<h3 id="fully-customizable-outfits-easily-change-or-remove-her-canonical-costume!">## Fully Customizable Outfits — Easily Change or Remove Her Canonical Costume!</h3>If you are already familiar with my *[Kira](<a target="_blank" rel="ugc" href="https://civarchive.com/models/3029127">https://civarchive.com/models/3029127</a>)** model, you will definitely recognize the signature facial features — be sure to check out her page as well if you haven't tried her yet!*📦 Liked this? <a target="_blank" rel="ugc" href="https://civarchive.com/user/bazya022020810">Come grab more</a>.<h3 id="update-v2.0-(recommended)">### Update v2.0 (Recommended)</h3>Version 1.0 was completely unsuccessful — the model struggled with facial consistency and outfit flexibility. **Version 2.0 has been retrained from scratch** on a brand-new, clean 55-image dataset.### Usage &amp; Outfit Switching* **Canonical Outfit:** Add <code>yenoutfit</code> at the beginning of your prompt to generate her classic black leather corset, fur collar, obsidian star choker, and high-slit skirt.* **Custom / Casual Outfits:** Simply **omit <code>yenoutfit</code>** from your prompt and describe any clothing you want (or leave clothing unspecified for casual black attire). Her face, black wavy hair, violet eyes, and star choker remain consistent across any outfit changes.* **Tip:** Do not include conflicting hair or eye colors in your prompt, and remove any unwanted clothing descriptions when using <code>yenoutfit</code>.* **Recommended LoRA Weight:** <code>0.85 – 1.0</code>### Example Prompts* **With Trigger (Canonical Outfit):**<code>yenoutfit, full-length candid portrait of a woman standing on wet medieval cobblestone courtyard at dusk, bare thigh visible through high skirt slit, black leather high-heeled boots, dramatic warm lantern rim light and cool blue twilight shadows, natural skin texture, 85mm lens.</code>* **Without Trigger (Custom Outfit):**<code>Full-length candid photo of a woman walking through a modern glass office lobby, wearing a crisp white button-up shirt, high-waisted black pencil skirt with a thigh slit, and black leather ankle boots, natural daylight, 50mm lens.</code>### Feedback &amp; SupportYour ratings, likes, and generation showcases in the gallery are always appreciated! Feel free to leave a comment if you have feedback or suggestions on which character to train next.This model received technical support from ModelScope.

Witches:  Yennefer cosplay  | MOSS / Krea 2 /Qwen 2.1

kira_yenneferV5M.safetensors

v2.0 MOSS

Krea2\v2.0\Yennefer

v1.0 Qwen 2.1

v1.0 Krea 2

v1.0 MOSS


# 

## Model Introduction

This model is trained using the [Civision/model training](https://modelscope.ai/civision/modelTraining) environment and computing resources provided by ModelScope.

* Model Type: LoRA
* Base Model: [openmoss/MOSS-Image-SFT](https://modelscope.ai/models/openmoss/MOSS-Image-SFT)
* Training Code: [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio)
* Data amount：55
* Total training steps：2000
* License: Apache-2.0

## Inference Code

Install [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio):

```bash
pip install diffsynth
```

Inference:

```python
from diffsynth.pipelines.moss_image import MOSSImagePipeline, ModelConfig
import torch

pipe = MOSSImagePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="openmoss/MOSS-Image-SFT", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
        ModelConfig(model_id="openmoss/MOSS-Image-SFT", origin_file_pattern="text_encoder/model*.safetensors"),
        ModelConfig(model_id="openmoss/MOSS-Image-SFT", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", computation_dtype=torch.float32),
    ],
    tokenizer_config=ModelConfig(model_id="openmoss/MOSS-Image-SFT", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, ModelConfig(model_id="bazya75/kira_yenneferV5M", origin_file_pattern="kira_yenneferV5M_c1-st2000.safetensors"))

prompt = "a cat"
image = pipe(prompt, height=1024, width=1024, seed=0, num_inference_steps=50, cfg_scale=3.0)
image.save("image.jpg")
```


kira_yenneferV5M

kira_yenneferV5M_c1-st2000.safetensors

c1-st2000

c1-st1000

Files with SHA256: c586dfbc46bc2038f5ef132504f37fa2c0eb77a0e14209a74f6635a15648d74e