My first attempt at creating my own LORA.
Style of Jacek Yerka, a modern surrealist painter from Poland.
230 images were used to train the model.
The training of the LoRA was on stable-diffusion-v1-5 model.
Max resolution was 1600x1600.
For best results, it's better not to use any negative prompts related to people. Also, I recommend to set the strength of this LORA to 0.1-0.3. Higher values tend to ignore your positive prompts related to anything that Yerka's paintings don't have.