Sorrentino Diffusion
Stable Diffusion model trained on images by the artists Andrea Sorrentino <div style="display: flex; flex-direction: row; flex-wrap: wrap"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959158-6303f37c3926de1f7ec42d3e.png" width="256"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959179-6303f37c3926de1f7ec42d3e.png" width="256"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959240-6303f37c3926de1f7ec42d3e.png" width="256"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959181-6303f37c3926de1f7ec42d3e.png" width="256"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959118-6303f37c3926de1f7ec42d3e.png" width="256"> <img src="https://s3.amazonaws.com/moonup/production/uploads/1667417959047-6303f37c3926de1f7ec42d3e.png" width="256"> </div>
How to use
- Download the model and use it on your desired UI (Tested on AUTOMATIC1111's) Currently only .ckpt version is supported
- Trigger the style in your prompt with the andreasorrentino token, look at the next section for more examples
Versions
- v1: Trained on 25 images over 3000 Dreambooth steps. 1000, 1500, 2000, 2500 and 3000 steps checkpoints available to download
We currently provide multiple checkpoints at different steps where you can compare results. v1 is only an experiment with a low quality dataset, results indicate the model might be overfitted. v2 will improve on dataset quality and quantity.
Examples
andreasorrentino style, a picture of a shiba inu
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 2207496243, Size: 512x512, Comparing v1 checkpoints
<hr />
drawing of a porsche, andreasorrentino style
Steps: 20, Sampler: Euler a, CFG scale: 7-15, Seed: 1734310449, Size: 512x512, andrea-sorrentino-v1_step_3000.ckpt
Tips
- Use different ways to trigger the style: andreasorrentino style, YOUR_PROMPT | YOUR_PROMPT in the style of andreasorrentino | YOUR_PROMPT, andreasorrentino style
https://twitter.com/nerijs