Training environment: <a target="_blank" rel="ugc" href="http://pixai.art">pixai.art</a> LoRA learningBase model: Anything V5Training images: 512×768 60 imagesChanges to Ver.2.0I remade it from scratch because the XL version didn't work out. The trigger word hasn't changed at all, except that it's now "nicole_a_v2,".Trigger word(v2.0): nicole_a_v2, green eyes, pink hair, long hair, two side up, hair ornament, hair ribbon, hairclip, mole under eye, earrings, jewelry, crop top, cleavage, shorts, black shorts, black thighhighs, black ribbon, single kneehigh, single thighhigh, thigh strap, boots, black footwear,Reference settings(v2.0):Checkpoint: anylora Sampler: DPM++ 2M Karras、Steps 40 CFG Scale 7、VAE：vae-ft-mse-840000The LoRA strength is about 6-8.Trigger word(v1.0): nicole_a, green eyes, pink hair, long hair, two side up, hair ornament, hair ribbon, hairclip, mole under eye, earrings, jewelry, crop top, cleavage, shorts, black shorts, black thighhighs, black ribbon, single kneehigh, single thighhigh, thigh strap, boots, black footwear,Reference settings(v1.0):Checkpoint: anylora, Sampler: DPM++ 2M Karras, Steps 30, CFG Scale 6, VAE: kl-f8-anime2EasyNegative used for embedded. This is just my environment, so please adjust it for yourself.

Files with SHA256: 9926b3135606a46fa72c9788385ba3e71b1420b59097600cd9e0dd9f006435d8