": mask
}
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 2, dtype family float, shape (2, 4) — shape (2, 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
mask_mags Ran
ran, violated the contract we drafted
alercebroker/atat/datasets.py
code served (permissive licence) · get_code("48af8335729d239c")
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 (2, 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 numpy as np

# Create small magnitude array with shape (batch=2, features=4)
mag = np.random.randn(2, 4).astype(np.float32)

# Create corresponding mask array with same shape
mask = np.ones((2, 4), dtype=np.float32)

ARGS = {
    "mag": mag,
    "mask": mask,
}
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 2, dtype family float, shape (2, 4) — shape (2, 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
obtain_valid_mask Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/get_metrics.py
code served (permissive licence) · get_code("e88492ba48e57733")
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 torch

torch.manual_seed(0)

# Construct sample dict with 'mask' key
sample = {
    'mask': torch.ones(2, 4)
}

# Construct mask tensor (2, 4)
mask = torch.ones(2, 4)

# Construct time_alert tensor (2, 4) with values
time_alert = torch.tensor([[1.0, 2.0, 3.0, 4.0],
                            [0.5, 1.5, 2.5, 3.5]])

# Construct eval_time as a scalar
eval_time = 2.5

ARGS = {
    'sample': sample,
    'mask': mask,
    'time_alert': time_alert,
    'eval_time': eval_time
}
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 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
trial_name_string Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/loading.py
code served (permissive licence) · get_code("8aac396beeb74be1")
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.

class Trial:
    def __init__(self, config):
        self.config = config

ARGS = {
    'trial': Trial({
        'learning_rate': 0.001,
        'batch_size': 32,
        'num_layers': 5,
        'dropout': 0.5,
        'model_type': 'transformer',
        'seed': 42
    })
}
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
update_input_str Ran
ran; our contract draft was wrong, not the code
alercebroker/atat/loading.py
code served (permissive licence) · get_code("37e9f414d2e70de7")
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 torch
import torch.nn as nn

# The function accesses arch_spec['dict_tune'][key][config[key]]
# and iterates through config.keys(), treating 'seed' specially
# So we need:
# - arch_spec: dict with 'dict_tune' key containing nested dicts
# - config: dict with string keys and values that can index into arch_spec['dict_tune']

# Build arch_spec with dict_tune structure
arch_spec = {
    'dict_tune': {
        'learning_rate': {
            0: ' --learning_rate 0.001',
            1: ' --learning_rate 0.01',
        },
        'batch_size': {
            0: ' --batch_size 32',
            1: ' --batch_size 64',
        },
        'seed': {
            42: ' --seed 42',
            123: ' --seed 123',
        },
        'dropout': {
            0: ' --dropout 0.1',
            1: ' --dropout 0.5',
        }
    }
}

# Build config with keys that exist in arch_spec['dict_tune']
# and values that can index into those nested dicts
config = {
    'learning_rate': 0,
    'batch_size': 1,
    'seed': 42,
    'dropout': 0,
}

ARGS = {
    'arch_spec': arch_spec,
    'config': config,
}
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 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
get_tabular_data Not yet run
unverified (harvested, no recorded run)
alercebroker/atat/get_metrics.py
code served (permissive licence) · get_code("256abfbef79e3d4a")
How it ran

no invocation recorded.

load_data Not yet run
unverified (harvested, no recorded run)
alercebroker/atat/get_metrics.py
code served (permissive licence) · get_code("b9bc31cbaee4ee1e")
How it ran

no invocation recorded.

obtain_optimizer Not yet run
unverified (harvested, no recorded run)
alercebroker/atat/layers/optimizers.py
code served (permissive licence) · get_code("d507fb2b0b68321a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 14 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
find_classes 30 other papers
28 dated earlier, 2 later
1406.2661 (2014-06)
Generative Adversarial Networks
1612.06321 (2016-12)
Large-Scale Image Retrieval with Attentive Deep Local Features
1809.04729 (2018-09)
A Less Biased Evaluation of Out-of-distribution Sample Detectors
and 27 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

The advent of next-generation survey instruments, such as the Vera C. Rubin Observatory and its Legacy Survey of Space and Time (LSST), is opening a window for new research in time-domain astronomy. The Extended LSST Astronomical Time-Series Classification Challenge (ELAsTiCC) was created to test the capacity of brokers to deal with a simulated LSST stream. We describe ATAT, the Astronomical Transformer for time series And Tabular data, a classification model conceived by the ALeRCE alert broker to classify light-curves from next-generation alert streams. ATAT was tested in production during the first round of the ELAsTiCC campaigns. ATAT consists of two Transformer models that encode light curves and features using novel time modulation and quantile feature tokenizer mechanisms, respectively. ATAT was trained on different combinations of light curves, metadata, and features calculated over the light curves. We compare ATAT against the current ALeRCE classifier, a Balanced Hierarchical Random Forest (BHRF) trained on human-engineered features derived from light curves and metadata. When trained on light curves and metadata, ATAT achieves a macro F1-score of 82.9 +- 0.4 in 20 classes, outperforming the BHRF model trained on 429 features, which achieves a macro F1-score of 79.4 +- 0.1. The use of Transformer multimodal architectures, combining light curves and tabular data, opens new possibilities for classifying alerts from a new generation of large etendue telescopes, such as the Vera C. Rubin Observatory, in real-world brokering scenarios.

For agents

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

get_harvested_code_for_paper("2405.03078")
get_code_for_paper("2405.03078")
have("2405.03078")

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

curl https://syntology.ai/api/ran/2405.03078.json

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

[![Syntology run record](https://syntology.ai/api/ran/2405.03078.svg)](https://syntology.ai/paper/2405.03078)

Connect an agent — have() is free.