This is a finetune that is made to create Sillytavern characters.
This model ist trained on ~1200 Samples, about 1/3 multiturn. It has a strong prebaked structure.
If you want this finetune on a different model, just comment and i will do it.
Plans for V2: Shorter answers, a bit more NSFW, better image prompts, more samples
This gemma3 model was trained 2x faster with
Unsloth and Huggingface's TRL library.
