Krea 2 LoRA — luzai-9527/luzai-krea2-character-v1

Prompt

A close-up photorealistic portrait of luzai_k2v1, an adult East Asian deer woman with dark brown twin ponytails, straight bangs, symmetrical deer ears and small three-tined antlers, neutral expression, natural window light, sharp facial detail.

Prompt

A full-body candid street photograph of luzai_k2v1 walking through a rainy modern city in a red raincoat and black boots, natural stride, complete body in frame, realistic skin and hair.

Prompt

A medium portrait of luzai_k2v1 seated in a quiet cafe wearing a pale blue blouse, looking slightly screen-left with a subtle closed-mouth smile, warm natural light, photorealistic facial detail.

A DreamBooth-LoRA for Krea 2, trained on Krea 2 RAW and shown on Krea 2 Turbo. The samples below were generated with this LoRA on Turbo (8 steps).

Trigger

Use the token luzai_k2v1 to invoke the concept.

Samples

"A close-up photorealistic portrait of luzai_k2v1, an adult East Asian deer woman with dark brown twin ponytails, straight bangs, symmetrical deer ears and small three-tined antlers, neutral expression, natural window light, sharp facial detail."

"A full-body candid street photograph of luzai_k2v1 walking through a rainy modern city in a red raincoat and black boots, natural stride, complete body in frame, realistic skin and hair."

"A medium portrait of luzai_k2v1 seated in a quiet cafe wearing a pale blue blouse, looking slightly screen-left with a subtle closed-mouth smile, warm natural light, photorealistic facial detail."

Use it with diffusers

import torch
from diffusers import Krea2Pipeline

pipe = Krea2Pipeline.from_pretrained("krea/Krea-2-Turbo", torch_dtype=torch.bfloat16).to("cuda")
pipe.load_lora_weights("luzai-9527/luzai-krea2-character-v1")
image = pipe("A close-up photorealistic portrait of luzai_k2v1, an adult East Asian deer woman with dark brown twin ponytails, straight bangs, symmetrical deer ears and small three-tined antlers, neutral expression, natural window light, sharp facial detail.", num_inference_steps=8, guidance_scale=0.0).images[0]
image.save("output.png")