|
--- |
|
license: apache-2.0 |
|
|
|
tags: |
|
- text-to-image |
|
- template:sd-lora |
|
- flux |
|
- lora |
|
- flux dev |
|
- image-generation |
|
- diffusers |
|
- painting |
|
- art |
|
- illustration |
|
- pop art |
|
- sots art |
|
pipeline_tag: text-to-image |
|
emoji: 🔜 |
|
language: |
|
- en |
|
base_model: black-forest-labs/FLUX.1-dev |
|
|
|
instance_prompt: SOTS, SOTS art style, SOTS art painting |
|
|
|
widget: |
|
- text: SOTS art style painting |
|
output: |
|
url: sotsart1.png |
|
- text: SOTS art style painting |
|
output: |
|
url: sotsart2.webp |
|
- text: SOTS art style painting |
|
output: |
|
url: sotsart3.webp |
|
- text: SOTS art style painting |
|
output: |
|
url: sotsart4.webp |
|
- text: SOTS art style painting |
|
output: |
|
url: sotsart5.png |
|
|
|
--- |
|
--- |
|
# SOTS ART LoRA by A.C.T. SOON® |
|
A style LoRA fine-tuned for [FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-Dev). |
|
Trained on 70 paintings by Dubosarskiy and Vinogradov. |
|
This LoRA was fine-tuned via a combination of two experimental methods suggested by theLastBen: |
|
Firstly, targetting only select layers of the base model. |
|
In this instance, the layers 7, 9, 16, and 25 were targetted. |
|
Secondly, training on four 512x512 images per input file (within a 1024x1024 spread), thereby reducing the dataset size and quickening the training (aka homebrew express batching technique). |
|
|
|
<Gallery /> |
|
|
|
## Trigger words |
|
You should use `SOTS` or 'SOTS art style painting' to trigger the image generation. |
|
|
|
- base model: [black-forest-labs/FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev) |
|
|
|
|
|
## Download model |
|
|
|
Weights for this model are available in Safetensors format. |
|
[Download](/AlekseyCalvin/SOTS_art/tree/main) them in the Files & versions tab. |
|
|
|
|
|
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers) |