Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,56 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# FLUX 艺术增强 LoRA
|
| 6 |
+
|
| 7 |
+
## 简介
|
| 8 |
+
|
| 9 |
+
这是一个为 FLUX.1-dev 训练的 LoRA 模型，能够使模型生成的图像更符合人类的审美，包括但不限于：丰富的细节、唯美的光影、美学的构图、清晰的画面。本模型不需要任何触发词。
|
| 10 |
+
|
| 11 |
+
* 论文：coming soon
|
| 12 |
+
* 开源代码：https://github.com/modelscope/DiffSynth-Studio
|
| 13 |
+
* 模型：
|
| 14 |
+
* ModelScope: https://www.modelscope.cn/models/DiffSynth-Studio/ArtAug-lora-FLUX.1dev-v1
|
| 15 |
+
* HuggingFace: coming soon
|
| 16 |
+
* 在线体验：点击右上角“一键生成”
|
| 17 |
+
|
| 18 |
+
## 使用方式
|
| 19 |
+
|
| 20 |
+
本模型使用 DiffSynth-Studio 训练而来，我们推荐使用 DiffSynth-Studio 进行生成。
|
| 21 |
+
|
| 22 |
+
```shell
|
| 23 |
+
git clone https://github.com/modelscope/DiffSynth-Studio.git
|
| 24 |
+
cd DiffSynth-Studio
|
| 25 |
+
pip install -e .
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
```python
|
| 29 |
+
import torch
|
| 30 |
+
from diffsynth import ModelManager, FluxImagePipeline, download_customized_models
|
| 31 |
+
|
| 32 |
+
lora_path = download_customized_models(
|
| 33 |
+
model_id="DiffSynth-Studio/ArtAug-lora-FLUX.1dev-v1",
|
| 34 |
+
origin_file_path="merged_lora.safetensors",
|
| 35 |
+
local_dir="models/lora"
|
| 36 |
+
)[0]
|
| 37 |
+
model_manager = ModelManager(torch_dtype=torch.bfloat16, device="cuda", model_id_list=["FLUX.1-dev"])
|
| 38 |
+
model_manager.load_lora(lora_path, lora_alpha=1.0)
|
| 39 |
+
pipe = FluxImagePipeline.from_model_manager(model_manager)
|
| 40 |
+
|
| 41 |
+
image = pipe(prompt="a house", seed=0)
|
| 42 |
+
image.save("image_artaug.jpg")
|
| 43 |
+
```
|
| 44 |
+
|
| 45 |
+
由于本模型使用了通用的 FLUX LoRA 格式封装，可以被大多数 LoRA 加载器加载，你可以将这个 LoRA 模型接入到你的工作流中。
|
| 46 |
+
|
| 47 |
+
## 图像样例
|
| 48 |
+
|
| 49 |
+
|FLUX.1-dev|FLUX.1-dev + ArtAug LoRA|
|
| 50 |
+
|-|-|
|
| 51 |
+
|||
|
| 52 |
+
|||
|
| 53 |
+
|||
|
| 54 |
+
|||
|
| 55 |
+
|||
|
| 56 |
+
|||
|