BTSbot is a multi-modal deep vision model for automating supernova identification and follow-up in the Zwicky Transient Facility (ZTF) Bright Transient Survey (BTS).
BTSbot contributed to the first supernova to be fully automatically discovered, confirmed, classified, and shared (AstroNote, press release) as well automated space-based supernova follow-up (AstroNote). See this animated walkthrough of the fully-automated BTS workflow.
Presented at the ML for Astrophysics workshop at ICML 2023 (Extended abstract)
The training set for the original production model is available on Zenodo.
Installation
Install with pip:
pip install btsbot
Usage
BTSbot models are now available on the HuggingFace Hub and can be downloaded and loaded into Python automatically. Here's how simple it is to get started:
import btsbot
# This will automatically download the model if not already present locally
model = btsbot.load_HF_model(
architecture="convnext", # or "maxvit"
multi_modal=True, # Set to False for image-only models
pretrain="galaxyzoo" # or "imagenet", "randinit"
)
The model object can then be used for inference. For a complete inference example, see inference_example.py which demonstrates:
- Loading example data (triplets and metadata)
- Running inference on batches
- Processing predictions
Citing BTSbot
If you use or reference BTSbot please cite Rehemtulla et al. 2024 (ADS).
@ARTICLE{Rehemtulla+2024,
author = {{Rehemtulla}, Nabeel and {Miller}, Adam A. and {Jegou Du Laz}, Theophile and {Coughlin}, Michael W. and {Fremling}, Christoffer and {Perley}, Daniel A. and {Qin}, Yu-Jing and {Sollerman}, Jesper and {Mahabal}, Ashish A. and {Laher}, Russ R. and {Riddle}, Reed and {Rusholme}, Ben and {Kulkarni}, Shrinivas R.},
title = "{The Zwicky Transient Facility Bright Transient Survey. III. BTSbot: Automated Identification and Follow-up of Bright Transients with Deep Learning}",
journal = {\apj},
keywords = {Time domain astronomy, Sky surveys, Supernovae, Convolutional neural networks, 2109, 1464, 1668, 1938, Astrophysics - Instrumentation and Methods for Astrophysics},
year = 2024,
month = sep,
volume = {972},
number = {1},
eid = {7},
pages = {7},
doi = {10.3847/1538-4357/ad5666},
archivePrefix = {arXiv},
eprint = {2401.15167},
primaryClass = {astro-ph.IM},
adsurl = {https://ui.adsabs.harvard.edu/abs/2024ApJ...972....7R},
adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}
If you use or reference a pre-trained BTSbot model like the updated ConvNeXt-based BTSbot
or any BTSbot output from 2026 or beyond, please also cite our follow-up publication.
@ARTICLE{Rehemtulla+2025,
title={Pre-training vision models for the classification of alerts from wide-field time-domain surveys},
author={Nabeel Rehemtulla and Adam A. Miller and Mike Walmsley and Ved G. Shah and Theophile Jegou du Laz and Michael W. Coughlin and Argyro Sasli and Joshua Bloom and Christoffer Fremling and Matthew J. Graham and Steven L. Groom and David Hale and Ashish A. Mahabal and Daniel A. Perley and Josiah Purdum and Ben Rusholme and Jesper Sollerman and Mansi M. Kasliwal},
year={2025},
eprint={2512.11957},
archivePrefix={arXiv},
primaryClass={astro-ph.IM},
url={https://arxiv.org/abs/2512.11957},
}
Metadata
Release files for btsbot 2.0.7
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| btsbot-2.0.7.tar.gz | 32.8 kB | Without provenance | Details for btsbot-2.0.7.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| btsbot-2.0.7-py3-none-any.whl | Python 3 | none | any | Details for btsbot-2.0.7-py3-none-any.whl |
Total release size: 67.4 kB
Release files / btsbot-2.0.7.tar.gz
| Download URL | btsbot-2.0.7.tar.gz |
|---|---|
| Size | 32.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
6e16e132f617f4acac5a89f7051f3e1fd5f8d0be09b1a21e09f86107552e8573
|
|
BLAKE2b-256 checksum How to use checksums |
7e1dd65ebdaef1c4abc2728319b802b0c2b22c3ae373bfd5035bb8f2d610a462
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.14
|
Release files / btsbot-2.0.7-py3-none-any.whl
| Download URL | btsbot-2.0.7-py3-none-any.whl |
|---|---|
| Size | 34.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
de33691e73c42090928f84a56aabb9127c000a6ceb529d2a48b20a38b78b9447
|
|
BLAKE2b-256 checksum How to use checksums |
2cd15f3b42811c3bfd7e18508143841dced1e1f022677414f65ce97836a86b5e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.14
|
Security for btsbot 2.0.7
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.