基于中国知名摄影师 Luna 的风格训练的一个摄影风格 Lora，Luna 专注于拍摄人像，注重光影塑造，风格纤细敏感，擅长捕捉转瞬即逝的女性瞬间，她也很喜欢拍摄相同风格的空镜。因为摄影师的取材范围大多来自中国，所以不可避免以亚洲女性的面孔为主。 摄影师 Instagram：@luna_atlantis 小红书同号 --- Photographer Style LoRA — LunaA photographic-style LoRA trained on the distinctive photography of Luna, one of China's most celebrated portrait photographers.Luna's work is defined by her masterful command of light — she sculpts and shapes illumination with a delicate, sensitive touch, using it to caress form, carve out mood, and bring each scene to life. Her portraiture is both refined and intimate, deftly capturing fleeting, spontaneous moments of her female subjects — the quiet glance, the breath held mid-laugh, the unguarded instant when emotion flickers across a face.Though primarily a portrait photographer, she is equally drawn to the same ambient stills within her visual world — the same light, the same mood, echoed in empty frames and atmospheric shots that match her portraiture in feeling and tone.Because Luna's subjects are overwhelmingly drawn from across China, the faces in her work are, by nature, predominantly Asian women — each rendered with the same sensitivity, grace, and luminous control of light that defines her signature style.

Luna Atlantis Photography Style | 摄影风格

luna_style_z_c1-st6000.safetensors

v1.0 ZIT

v1.0 Moss Base

v1.0 for Qwen Image 2.1

v1.0 for Krea 2


# 

## 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: [Tongyi-MAI/Z-Image](https://modelscope.ai/models/Tongyi-MAI/Z-Image)
* Training Code: [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio)
* Data amount：100
* Total training steps：6000
* License: Apache-2.0

## Inference Code

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

```bash
pip install diffsynth
```

Inference:

```python
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig
import torch

pipe = ZImagePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="Tongyi-MAI/Z-Image", origin_file_pattern="transformer/*.safetensors"),
        ModelConfig(model_id="Tongyi-MAI/Z-Image", origin_file_pattern="text_encoder/*.safetensors"),
        ModelConfig(model_id="Tongyi-MAI/Z-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
    ],
    tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, ModelConfig(model_id="bakamaru/luna_style_z", origin_file_pattern="luna_style_z_c1-st6000.safetensors"))
prompt = "a cat"
image = pipe(prompt=prompt, num_inference_steps=50, cfg_scale=4)
image.save("image.jpg")
```


Files with SHA256: 8170ef6823e71d45c4c2ba767f792abe03137f54fcb9237eedd016ba3e00f713