---
language: en
license: mit
thumbnail: https://i.ibb.co/6NLyc1P/stellar-diffusion.png
tags:
- stable-diffusion
- text-to-image
---

# Stellar Diffusion

Stellar Diffusion v0.1 vs Base Stable Diffusion v1.5
![Galaxy - Stellar Diffusion](images/galaxy_std.png) ![Galaxy - Stable Diffusion](images/galaxy_sd.png)
```
prompt = A hubble photograph of a galaxy
seed = 42
size = 512x512
```


Version: 0.1

Stable Diffusion 1.5 finetuned on high quality processed space imagery.

---
# Example Results

![Barred Spiral - Stellar Diffusion](images/barredspiral.png)
```
prompt = A barred spiral galaxy
seed = 44
size = 512x512
```

![Bright Constellation - Stellar Diffusion](images/brightconstellation.png)
```
prompt = bright, constellation, hubble, clouds
seed = 42
size = 512x512
```

![Nebula - Stellar Diffusion](images/nebula.png)
```
prompt = a hubble photograph of a nebula
seed = 42
size = 512x512
```

![NGC 7714 - Stellar Diffusion](images/ngc7714.png)
```
prompt = ngc 7714
seed = 42
size = 512x512
```

---

# Suggested parameters

512x512

---

# Reconized Tags

All reconized tags can be found in the tags.txt file. They are generated from the annotated descriptions of the photograph.
Current Dataset is small and is poor at generating exact celestial bodies, but is better at generating common generic bodies like nebula, galaxies etc.

### Partial support for scientific celestial body tags as follows:

NGC - New General Catalogue of Nebulae and Clusters of Stars

M / Messier - A set of 110 astronomical objects catalogued by the French astronomer Charles Messier

UGC – (catalog) Uppsala General Catalogue, a catalog of galaxies

### Partial support for the following classification methods as follows:

By recording instrument/spacecraft (ex. Voyager, Hubble)

By Color

By Celestial Body type

# Python Usage

```py
from diffusers import StableDiffusionPipeline
import torch

model_id = "rexwang8/stellar-diffusion"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
pipe = pipe.to("cuda")

prompt = "A hubble photograph of a galaxy"
image = pipe(prompt).images[0]  
    
image.save("hubble_galaxy.png")
```

---

# Dataset and Credits

### Model

Rex Wang (me!)

RunwayML for their SD 1.5

### Compute

Coreweave - 2x A40s

### Dataset

91 of the 100 images from https://esahubble.org/ Top 100 Hubble Images
ESA/Hubble

---

# Planned

Expansion of dataset to include:

Solar system

Asteroids

More star types

Black holes

Exo planets

More data in general

---

# Version History

V0.1 - 91 image dataset