|
--- |
|
license: creativeml-openrail-m |
|
datasets: |
|
- mirav/artistic-imagery |
|
language: |
|
- en |
|
pipeline_tag: image-to-text |
|
tags: |
|
- stable diffusion |
|
--- |
|
|
|
## Description |
|
Stable Diffusion 1.5 model finetuned on a subset of [mirav/artistic-imagery](https://huggingface.co/datasets/mirav/artistic-imagery). Still a work in progress.<br> |
|
|
|
## Goals |
|
Providing a model that can produce a wide variety of styles and is highly responsive to certain traditional art terms. |
|
Current trained terms (as of version 1.0): |
|
* watercolor \(medium\) |
|
* watercolor pencil \(medium\) |
|
* sketch |
|
* traditional media |
|
* ink wash painting |
|
* impressionism |
|
* acrylic painting |
|
* oil painting |
|
* chiaroscuro<sup>1</sup> |
|
|
|
## Notes |
|
<sup>1</sup> Due to documented issues with the noise scheduler, this does not presently have quite the intended effect. |
|
|
|
## TODO |
|
* Improve caption quality. |
|
|
