---
title: CDA-Z-Image
canonical_url: "https://www.modelscope.cn/models/SecretDongGe/CDA-Z-Image"
md_url: "https://www.modelscope.cn/models/SecretDongGe/CDA-Z-Image.md"
repository: SecretDongGe/CDA-Z-Image
chinese_name: CDA-Z-Image
last_updated: 2026-06-05
license: "Apache License 2.0"
base_model:
  - Tongyi-MAI/Z-Image
base_model_relation: adapter
parameters: 158.8M
tensor_type:
  - BF16
library_name:
  - pytorch
  - lora
  - safetensors
supports_inference: txt2img
downloads: 33
stars: 1
tags:
  - LoRA
---

# CDA-Z-Image

> CDA-Z-Image - SecretDongGe 在 ModelScope 开源的模型。针对 CDA 进行专项训练的LoRA 推荐使用权重 0.60.8 本 LoRA 针对 CDA 精心训练而成，旨在提升 AI 生成女性人物时的面部美感、自然肤色、光影层次与整体气质。 经过大量高质量人像数据微调，CDA-Z-Image 能够有效增强人物的五官立体感、眼神神韵、皮肤质感与柔光氛围，适用于写实人像、影视角色、时尚摄影等多种场景。 特色包括： 自然清透的“奶油肌”肤质 柔和过渡的高光与阴影 高级感的服装与环境搭配…

SecretDongGe/CDA-Z-Image 是 ModelScope 魔搭社区上的 158.8M 参数机器学习模型，采用 Apache License 2.0 许可，基于 Tongyi-MAI/Z-Image 构建，并支持在线推理（txt2img）。

- **Repository**: SecretDongGe/CDA-Z-Image
- **License**: Apache License 2.0
- **Parameters**: 158.8M
- **Base model**: Tongyi-MAI/Z-Image
- **Online inference**: txt2img
- **Tags**: LoRA
- **Downloads**: 33
- **Stars**: 1
- **Last updated**: 2026-06-05

Source: https://www.modelscope.cn/models/SecretDongGe/CDA-Z-Image

---

# CDA

## 模型介绍

### 针对 CDA 进行专项训练的LoRA
**推荐使用权重 0.6~0.8**<br>
本 LoRA 针对 CDA 精心训练而成，旨在提升 AI 生成女性人物时的**面部美感、自然肤色、光影层次与整体气质**。
经过大量高质量人像数据微调，CDA-Z-Image 能够有效增强人物的**五官立体感、眼神神韵、皮肤质感与柔光氛围**，适用于写实人像、影视角色、时尚摄影等多种场景。<br>
特色包括：
- 自然清透的“奶油肌”肤质
- 柔和过渡的高光与阴影
- 高级感的服装与环境搭配
- 氛围感十足的“电影级”构图

### Style LoRA For CDA
**Recommended usage: Apply with weight 0.6–0.8**<br>
Trained for CDA, this LoRA enhances the visual appeal of female characters in AI-generated images.
CDA-Z-Image focuses on improving **facial harmony, natural skin texture, soft lighting, and overall aesthetic presence**, making it ideal for realistic portraits, character design, fashion photography, and cinematic scenes.<br>
Key features:
- Smooth "cream skin" texture with natural glow
- Soft highlight and shadow transitions
- Elegant poses and atmospheric backgrounds
- Cinematic composition and warm color tones

## 推理代码

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

```bash
pip install diffsynth
```

开始推理：

```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="SecretDongGe/CDA-Z-Image", origin_file_pattern="CDA_c1-st1000.safetensors"))
prompt = "a cat"
image = pipe(prompt=prompt, num_inference_steps=50, cfg_scale=4)
image.save("image.jpg")
```