Artist | LoRA for Stable Diffusion 1.5
A LoRA for Stable Diffusion 1.5 to create an artistic watercolor and sketch effect in Automatic1111
This LoRA has been trained on 100 images created using the Artist Photoshop Effect
The LoRA is compatible with stable diffusion 1.5 using the Automatic1111 interface. Make sure you have loaded a stable diffusion 1.5 based model into Automatic1111 before using this LoRA. It is not compatible with stable diffusion 2.x
How does it work?
Firstly, ensure you have the latest version of Automatic1111 installed. How to install Stable Diffusion locally using Automatic1111
After installing Automatic1111, copy the “7sArtist.safetensors” LoRA file found in this download, into your stable diffusion Lora folder directory – E.g.\stable-diffusion-webui\models\Lora
To create the effect, add <lora:7sArtist:1> anywhere in your prompt – E.g. “drawing of a woman <lora:7sArtist:1>“
To adjust the strength of the effect, change the number after the ‘:‘ – E.g. <lora:7sArtist:0.5>, <lora:7sArtist:0.6>, <lora:7sArtist:0.7>, <lora:7sArtist:0.8>, <lora:7sArtist:0.9>, <lora:7sArtist:1>, <lora:7sArtist:1.1>
The default strength is 1, so make sure to lower it if the effect is coming through too strong.
- Recommended CFG scale of 3-6
- Try using ‘UniPC’, ‘DPM ++2M’ , ‘DPM ++2m Karras’ and DDIM samplers to bring out more detail
- Add terms to your prompt such as ‘drawing’, ‘sketch art’, ‘concept art’, ‘bold colors, ‘contrast’, ‘watercolor’, ‘painted’, ‘line art’, ‘ink outlines’ to create various artistic looks.
- Use the Controlnet extension for Automatic1111 to apply the effect to your own images – or base your image composition off another image.
- Experiment with custom checkpoint models with this LoRA file instead of the base stable diffusion 1.5 model. Try out Deliberate, Realistic Vision or Dreamshaper to get started.
- Add the ‘epi_noiseoffset‘ LoRA in addition to this LoRA to your prompt to add varying amounts of contrast to your image
- Use the ‘Highres. fix’ option in Automatic1111 with a low denoising strength (0.1) to double the resolution of your image and bring out much finer details.
- Try combing this LoRA with other LoRA’s in your prompts
- Try adding color keywords into your prompt for more creative looks – e.g. “drawing of a man, grey, blue”
What was my process to create the LoRA?
Try it out
If you want to try out the effect or use it for your personal projects, download the file using the ‘Personal Use’ option.
- Re-trained with adjusted parameters (better results)
- Re-trained with an additional 40 images (100 total)
It´s really good
I must say that this is one of the easiest Loras that I tried. Usually a Lora takes some attempts to get decent outputs, but in this one is really good on it.