README.md

metadata

base_model:
  - stable-diffusion-v1-5/stable-diffusion-v1-5
license: apache-2.0
pipeline_tag: text-to-image
library_name: diffusers

🌌 GalaxySD: Can AI Dream of Unseen Galaxies?

Conditional Diffusion Model for Galaxy Morphology Augmentation

🔍 Model Description

This repository provides the trained weights for GalaxySD, a fine-tuned Stable Diffusion model designed to generate high-fidelity galaxy images from text prompts. The model demonstrates strong performance in synthesizing visually realistic galaxies that adhere closely to specified morphological characteristics, such as spiral arms, bulge prominence, and edge-on orientation.

💻 How to Use

The file galaxsd_final.safetensors contains the fine-tuned Stable Diffusion weights for the GalaxySD model. To use this model in your own pipeline:

1. Make sure you have diffusers, transformers, accelerate, and safetensors installed:

pip install diffusers transformers accelerate safetensors

2. Load and run the model:

from diffusers import StableDiffusionPipeline
import torch

pipe = StableDiffusionPipeline.from_single_file(
    "galaxsd_final.safetensors",
    torch_dtype=torch.float16
).to("cuda")

prompt = "sdss, spiral galaxy, loosely wound spiral arms, 2 spiral arms"
image = pipe(prompt).images[0]
image.save("output.png")

If you encounter any issues, feel free to raise them in the GalaxySD Repo or contact authors.

🔗 Project Resources

• 🏠 Homepage
• 🚀 GalaxySD Repository
• 🛠️ Trained Galaxy Embedding Tool
• 🗂️ Training Dataset
• 📊 A Contributed Catalog

📄 Citation

@misc{ma2025aidreamunseengalaxies,
      title={Can AI Dream of Unseen Galaxies? Conditional Diffusion Model for Galaxy Morphology Augmentation}, 
      author={Chenrui Ma and Zechang Sun and Tao Jing and Zheng Cai and Yuan-Sen Ting and Song Huang and Mingyu Li},
      year={2025},
      eprint={2506.16233},
      archivePrefix={arXiv},
      primaryClass={astro-ph.GA},
      url={https://arxiv.org/abs/2506.16233}, 
}