|
--- |
|
license: cc |
|
--- |
|
A Textual Inversion Embedding for Stable Diffusion, trained on SD2.1 |
|
|
|
Version 2 of this embedding now allows you to achieve that loose digital paint style with much greater flexibility and fewer artifacts. I highly recommend it over V1. |
|
|
|
Simply call on the keyword laxpeintV2 (or laxpeint for the original) as in, single solo skydiver by laxpeintV2, overhead view of English country, above a skydiver, excitement and adrenaline |
|
|
|
<strong>Interested in generating your own embeddings? <a href="https://docs.google.com/document/d/1JvlM0phnok4pghVBAMsMq_-Z18_ip_GXvHYE0mITdFE/edit?usp=sharing" target="_blank">My Google doc walkthrough might help</a></strong> |
|
|
|
One odd artifact (though I think it's a funny kind of blessing from the gods of the latent space) is that skydiver prompt and multiple variations all generated a horse. No horses were included in the training data, so where it came from? Who knows! But I love it. |
|
|
|
Generally, laxpeintV2 lets you prompt directly and simply for what you want and you'll be quickly on the right track, whether that's a cute animal, anthropomorphic animal, fantasy scene, sci-fi scene, landscape, interior, or whatever you dream up. This embedding works with your vision and rarely requires you to struggle with its quirks. |
|
|
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
