As I promised earlier, if I have bazz, I will create Loras. I thank everyone who likes and tips my images, supports me in every possible way. Your support helps me do this! This model is an attempt to synthesize a style imitating Frank Frazetta's painting technique. Lora is more suitable for fantasy. I recommend - Euler a and cfg 4.0upd: I'm currently looking at how the data in a dataset affects the results. Some experiments are giving me new insights into how to improve my style in the future. If you have any thoughts or results, I would really appreciate it if you shared them with me.

FRAZETTA technique (Ax) Syntez

I've expanded the dataset a bit, changed the settings. I want to find the best settings for style training. Let's see what I've got.

Files with SHA256: fb0ccf9154725664c74b31c0792be889fd934d91da69d0bb1bb646432a608c2c