SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.16829 · 2024

Maven: A Multimodal Foundation Model for Supernova Science

arXiv · PDF · Open in the Atlas

Code that ran

9 of 9 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 7 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/2408.16829.json · badge: /api/ran/2408.16829.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
thomashelfer/multimodal-supernovae canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
FunctionStatusWhere it lives
clip_loss Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/loss.py
code served (permissive licence) · get_code("ca4c88ac11df70b3")
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 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 torch.nn as nn

torch.manual_seed(0)

# Create small embedding tensors
# embs1 and embs2 should be 2D tensors (batch_size, embedding_dim)
# Using distinct batch sizes (2 and 3) and embedding dim (4)
embs1 = torch.randn(2, 4)
embs2 = torch.randn(3, 4)

# logit_scale should be a tensor that supports .exp()
logit_scale = torch.tensor(1.0)

# logit_bias is a scalar
logit_bias = 0.0

# image_encoder - not used in function body, can be None
image_encoder = None

# lightcurve_encoder - not used in function body, can be None
lightcurve_encoder = None

ARGS = {
    "embs1": embs1,
    "embs2": embs2,
    "logit_scale": logit_scale,
    "logit_bias": logit_bias,
    "image_encoder": image_encoder,
    "lightcurve_encoder": lightcurve_encoder,
}
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 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
clip_loss_multimodal Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/loss.py
code served (permissive licence) · get_code("8adc5be2a5d16bae")
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 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 torch.nn as nn

torch.manual_seed(0)

# Create three small embedding tensors with distinct sizes
# batch_size=2, embedding_dim=4
emb1 = torch.randn(2, 4)
emb2 = torch.randn(2, 4)
emb3 = torch.randn(2, 4)

# Pack embeddings into a list
embeddings = [emb1, emb2, emb3]

# Create logit_scales and logit_biases as tensors
# For 3 embeddings, we have 3*(3-1)/2 = 3 pairs
n_pairs = len(embeddings) * (len(embeddings) - 1) // 2
logit_scales = torch.tensor(1.0)
logit_biases = torch.tensor(0.0)

ARGS = {
    "embeddings": embeddings,
    "logit_scales": logit_scales,
    "logit_biases": logit_biases,
}
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 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
filter_files Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/utils.py
code served (permissive licence) · get_code("e57a00c69dd27571")
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 numpy as np

# Construct arguments for filter_files function

# Create a list of available filenames
filenames_avail = ['file_001.txt', 'file_002.txt', 'file_003.txt', 'file_004.txt']

# Create a list of filenames to filter (with some overlap and some non-overlapping)
filenames_to_filter = ['file_001.txt', 'file_002.txt', 'file_005.txt', 'file_006.txt']

# Create data to filter - a list of numpy arrays with different shapes
data_to_filter = [
    np.random.randn(4, 3),  # 4 samples, 3 features
    np.random.randn(4, 5),  # 4 samples, 5 features
]

ARGS = {
    'filenames_avail': filenames_avail,
    'filenames_to_filter': filenames_to_filter,
    'data_to_filter': data_to_filter,
}
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 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_indices_in_arrays Ran
ran, honoured the contract we drafted
thomashelfer/multimodal-supernovae/src/utils.py
code served (permissive licence) · get_code("86e7475bca034d65")
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 2, dtype family int; 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 numpy as np

# Based on the function, it takes two parameters:
# - st1: list or array of strings to find in st2
# - st2: list or array of strings to search within
# The function uses .index() method which works on lists

# Create simple test data
st1 = ["apple", "banana", "cherry", "date"]
st2 = ["banana", "apple", "elderberry", "fig", "date"]

ARGS = {
    "st1": st1,
    "st2": st2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family int — 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
get_continous_random_mask Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/models_pretraining.py
code served (permissive licence) · get_code("ab85ec5ba442c741")
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 3, 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 random

torch.manual_seed(0)

# Create a small padding mask tensor
# Shape: (B=2, T=4) - batch size 2, sequence length 4
# Values should be 1 (True) for valid positions, 0 (False) for padding
padding_mask = torch.tensor([
    [1, 1, 1, 1],
    [1, 1, 0, 0]
], dtype=torch.float32)

# Number of bands to divide the sequence into
nbands = 2

# Fraction of sequence to mask out (between 0 and 1)
f_mask = 0.15

ARGS = {
    "padding_mask": padding_mask,
    "nbands": nbands,
    "f_mask": f_mask
}
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 tuple, ndim 3, 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
get_random_mask Ran
ran, violated the contract we drafted
thomashelfer/multimodal-supernovae/src/models_pretraining.py
code served (permissive licence) · get_code("5eccc89f9ef96a26")
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 3, 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
from torch import Tensor

torch.manual_seed(0)

# Create a small padding mask tensor of shape (B, T) = (2, 4)
# True indicates valid positions, False indicates padding
padding_mask = torch.tensor([
    [True, True, True, False],
    [True, True, False, False]
], dtype=torch.bool)

ARGS = {
    "padding_mask": padding_mask,
    "f_mask": 0.15
}
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 tuple, ndim 3, 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
load_classes Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/dataloader.py
code served (permissive licence) · get_code("702728a86e02615f")
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 2, 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 numpy as np
import pandas as pd
import tempfile
import os

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

# Create a sample CSV file with the expected structure
csv_data = {
    'ZTFID': ['ZTF001', 'ZTF002', 'ZTF003', 'ZTF004', 'ZTF005', 'ZTF006'],
    'type': ['SN Ia', 'SN Ib', 'SN Ic', 'SN IIP', 'SLSN-I', 'SN IIn']
}
df = pd.DataFrame(csv_data)
csv_path = os.path.join(temp_dir, 'ZTFBTS_TransientTable.csv')
df.to_csv(csv_path, index=False)

ARGS = {
    'data_dir': temp_dir,
    'n_classes': 5,
    'filenames': None
}
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 tuple, ndim 2, 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
load_redshifts Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/dataloader.py
code served (permissive licence) · get_code("940f8b5955a417a7")
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 2, 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 numpy as np
import pandas as pd
import tempfile
import os

# Create a temporary directory with a mock CSV file
temp_dir = tempfile.mkdtemp()

# Create a mock ZTFBTS_TransientTable.csv with redshift data
mock_data = {
    'ZTFID': ['ZTF001', 'ZTF002', 'ZTF003', 'ZTF004', 'ZTF005'],
    'redshift': [0.5, 0.8, 1.2, np.nan, 0.3]
}
df_mock = pd.DataFrame(mock_data)
csv_path = os.path.join(temp_dir, 'ZTFBTS_TransientTable.csv')
df_mock.to_csv(csv_path, index=False)

# Define ARGS
ARGS = {
    'data_dir': temp_dir,
    'filenames': None
}
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 tuple, ndim 2, 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
sigmoid_loss Ran
ran; our contract draft was wrong, not the code
thomashelfer/multimodal-supernovae/src/loss.py
code served (permissive licence) · get_code("bedff2f0fe635f3f")
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 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

torch.manual_seed(0)

# Create embeddings with distinct dimensions
# embs1 should be (bs, dim) and embs2 should be (bs, dim)
bs = 2
dim = 4

embs1 = torch.randn(bs, dim, dtype=torch.float32)
embs2 = torch.randn(bs, dim, dtype=torch.float32)

# logit_scale must be a tensor that supports .exp()
logit_scale = torch.tensor(1.0, requires_grad=True)

# logit_bias is a scalar float
logit_bias = 2.73

ARGS = {
    "embs1": embs1,
    "embs2": embs2,
    "logit_scale": logit_scale,
    "logit_bias": logit_bias,
}
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 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

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 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 9 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

A common setting in astronomy is the availability of a small number of high-quality observations, and larger amounts of either lower-quality observations or synthetic data from simplified models. Time-domain astrophysics is a canonical example of this imbalance, with the number of supernovae observed photometrically outpacing the number observed spectroscopically by multiple orders of magnitude. At the same time, no data-driven models exist to understand these photometric and spectroscopic observables in a common context. Contrastive learning objectives, which have grown in popularity for aligning distinct data modalities in a shared embedding space, provide a potential solution to extract information from these modalities. We present Maven, the first foundation model for supernova science. To construct Maven, we first pre-train our model to align photometry and spectroscopy from 0.5M synthetic supernovae using a constrastive objective. We then fine-tune the model on 4,702 observed supernovae from the Zwicky Transient Facility. Maven reaches state-of-the-art performance on both classification and redshift estimation, despite the embeddings not being explicitly optimized for these tasks. Through ablation studies, we show that pre-training with synthetic data improves overall performance. In the upcoming era of the Vera C. Rubin Observatory, Maven serves as a Rosetta Stone for leveraging large, unlabeled and multimodal time-domain datasets.

For agents

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

get_harvested_code_for_paper("2408.16829")
get_code_for_paper("2408.16829")
have("2408.16829")

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

curl https://syntology.ai/api/ran/2408.16829.json

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

[![Syntology run record](https://syntology.ai/api/ran/2408.16829.svg)](https://syntology.ai/paper/2408.16829)

Connect an agent — have() is free.