|
--- |
|
license: cc0-1.0 |
|
pipeline_tag: text-to-image |
|
tags: |
|
- art |
|
- not_for_all_eyes |
|
--- |
|
|
|
# pleroma |
|
|
|
noun |
|
1.(in Gnosticism) the spiritual universe as the abode of God and of the totality of the divine powers and emanations. |
|
2. a visual system for achieving beauty, created by me. |
|
|
|
the model itself is an artistic statement, along with its outputs. |
|
|
|
while the model and the embeddings are cc0, any outputs that come out can be commercialized by the output's creator, the end user, with no reference needed to me, the author. |
|
have fun, do anything you like with it. i can't impose restrictions on beauty! |
|
|
|
|
|
## Intended uses & limitations |
|
|
|
these inversion embeddings are the second layer of achieving "pleroma art", the first layer being [the pleroma init model](https://huggingface.co/srodan/pleroma) |
|
#### How to use |
|
copy the contents to your inversion embedding folders for your local stable diffusion. |
|
|
|
#### Limitations and bias |
|
the embeddings have a limited palette of blue white and black, with red as a rare accent color, in which mixing with analog diffusion or stable diffusion 1.4-1.5 has achieved best results. |
|
the model is very limited to create a specific artistic style, and was made as a proof-of-concept to show the possibility of creating new artstyles with latent space. |
|
no artist names were used in the creation of synthetic dataset. |
|
the model has an affinity to create people and buildings, but each model differs in scene creation, and the used .ckpt, so do share your experiments ^^ |
|
|
|
## Training data |
|
|
|
the training data consists of 10 mixed media digital artworks that were synthetically made. a curated dataset of what was used to train the diffusers model referred in "intended use & limitations" |
|
the training data was changed in specific steps, so that three versions of pleroma embeddings were born out of it, as an unholy trinity symbolism. |
|
another symbolism that was made in training was using some outputs to be fed into the training data, which ritualistically created the ouroboros. |
|
## pleroma, pleroma-1800 and ploqoso |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma2.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma3.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma1800.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma1800-2.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/pleroma1800-3.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/ploqoso.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/ploqoso2.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/ploqoso3.png) |
|
## Mixed embeddings X Wavymulder's [Analog Diffusion](https://huggingface.co/wavymulder/Analog-Diffusion) ckpt |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/01.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/02.jpg) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/03.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/04.png) |
|
![Example](https://huggingface.co/srodan/pleroma-inversion-embeddings/resolve/main/images/05.png) |