File size: 8,710 Bytes
c4b60d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
---
license: openrail++
tags:
- text-to-image
- stable-diffusion
duplicated_from: stabilityai/stable-diffusion-xl-base-1.0
---
# SD-XL 1.0-base Model Card
![row01](01.png)

## Model

![pipeline](pipeline.png)

[SDXL](https://arxiv.org/abs/2307.01952) consists of an [ensemble of experts](https://arxiv.org/abs/2211.01324) pipeline for latent diffusion: 
In a first step, the base model is used to generate (noisy) latents, 
which are then further processed with a refinement model (available here: https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/) specialized for the final denoising steps.
Note that the base model can be used as a standalone module.

Alternatively, we can use a two-stage pipeline as follows: 
First, the base model is used to generate latents of the desired output size. 
In the second step, we use a specialized high-resolution model and apply a technique called SDEdit (https://arxiv.org/abs/2108.01073, also known as "img2img") 
to the latents generated in the first step, using the same prompt. This technique is slightly slower than the first one, as it requires more function evaluations.

Source code is available at https://github.com/Stability-AI/generative-models .

### Model Description

- **Developed by:** Stability AI
- **Model type:** Diffusion-based text-to-image generative model
- **License:** [CreativeML Open RAIL++-M License](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/LICENSE.md)
- **Model Description:** This is a model that can be used to generate and modify images based on text prompts. It is a [Latent Diffusion Model](https://arxiv.org/abs/2112.10752) that uses two fixed, pretrained text encoders ([OpenCLIP-ViT/G](https://github.com/mlfoundations/open_clip) and [CLIP-ViT/L](https://github.com/openai/CLIP/tree/main)).
- **Resources for more information:** Check out our [GitHub Repository](https://github.com/Stability-AI/generative-models) and the [SDXL report on arXiv](https://arxiv.org/abs/2307.01952).

### Model Sources

For research purposes, we recommend our `generative-models` Github repository (https://github.com/Stability-AI/generative-models), which implements the most popular diffusion frameworks (both training and inference) and for which new functionalities like distillation will be added over time.
[Clipdrop](https://clipdrop.co/stable-diffusion) provides free SDXL inference.

- **Repository:** https://github.com/Stability-AI/generative-models
- **Demo:** https://clipdrop.co/stable-diffusion


## Evaluation
![comparison](comparison.png)
The chart above evaluates user preference for SDXL (with and without refinement) over SDXL 0.9 and Stable Diffusion 1.5 and 2.1. 
The SDXL base model performs significantly better than the previous variants, and the model combined with the refinement module achieves the best overall performance.


### 🧨 Diffusers 

Make sure to upgrade diffusers to >= 0.19.0:
```
pip install diffusers --upgrade
```

In addition make sure to install `transformers`, `safetensors`, `accelerate` as well as the invisible watermark:
```
pip install invisible_watermark transformers accelerate safetensors
```

To just use the base model, you can run:

```py
from diffusers import DiffusionPipeline
import torch

pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16, use_safetensors=True, variant="fp16")
pipe.to("cuda")

# if using torch < 2.0
# pipe.enable_xformers_memory_efficient_attention()

prompt = "An astronaut riding a green horse"

images = pipe(prompt=prompt).images[0]
```

To use the whole base + refiner pipeline as an ensemble of experts you can run:

```py
from diffusers import DiffusionPipeline
import torch

# load both base & refiner
base = DiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16, variant="fp16", use_safetensors=True
)
base.to("cuda")
refiner = DiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-refiner-1.0",
    text_encoder_2=base.text_encoder_2,
    vae=base.vae,
    torch_dtype=torch.float16,
    use_safetensors=True,
    variant="fp16",
)
refiner.to("cuda")

# Define how many steps and what % of steps to be run on each experts (80/20) here
n_steps = 40
high_noise_frac = 0.8

prompt = "A majestic lion jumping from a big stone at night"

# run both experts
image = base(
    prompt=prompt,
    num_inference_steps=n_steps,
    denoising_end=high_noise_frac,
    output_type="latent",
).images
image = refiner(
    prompt=prompt,
    num_inference_steps=n_steps,
    denoising_start=high_noise_frac,
    image=image,
).images[0]
```

When using `torch >= 2.0`, you can improve the inference speed by 20-30% with torch.compile. Simple wrap the unet with torch compile before running the pipeline:
```py
pipe.unet = torch.compile(pipe.unet, mode="reduce-overhead", fullgraph=True)
```

If you are limited by GPU VRAM, you can enable *cpu offloading* by calling `pipe.enable_model_cpu_offload`
instead of `.to("cuda")`:

```diff
- pipe.to("cuda")
+ pipe.enable_model_cpu_offload()
```

For more information on how to use Stable Diffusion XL with `diffusers`, please have a look at [the Stable Diffusion XL Docs](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/stable_diffusion_xl).

### Optimum
[Optimum](https://github.com/huggingface/optimum) provides a Stable Diffusion pipeline compatible with both [OpenVINO](https://docs.openvino.ai/latest/index.html) and [ONNX Runtime](https://onnxruntime.ai/).

#### OpenVINO

To install Optimum with the dependencies required for OpenVINO :

```bash
pip install optimum[openvino]
```

To load an OpenVINO model and run inference with OpenVINO Runtime, you need to replace `StableDiffusionXLPipeline` with Optimum `OVStableDiffusionXLPipeline`. In case you want to load a PyTorch model and convert it to the OpenVINO format on-the-fly, you can set `export=True`.

```diff
- from diffusers import StableDiffusionPipeline
+ from optimum.intel import OVStableDiffusionPipeline

model_id = "stabilityai/stable-diffusion-xl-base-1.0"
- pipeline = StableDiffusionPipeline.from_pretrained(model_id)
+ pipeline = OVStableDiffusionPipeline.from_pretrained(model_id)
prompt = "A majestic lion jumping from a big stone at night"
image = pipeline(prompt).images[0]
```

You can find more examples (such as static reshaping and model compilation) in optimum [documentation](https://huggingface.co/docs/optimum/main/en/intel/inference#stable-diffusion-xl).


#### ONNX

To install Optimum with the dependencies required for ONNX Runtime inference :

```bash
pip install optimum[onnxruntime]
```

To load an ONNX model and run inference with ONNX Runtime, you need to replace `StableDiffusionXLPipeline` with Optimum `ORTStableDiffusionXLPipeline`. In case you want to load a PyTorch model and convert it to the ONNX format on-the-fly, you can set `export=True`.

```diff
- from diffusers import StableDiffusionPipeline
+ from optimum.onnxruntime import ORTStableDiffusionPipeline

model_id = "stabilityai/stable-diffusion-xl-base-1.0"
- pipeline = StableDiffusionPipeline.from_pretrained(model_id)
+ pipeline = ORTStableDiffusionPipeline.from_pretrained(model_id)
prompt = "A majestic lion jumping from a big stone at night"
image = pipeline(prompt).images[0]
```

You can find more examples in optimum [documentation](https://huggingface.co/docs/optimum/main/en/onnxruntime/usage_guides/models#stable-diffusion-xl).


## Uses

### Direct Use

The model is intended for research purposes only. Possible research areas and tasks include

- Generation of artworks and use in design and other artistic processes.
- Applications in educational or creative tools.
- Research on generative models.
- Safe deployment of models which have the potential to generate harmful content.
- Probing and understanding the limitations and biases of generative models.

Excluded uses are described below.

### Out-of-Scope Use

The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.

## Limitations and Bias

### Limitations

- The model does not achieve perfect photorealism
- The model cannot render legible text
- The model struggles with more difficult tasks which involve compositionality, such as rendering an image corresponding to “A red cube on top of a blue sphere”
- Faces and people in general may not be generated properly.
- The autoencoding part of the model is lossy.

### Bias
While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.