
# ZL

## 模型介绍

本模型依托魔搭社区（ModelScope）AIGC专区[模型训练](https://modelscope.cn/aigc/modelTraining)环境与算力完成训练。

* 模型类型：LoRA
* 基础模型：[krea/Krea-2-Turbo](https://modelscope.cn/models/krea/Krea-2-Turbo)
* 训练代码：[DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio)
* 训练数据量：46
* 总训练步数：3000
* 开源协议：Apache-2.0

## 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.krea2 import Krea2Pipeline, ModelConfig
import torch

pipe = Krea2Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="krea/Krea-2-Turbo", origin_file_pattern="turbo.safetensors"),
        ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern="*.safetensors"),
        ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
    ],
    tokenizer_config=ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern=""),
)
pipe.load_lora(pipe.dit, ModelConfig(model_id="Neeshck/Krea2Turbo-ZL-V3", origin_file_pattern="Krea2Turbo-ZL-V3_c1-st3000.safetensors"))
prompt = "a cat"
image = pipe(
    prompt, seed=0,
    num_inference_steps=8, cfg_scale=1, mu=1.15,
)
image.save("image.jpg")
```


Files with SHA256: efc930f1ffd03e08ab8ef8d16e12434f73d902d724c5c60abd29d1921b78c290