datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
galaxies
Galaxies for training astroPT
Here we have ~8.5 million galaxy cutouts from the DESI legacy survey DR8.
The cut outs are 512x512 pixel jpg images centred on the galaxy source.
I've split away 1% of the images into a test set, and 1% into a validation set.
The remaining 98% of the images comprise the training set.
Metadata (galaxy properties)
There is also accompanying metadata!
To combine the metadata with the galaxy images you can do (for example):
from datasets import… See the full description on the dataset page: https://huggingface.co/datasets/Smith42/galaxies.PAPO_evalgalaxy_manifoldautotrain-data-galaxy_classification
AutoTrain Dataset for project: galaxy_classification
Dataset Description
This dataset has been automatically processed by AutoTrain for project galaxy_classification.
Languages
The BCP-47 code for the dataset's language is unk.
Dataset Structure
Data Instances
A sample from this dataset looks as follows:
[
{
"image": "<256x256 RGB PIL image>",
"target": 0
},
{
"image": "<256x256 RGB PIL image>",
"target": 0
}]… See the full description on the dataset page: https://huggingface.co/datasets/Xanadu00/autotrain-data-galaxy_classification.galaxy10_decals
Dataset Card for "galaxy10_decals"
Galaxy10 DECals is a dataset contains 17736 256x256 pixels colored galaxy images separated in 10 classes.
Galaxy10 DECals images come from DESI Legacy Imaging Surveys which includes the Beijing-Arizona Sky Survey (BASS), the DECam Legacy Survey (DECaLS) and the Mayall z-band Legacy Survey and labels come from Galaxy Zoo.
Dataset create from
https://astronn.readthedocs.io/en/latest/galaxy10.html
https://github.com/henrysky/Galaxy10… See the full description on the dataset page: https://huggingface.co/datasets/matthieulel/galaxy10_decals.galaxy-zoo-dataset
Galaxy Morphological Classification (Derived from Galaxy Zoo)
Dataset Summary
This dataset is designed for Galaxy Morphological Classification tasks, specifically tailored for training deep learning models such as Efficient Vision Transformers (ViT).
The dataset contains galaxy images and their corresponding morphological labels (8 classes). The underlying images and ground-truth classifications are derived from the Galaxy Zoo DECaLS project (Walmsley et al., 2022). The… See the full description on the dataset page: https://huggingface.co/datasets/mrJordi0/galaxy-zoo-dataset.galahad-deconf-verb
Deconfounded verb set (RoboCasa)
Part of the Galahad release · Project page · Code + battery + generator
Lift-versus-slide with the same object under both verbs, role-balanced, so the verb is the only cue. Behind the verb-selectivity result (+44.3pp, told-slide→lift 2.0%).
LeRobot v3 format. Generated by the deconfounding generators in the release repo; the generator produces the exam (the battery) and the medicine (the training set) from the same code.
galaxy-descriptions
Galaxy Descriptions
Project Page | Code
This dataset provides galaxy cutout images, natural-language descriptions, text embeddings, and image embeddings for galaxies drawn from multiple imaging surveys (specifically Legacy DR10 and HSC PDR3 Wide).
Each row corresponds to a single galaxy and contains:
A preprocessed RGB galaxy image
A caption generated by gpt-4.1-mini
A single-sentence summary of the caption generated by gpt-4.1-nano
Text embeddings for the caption and summary… See the full description on the dataset page: https://huggingface.co/datasets/astronolan/galaxy-descriptions.gala-article-assets
gala-article-assets
Figures for the Gala article (see junafinity/Gala-598M-MLX).
All plots are generated directly from the training/eval logs shipped in the model repo.
file
shows
fig_10m.png
10.5M-token stream: decode speed and next-2k loss flat at every depth
fig_seeds.png
3-seed comparison, Gala vs Transformer, matched 120M tokens
fig_scale.png
120M→300M scale study: the curves diverge
fig_passkey.png
recall boundary: 100% in-window, chance beyond… See the full description on the dataset page: https://huggingface.co/datasets/junafinity/gala-article-assets.galaxies_captioned
Galaxy captions for pairing with `smith42/galaxies' dataset
Here we have gemini-generated captions for 160k galaxies.
Useful links
Paper here: https://arxiv.org/abs/2405.14930
Models here: https://huggingface.co/Smith42/astroPT
Code here: https://github.com/smith42/astroPT
galaxies-with-hats
Galaxies with HATS
This is a HATS (HEALPix Adaptive Tiling Scheme)
version of Smith42/galaxies
(revision v2.0): ~8.5 million 256×256 pixel PNG galaxy cutouts from the
DESI Legacy Survey DR8, centred on
the galaxy source, spatially partitioned on the sky, and bundled with all ~160
metadata columns (Galaxy Zoo DESI morphologies, NSA photometry, redshifts, and
OSSY / ALFALFA / JHU-MPA cross-matches) in every row.
Where the original dataset is ordered for ML training, this… See the full description on the dataset page: https://huggingface.co/datasets/Smith42/galaxies-with-hats.Cityscapes_Style_2048_1024HumanoidDatagalaxy10-aion
Galaxy10 AION-1 Benchmark
This dataset provides the exact train/test split used to produce the Galaxy Morphology Classification results in the AION-1 paper (Table 2, Section 7.2.2) and used in the AION-Search paper.
Task
Classify galaxy images into 10 morphology classes from Galaxy Zoo DECaLS:
Label
Class Name
0
Disturbed Galaxies
1
Merging Galaxies
2
Round Smooth Galaxies
3
In-between Round Smooth Galaxies
4
Cigar Shaped Smooth Galaxies
5… See the full description on the dataset page: https://huggingface.co/datasets/astronolan/galaxy10-aion.galahad-deconf-category
Deconfounded category set (RoboCasa)
Part of the Galahad release · Project page · Code + battery + generator
500 episodes, 100 per category, category ⟂ instance ⟂ co-appearance ⟂ position: per-episode mesh resampling breaks single-instance memorization, random distractor subsets break co-appearance.
LeRobot v2.1 format. Generated by the deconfounding generators in the release repo; the generator produces the exam (the battery) and the medicine (the training set) from the same… See the full description on the dataset page: https://huggingface.co/datasets/phi-monster/galahad-deconf-category.galaxies_embeddings
Embeddings for 🤗 smith42/galaxies
This dataset contains embeddings generated by AstroPTv2.0 🐙.
The embeddings are generated from smith42/galaxies, and can be paired with accompanying metadata from that dataset (the rows are in the same order!).
Check out the github for more docs and info on AstroPT.
galaxies_jetformer_reconJetFormer reconstructions + embeddings of Smith42/galaxies (test + validation splits).
Model: JetFormer (sogol_checkpoint_step_0079999)
reconstructed_image: deterministic JetFormer reconstruction
jetformer_embedding (768-d): JetFormer layer-6 embedding of the original galaxy image
resnet50_embedding (2048-d): ResNet-50 embedding of the JetFormer-reconstructed image
resnet50_original_embedding (2048-d): ResNet-50 embedding of the original galaxy image
resnet50_vqvae_embedding (2048-d):… See the full description on the dataset page: https://huggingface.co/datasets/SogolS/galaxies_jetformer_recon.galaxy_mnist
Dataset GZMNIST
Dataset create from Mike Walmsley repo : https://github.com/mwalmsley/galaxy_mnist
Contains 10,000 images of galaxies, only HiRes 3x224x224.
Confidently labelled by Galaxy Zoo volunteers as belonging to one of four morphology classes.
Classes
0 : 'smooth and round'
1 : 'smooth and cigar-shaped'
2 : 'edge-on-disk'
3 : 'unbarred spiral
Data from / Citation :
Walmsley et al.(2022) - 2022MNRAS.509.3966W
Walmsley, M. and 17 colleagues 2022.
Galaxy… See the full description on the dataset page: https://huggingface.co/datasets/matthieulel/galaxy_mnist.galaxy-zoo-dataset
Galaxy Morphological Classification (Derived from Galaxy Zoo)
Dataset Summary
This dataset is designed for Galaxy Morphological Classification tasks, specifically tailored for training deep learning models such as Efficient Vision Transformers (ViT).
The dataset contains galaxy images and their corresponding morphological labels (8 classes). The underlying images and ground-truth classifications are derived from the Galaxy Zoo DECaLS project (Walmsley et al., 2022). This… See the full description on the dataset page: https://huggingface.co/datasets/Greate123/galaxy-zoo-dataset.galahad-deconf-object
Deconfounded object-identity set (LIBERO)
Part of the Galahad release · Project page · Code + battery + generator
900 episodes, 90 per object across all 10 LIBERO objects, perfect role balance: every object appears as the target and as a distractor across randomized positions, so the instruction is the only predictor of the target. The training set behind the object-substitution cure (base 13.5% → 94.5%).
LeRobot v3 format. Generated by the deconfounding generators in the… See the full description on the dataset page: https://huggingface.co/datasets/phi-monster/galahad-deconf-object.TeTIm-EvalText To Image Evaluation (TeTIm-Eval)galahad-demo-sim
Demo video assets
Part of the Galahad release · Project page · Code + battery + generator
Simulation demo clips: change one word and the arm changes targets.
LeRobot v3 format. Generated by the deconfounding generators in the release repo; the generator produces the exam (the battery) and the medicine (the training set) from the same code.
galahad-demo-chain
Chain demo frames
Part of the Galahad release · Project page · Code + battery + generator
Same scene, one word changed: the predicted motion region and the action endpoint flip together to the newly named object on a single set of weights (EEF lands 0.3–0.5 cm away). Frames and episodes for the dual-output demonstration.
LeRobot v3 format. Generated by the deconfounding generators in the release repo; the generator produces the exam (the battery) and the medicine (the training… See the full description on the dataset page: https://huggingface.co/datasets/phi-monster/galahad-demo-chain.contest-frames-fullindex-cards-cas-galapagos-stewart-specimens
Alban Stewart's Galápagos Expedition Specimen Cards (CAS Archives, 1905–1906)
1,059 specimen index cards from the California Academy of Sciences Archives,
documenting Alban Stewart's botanical specimens collected on the 1905–1906
California Academy of Sciences Galápagos Expedition. One card per specimen with
species, locality on the islands, collection date and field notes; companion to
Stewart's expedition journal.
Harvested from the 6 IA items
csfa788562626 +
csfa788562626Alpha +… See the full description on the dataset page: https://huggingface.co/datasets/biglam/index-cards-cas-galapagos-stewart-specimens.GalaxyMergerPreprocessedImgOnlyResizegalaxy10_decalsgalaxy10_decals_astropilegalaxy-description-benchmark
Description Benchmark based on GalaxyZoo
A small sample of the GZ-DECaLS dataset to cheaply test the ability of vision-language models to describe galaxy images.
Data source
The underlying catalog data are derived from the Galaxy Zoo DECaLS: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314,000 Galaxies and remain subject to their original data-use policy and required acknowledgements.
Embeddings Citation… See the full description on the dataset page: https://huggingface.co/datasets/astronolan/galaxy-description-benchmark.instructPix2Pix_cityscapes_512