README_from_modelscope.md

---
frameworks:
- Pytorch
license: Apache License 2.0
tasks:
- text-to-image-synthesis

#model-type:
##如 gpt、phi、llama、chatglm、baichuan 等
#- gpt

#domain:
##如 nlp、cv、audio、multi-modal
#- nlp

#language:
##语言代码列表 https://help.aliyun.com/document_detail/215387.html?spm=a2c4g.11186623.0.0.9f8d7467kni6Aa
#- cn 

#metrics:
##如 CIDEr、Blue、ROUGE 等
#- CIDEr

#tags:
##各种自定义，包括 pretrained、fine-tuned、instruction-tuned、RL-tuned 等训练方法和其他
#- pretrained

#tools:
##如 vllm、fastchat、llamacpp、AdaSeq 等
#- vllm
---
# AttriCtrl 数值型生图控制模型

## 简介

AttriCtrl 可以实现数值型图像指标的生图控制。

更多细节请参考我们的论文: [AttriCtrl: Fine-Grained Control of Aesthetic Attribute Intensity in Diffusion Models](https://arxiv.org/abs/2508.02151)

## 效果展示

### 亮度（Brightness）

|scale = 0.1|scale = 0.3|scale = 0.5|scale = 0.7|scale = 0.9|
|-|-|-|-|-|
|![](./assets/brightness/value_control_0.1.jpg)|![](./assets/brightness/value_control_0.3.jpg)|![](./assets/brightness/value_control_0.5.jpg)|![](./assets/brightness/value_control_0.7.jpg)|![](./assets/brightness/value_control_0.9.jpg)|

### 细节（Detail）

|scale = 0.1|scale = 0.3|scale = 0.5|scale = 0.7|scale = 0.9|
|-|-|-|-|-|
|![](./assets/detail/value_control_0.1.jpg)|![](./assets/detail/value_control_0.3.jpg)|![](./assets/detail/value_control_0.5.jpg)|![](./assets/detail/value_control_0.7.jpg)|![](./assets/detail/value_control_0.9.jpg)|

### 摄影感（Realism）

|scale = 0.1|scale = 0.3|scale = 0.5|scale = 0.7|scale = 0.9|
|-|-|-|-|-|
|![](./assets/realism/value_control_0.1.jpg)|![](./assets/realism/value_control_0.3.jpg)|![](./assets/realism/value_control_0.5.jpg)|![](./assets/realism/value_control_0.7.jpg)|![](./assets/realism/value_control_0.9.jpg)|

## 推理代码

```shell
git clone https://github.com/modelscope/DiffSynth-Studio.git  
cd DiffSynth-Studio
pip install -e .
```

```python
import torch
from diffsynth.pipelines.flux_image_new import FluxImagePipeline, ModelConfig


pipe = FluxImagePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/"),
        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
        ModelConfig(model_id="DiffSynth-Studio/AttriCtrl-FLUX.1-Dev", origin_file_pattern="models/detail.safetensors")
    ],
)

for i in [0.1, 0.3, 0.5, 0.7, 0.9]:
    image = pipe(prompt="a cat on the beach", seed=2, value_controller_inputs=[i])
    image.save(f"value_control_{i}.jpg")
```