File size: 8,965 Bytes
3575584
 
882b239
 
 
 
 
 
 
3575584
882b239
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
---
license: apache-2.0
base_model: stabilityai/stable-diffusion-xl-base-1.0
tags:
- art
- t2i-adapter
- image-to-image
- stable-diffusion-xl-diffusers
- stable-diffusion-xl
---

# T2I-Adapter-SDXL - Openpose

T2I Adapter is a network providing additional conditioning to stable diffusion. Each t2i checkpoint takes a different type of conditioning as input and is used with a specific base stable diffusion checkpoint.

This checkpoint provides conditioning on openpose for the StableDiffusionXL checkpoint. This was a collaboration between **Tencent ARC** and [**Hugging Face**](https://huggingface.co/).

## Model Details
- **Developed by:** T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models
- **Model type:** Diffusion-based text-to-image generation model
- **Language(s):** English
- **License:** Apache 2.0
- **Resources for more information:** [GitHub Repository](https://github.com/TencentARC/T2I-Adapter), [Paper](https://arxiv.org/abs/2302.08453).
- **Model complexity:**
|   | SD-V1.4/1.5 | SD-XL | T2I-Adapter | T2I-Adapter-SDXL |
| --- | --- |--- |--- |--- |
| Parameters | 860M | 2.6B |77 M | 77/79 M | |
- **Cite as:**

  @misc{
    title={T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models}, 
    author={Chong Mou, Xintao Wang, Liangbin Xie, Yanze Wu, Jian Zhang, Zhongang Qi, Ying Shan, Xiaohu Qie},
    year={2023},
    eprint={2302.08453},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
  }


### Checkpoints

| Model Name | Control Image Overview| Control Image Example | Generated Image Example |
|---|---|---|---|
|[TencentARC/t2i-adapter-canny-sdxl-1.0](https://huggingface.co/TencentARC/t2i-adapter-canny-sdxl-1.0)<br/> *Trained with canny edge detection* | A monochrome image with white edges on a black background.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_canny.png"><img width="64" style="margin:0;padding:0;" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_canny.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_canny.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_canny.png"/></a>|
|[TencentARC/t2i-adapter-sketch-sdxl-1.0](https://huggingface.co/TencentARC/t2i-adapter-sketch-sdxl-1.0)<br/> *Trained with [PidiNet](https://github.com/zhuoinoulu/pidinet) edge detection* | A hand-drawn monochrome image with white outlines on a black background.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_sketch.png"><img width="64" style="margin:0;padding:0;" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_sketch.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_sketch.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_sketch.png"/></a>|
|[TencentARC/t2i-adapter-lineart-sdxl-1.0](https://huggingface.co/TencentARC/t2i-adapter-lineart-sdxl-1.0)<br/> *Trained with lineart edge detection* | A hand-drawn monochrome image with white outlines on a black background.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_lin.png"><img width="64" style="margin:0;padding:0;" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_lin.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_lin.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_lin.png"/></a>|
|[TencentARC/t2i-adapter-depth-midas-sdxl-1.0](https://huggingface.co/TencentARC/t2i-adapter-depth-midas-sdxl-1.0)<br/> *Trained with Midas depth estimation*  | A grayscale image with black representing deep areas and white representing shallow areas.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_mid.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_mid.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_mid.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_mid.png"/></a>|
|[TencentARC/t2i-adapter-depth-zoe-sdxl-1.0](https://huggingface.co/TencentARC/t2i-adapter-depth-zoe-sdxl-1.0)<br/> *Trained with Zoe depth estimation*  | A grayscale image with black representing deep areas and white representing shallow areas.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_zeo.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_zeo.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_zeo.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_zeo.png"/></a>|
|[Adapter/t2iadapter_openpose_sdxlv1](https://huggingface.co/Adapter/t2iadapter_openpose_sdxlv1)<br/> *Trained with OpenPose bone image*  | A [OpenPose bone](https://github.com/CMU-Perceptual-Computing-Lab/openpose) image.|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/openpose.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/openpose.png"/></a>|<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/res_pose.png"><img width="64" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/res_pose.png"/></a>|


## Example

To get started, first install the required dependencies:

```bash
pip install git+https://github.com/huggingface/diffusers.git@t2iadapterxl # for now
pip install -U controlnet_aux==0.0.7 # for conditioning models and detectors
pip install transformers accelerate safetensors
```

1. Images are first downloaded into the appropriate *control image* format.
 2. The *control image* and *prompt* are passed to the [`StableDiffusionXLAdapterPipeline`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py#L125).

Let's have a look at a simple example using the [Canny Adapter](https://huggingface.co/TencentARC/t2i-adapter-lineart-sdxl-1.0).

- Dependency
```py
from diffusers import StableDiffusionXLAdapterPipeline, T2IAdapter, EulerAncestralDiscreteScheduler, AutoencoderKL
from diffusers.utils import load_image, make_image_grid
from controlnet_aux.midas import MidasDetector
import torch

# load adapter
adapter = T2IAdapter.from_pretrained(
  "TencentARC/t2i-adapter-depth-midas-sdxl-1.0", torch_dtype=torch.float16, varient="fp16"
).to("cuda")

# load euler_a scheduler
model_id = 'stabilityai/stable-diffusion-xl-base-1.0'
euler_a = EulerAncestralDiscreteScheduler.from_pretrained(model_id, subfolder="scheduler")
vae=AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
pipe = StableDiffusionXLAdapterPipeline.from_pretrained(
    model_id, vae=vae, adapter=adapter, scheduler=euler_a, torch_dtype=torch.float16, variant="fp16", 
).to("cuda")
pipe.enable_xformers_memory_efficient_attention()

midas_depth = MidasDetector.from_pretrained(
  "valhalla/t2iadapter-aux-models", filename="dpt_large_384.pt", model_type="dpt_large"
).to("cuda")
```

- Condition Image
```py
url = "https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/org_mid.jpg"
image = load_image(url)
image = midas_depth(
  image, detect_resolution=512, image_resolution=1024
)
```
<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_mid.png"><img width="480" style="margin:0;padding:0;" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/cond_depth_mid.png"/></a>

- Generation
```py
prompt = "A photo of a room, 4k photo, highly detailed"
negative_prompt = "anime, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured"

gen_images = pipe(
  prompt=prompt,
  negative_prompt=negative_prompt,
  image=image,
  num_inference_steps=30,
  adapter_conditioning_scale=1,
  guidance_scale=7.5,  
).images[0]
gen_images.save('out_mid.png')
```
<a href="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_mid.png"><img width="480" style="margin:0;padding:0;" src="https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/res_depth_mid.png"/></a>

### Training

Our training script was built on top of the official training script that we provide [here](https://github.com/huggingface/diffusers/blob/main/examples/t2i_adapter/README_sdxl.md). 

The model is trained on 3M high-resolution image-text pairs from LAION-Aesthetics V2 with

- Training steps: 35000
- Batch size: Data parallel with a single gpu batch size of `16` for a total batch size of `256`.
- Learning rate: Constant learning rate of `1e-5`.
- Mixed precision: fp16