v3.0 is strongly recommended. I have experimented and improved hyper parameters, dataset size and quality as well dataset captioning.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/86443c4c-07c9-4c68-bd28-d9a766042214/original=true/86443c4c-07c9-4c68-bd28-d9a766042214.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/431c44b7-dbad-41eb-af27-b0096d4f0a52/original=true/431c44b7-dbad-41eb-af27-b0096d4f0a52.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/61b53eed-4a6e-4c60-ade5-38498a1c2d78/original=true/61b53eed-4a6e-4c60-ade5-38498a1c2d78.jpeg" />Triggerword: kchisara

Kchisara - Krea2

Dataset Optimization and more chackpoints saved allowed me to better cherry picking.

Files with SHA256: ba75c0a2e99df27b9119b39b678aebdb6b288c43035d07491e9d399b7e5359a5