kwazzi-jack/mirabest-radio-astronomy-unofficial

MiraBest Radio Astronomy Dataset (Unofficial)

⚠️ IMPORTANT: This is an unofficial repository containing a processed version of the MiraBest dataset formatted for stable diffusion fine-tuning. This repository is not affiliated with the original authors.

Unofficial processing of the MiraBest radio astronomy dataset with original classification labels and natural language captions for diffusion fine-tuning. Original dataset by Porter & Scaife (2023).

Original Dataset

The original MiraBest dataset was created by:

• Authors: Fiona Porter, Anna M. M. Scaife
• Paper: "MiraBest: A Dataset of Morphologically Classified Radio Sources for Machine Learning"
• Source: https://academic.oup.com/rasti/article/2/1/293/7202349
• Original Source: https://doi.org/10.5281/zenodo.5588282

This Repository

This unofficial repository contains:

• Converted PNG images (150x150 grayscale) from the original pickle format
• Natural language captions generated for stable diffusion training
• HuggingFace datasets-compatible structure with PARQUET files
• Three separate datasets: full (combined), F_batches only, and N_batches only

Dataset Details

• Version: 1.0.0
• Total Images: 2,512
• Image Format: PARQUET (optimized for HuggingFace Hub)
• Original Source: https://doi.org/10.5281/zenodo.5588282
• Original Version: 2.0.0
• Processing: Unofficial conversion for stable diffusion training
• Format: PARQUET files optimized for HuggingFace Hub

Available Datasets

• Full Dataset (data_dir="full"): 2,512 images (both F and N batches combined)
• F Batch Dataset (data_dir="F_batches"): 1,256 images (F batch only)
• N Batch Dataset (data_dir="N_batches"): 1,256 images (N batch only)

Galaxy Type Distribution

• FRI: 1,182 (47.1%)
• FRII: 1,262 (50.2%)
• hybrid: 68 (2.7%)

Usage

from datasets import load_dataset

# Load the full dataset (both F and N batches)
dataset = load_dataset("kwazzi-jack/mirabest-radio-astronomy-unofficial", data_dir="full")

# Or load specific batch datasets:
# f_dataset = load_dataset("kwazzi-jack/mirabest-radio-astronomy-unofficial", data_dir="F_batches")  # F batch only
# n_dataset = load_dataset("kwazzi-jack/mirabest-radio-astronomy-unofficial", data_dir="N_batches")  # N batch only

# Access the training data
train_data = dataset['train']

# Access images and captions
for item in train_data:
    image = item['image']
    description = item['description']
    metadata = item['metadata']
    print("Description:", description)
    print("Galaxy type:", metadata['galaxy_type'])
    print("Batch type:", metadata['batch_type'])

# Check dataset size
print(f"Total samples: {len(train_data)}")
```## Citation

**IMPORTANT**: If you use this dataset, please cite the original MiraBest paper:

```bibtex
@article{10.1093/rasti/rzad017,
    author = {Porter, Fiona A M and Scaife, Anna M M},
    title = {MiraBest: a data set of morphologically classified radio galaxies for machine learning},
    journal = {RAS Techniques and Instruments},
    volume = {2},
    number = {1},
    pages = {293-306},
    year = {2023},
    month = {06},
    abstract = {The volume of data from current and future observatories has motivated the increased development and application of automated machine learning methodologies for astronomy. However, less attention has been given to the production of standardized data sets for assessing the performance of different machine learning algorithms within astronomy and astrophysics. Here we describe in detail the MiraBest data set, a publicly available batched data set of 1256 radio-loud AGN from NVSS and FIRST, filtered to 0.03 \< z \< 0.1, manually labelled by Miraghaei and Best according to the Fanaroff-Riley morphological classification, created for machine learning applications and compatible for use with standard deep learning libraries. We outline the principles underlying the construction of the data set, the sample selection and pre-processing methodology, data set structure and composition, as well as a comparison of MiraBest to other data sets used in the literature. Existing applications that utilize the MiraBest data set are reviewed, and an extended data set of 2100 sources is created by cross-matching MiraBest with other catalogues of radio-loud AGN that have been used more widely in the literature for machine learning applications.},
    issn = {2752-8200},
    doi = {10.1093/rasti/rzad017},
    url = {https://doi.org/10.1093/rasti/rzad017},
    eprint = {https://academic.oup.com/rasti/article-pdf/2/1/293/50696568/rzad017.pdf},
}

License

This processed version is released under CC BY 4.0 license, following the same terms as the original MiraBest dataset. Please refer to the original dataset documentation for full licensing details.

Disclaimer

This is an unofficial processing of the MiraBest dataset. For the authoritative version and latest updates, please refer to the original dataset at https://doi.org/10.5281/zenodo.5588282.