Calling All Style LoRA Creators – I Need Your Help!

Hi everyone! I wanted to try out and train a second version of my first style lora https://civitai.com/models/1466088/fsmix-style-lora.

I have already tried once but wasn't really happy with the result so i'm coming for your help 😆

I'm looking for any pointers, suggestions, training parameters, tagging tips or whatever else you can think of to end up with a good result with the training!

I should mention I'm using the on/site training and I've already saved up around 300 images in my dataset.

Thanks in advance to anyone that's going to take some time to give me an hand. 🙏