README.md

---
base_model: krea/Krea-2-Raw
tags:
- text-to-image
- diffusers
- lora
- krea2
- template:sd-lora
license: apache-2.0
instance_prompt: "Sofia"
widget:
- text: "A cinematic close-up of Sofia as a cyberpunk hacker in a neon-drenched Tokyo alley, surrounded by holographic screens and floating data streams."
  output:
    url: sample_0.png
- text: "A lush oil painting of Sofia dressed as a Victorian noblewoman, reading a leather-bound book in a sun-drenched conservatory filled with exotic ferns."
  output:
    url: sample_1.png
- text: "An epic wide shot of Sofia as an astronaut standing on the crystalline surface of a distant purple planet, gazing at a massive swirling galaxy above."
  output:
    url: sample_2.png
---

# Krea 2 LoRA — Wonky468/sofia

<Gallery />

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 `Sofia` to invoke the concept.

## Samples

![sample](./sample_0.png)

> *"A cinematic close-up of Sofia as a cyberpunk hacker in a neon-drenched Tokyo alley, surrounded by holographic screens and floating data streams."*

![sample](./sample_1.png)

> *"A lush oil painting of Sofia dressed as a Victorian noblewoman, reading a leather-bound book in a sun-drenched conservatory filled with exotic ferns."*

![sample](./sample_2.png)

> *"An epic wide shot of Sofia as an astronaut standing on the crystalline surface of a distant purple planet, gazing at a massive swirling galaxy above."*

## Use it with diffusers

```py
import torch
from diffusers import Krea2Pipeline

pipe = Krea2Pipeline.from_pretrained("krea/Krea-2-Turbo", torch_dtype=torch.bfloat16).to("cuda")
pipe.load_lora_weights("Wonky468/sofia")
image = pipe("A cinematic close-up of Sofia as a cyberpunk hacker in a neon-drenched Tokyo alley, surrounded by holographic screens and floating data streams.", num_inference_steps=8, guidance_scale=0.0).images[0]
image.save("output.png")
```