File size: 17,239 Bytes
9c24957
7c3034b
9c24957
 
9852c31
9c24957
54e78ff
 
 
 
 
 
 
 
 
 
 
f99eae7
9b8670f
 
 
f99eae7
f10ab74
f99eae7
f10ab74
 
 
9c24957
 
72b584c
acd94cc
 
48a1bd8
acd94cc
36ae9ff
 
 
48a1bd8
9c24957
 
 
 
 
cf7033c
9c24957
 
 
 
 
 
 
 
 
 
 
 
 
412dbda
9c24957
72b584c
 
f15bc76
 
9c24957
 
 
 
 
72b584c
9c24957
 
 
 
2b64332
 
9c24957
412dbda
9c24957
 
2881c08
412dbda
 
2b64332
114c79c
412dbda
11a3326
 
114c79c
11a3326
 
2b64332
9c24957
2b64332
114c79c
2b64332
 
 
 
 
114c79c
2b64332
114c79c
11a3326
 
114c79c
11a3326
 
9c24957
 
 
 
 
114c79c
9c24957
72b584c
114c79c
 
 
 
72b584c
11a3326
 
114c79c
11a3326
 
9c24957
 
f15bc76
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bd73f2a
f15bc76
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bd73f2a
f15bc76
 
 
 
 
 
9c24957
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
72b584c
9c24957
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26c6da0
 
9c24957
 
72b584c
9c24957
 
 
 
 
 
 
e9bde7f
 
 
 
 
 
 
 
9c24957
 
 
 
 
 
 
 
 
72b584c
9c24957
 
 
 
 
 
cc66b03
 
72b584c
cc66b03
72b584c
9c24957
0b2e68a
 
9c24957
 
 
 
 
 
 
 
 
 
 
 
21b2071
9c24957
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
---
license: creativeml-openrail-m
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
widget:
- text: "A high tech solarpunk utopia in the Amazon rainforest"
  example_title: Amazon rainforest
- text: "A pikachu fine dining with a view to the Eiffel Tower"
  example_title: Pikachu in Paris
- text: "A mecha robot in a favela in expressionist style"
  example_title: Expressionist robot
- text: "an insect robot preparing a delicious meal"
  example_title: Insect robot
- text: "A small cabin on top of a snowy mountain in the style of Disney, artstation"
  example_title: Snowy disney cabin
extra_gated_prompt: |-
  This model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage.
  The CreativeML OpenRAIL License specifies: 

  1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content 
  2. The authors claim no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
  3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully)
  Please read the full license carefully here: https://huggingface.co/spaces/CompVis/stable-diffusion-license
      
extra_gated_heading: Please read the LICENSE to access this model
---

# Stable Diffusion v1-4 Model Card

Stable Diffusion is a latent text-to-image diffusion model capable of generating photo-realistic images given any text input.
For more information about how Stable Diffusion functions, please have a look at [🤗's Stable Diffusion with 🧨Diffusers blog](https://huggingface.co/blog/stable_diffusion).

The **Stable-Diffusion-v1-4** checkpoint was initialized with the weights of the [Stable-Diffusion-v1-2](https:/steps/huggingface.co/CompVis/stable-diffusion-v1-2) 
checkpoint and subsequently fine-tuned on 225k steps at resolution 512x512 on "laion-aesthetics v2 5+" and 10% dropping of the text-conditioning to improve [classifier-free guidance sampling](https://arxiv.org/abs/2207.12598).

This weights here are intended to be used with the 🧨 Diffusers library. If you are looking for the weights to be loaded into the CompVis Stable Diffusion codebase, [come here](https://huggingface.co/CompVis/stable-diffusion-v-1-4-original)

## Model Details
- **Developed by:** Robin Rombach, Patrick Esser
- **Model type:** Diffusion-based text-to-image generation model
- **Language(s):** English
- **License:** [The CreativeML OpenRAIL M license](https://huggingface.co/spaces/CompVis/stable-diffusion-license) is an [Open RAIL M license](https://www.licenses.ai/blog/2022/8/18/naming-convention-of-responsible-ai-licenses), adapted from the work that [BigScience](https://bigscience.huggingface.co/) and [the RAIL Initiative](https://www.licenses.ai/) are jointly carrying in the area of responsible AI licensing. See also [the article about the BLOOM Open RAIL license](https://bigscience.huggingface.co/blog/the-bigscience-rail-license) on which our license is based.
- **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 a fixed, pretrained text encoder ([CLIP ViT-L/14](https://arxiv.org/abs/2103.00020)) as suggested in the [Imagen paper](https://arxiv.org/abs/2205.11487).
- **Resources for more information:** [GitHub Repository](https://github.com/CompVis/stable-diffusion), [Paper](https://arxiv.org/abs/2112.10752).
- **Cite as:**

      @InProceedings{Rombach_2022_CVPR,
          author    = {Rombach, Robin and Blattmann, Andreas and Lorenz, Dominik and Esser, Patrick and Ommer, Bj\"orn},
          title     = {High-Resolution Image Synthesis With Latent Diffusion Models},
          booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
          month     = {June},
          year      = {2022},
          pages     = {10684-10695}
      }

## Examples

We recommend using [🤗's Diffusers library](https://github.com/huggingface/diffusers) to run Stable Diffusion.

### PyTorch

```bash
pip install --upgrade diffusers transformers scipy
```

Run this command to log in with your HF Hub token if you haven't before:

```bash
huggingface-cli login
```

Running the pipeline with the default PNDM scheduler:

```python
import torch
from diffusers import StableDiffusionPipeline

model_id = "CompVis/stable-diffusion-v1-4"
device = "cuda"


pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, revision="fp16")
pipe = pipe.to(device)

prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]  
    
image.save("astronaut_rides_horse.png")
```

**Note**:
If you are limited by GPU memory and have less than 4GB of GPU RAM available, please make sure to load the StableDiffusionPipeline in float16 precision instead of the default float32 precision as done above. You can do so by telling diffusers to expect the weights to be in float16 precision:


```py
import torch

pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, revision="fp16")
pipe = pipe.to(device)
pipe.enable_attention_slicing()

prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]  
    
image.save("astronaut_rides_horse.png")
```

To swap out the noise scheduler, pass it to `from_pretrained`:

```python
from diffusers import StableDiffusionPipeline, EulerDiscreteScheduler

model_id = "CompVis/stable-diffusion-v1-4"

# Use the Euler scheduler here instead
scheduler = EulerDiscreteScheduler.from_pretrained(model_id, subfolder="scheduler")
pipe = StableDiffusionPipeline.from_pretrained(model_id, scheduler=scheduler, torch_dtype=torch.float16, revision="fp16")
pipe = pipe.to("cuda")

prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]  
    
image.save("astronaut_rides_horse.png")
```

### JAX/Flax

To use StableDiffusion on TPUs and GPUs for faster inference you can leverage JAX/Flax.

Running the pipeline with default PNDMScheduler

```python
import jax
import numpy as np
from flax.jax_utils import replicate
from flax.training.common_utils import shard

from diffusers import FlaxStableDiffusionPipeline

pipeline, params = FlaxStableDiffusionPipeline.from_pretrained(
    "CompVis/stable-diffusion-v1-4", revision="flax", dtype=jax.numpy.bfloat16
)

prompt = "a photo of an astronaut riding a horse on mars"

prng_seed = jax.random.PRNGKey(0)
num_inference_steps = 50

num_samples = jax.device_count()
prompt = num_samples * [prompt]
prompt_ids = pipeline.prepare_inputs(prompt)

# shard inputs and rng
params = replicate(params)
prng_seed = jax.random.split(prng_seed, num_samples)
prompt_ids = shard(prompt_ids)

images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).images
images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
```

**Note**:
If you are limited by TPU memory, please make sure to load the `FlaxStableDiffusionPipeline` in `bfloat16` precision instead of the default `float32` precision as done above. You can do so by telling diffusers to load the weights from "bf16" branch.

```python
import jax
import numpy as np
from flax.jax_utils import replicate
from flax.training.common_utils import shard

from diffusers import FlaxStableDiffusionPipeline

pipeline, params = FlaxStableDiffusionPipeline.from_pretrained(
    "CompVis/stable-diffusion-v1-4", revision="bf16", dtype=jax.numpy.bfloat16
)

prompt = "a photo of an astronaut riding a horse on mars"

prng_seed = jax.random.PRNGKey(0)
num_inference_steps = 50

num_samples = jax.device_count()
prompt = num_samples * [prompt]
prompt_ids = pipeline.prepare_inputs(prompt)

# shard inputs and rng
params = replicate(params)
prng_seed = jax.random.split(prng_seed, num_samples)
prompt_ids = shard(prompt_ids)

images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).images
images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
```

# Uses

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

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

Excluded uses are described below.

 ### Misuse, Malicious Use, and Out-of-Scope Use
_Note: This section is taken from the [DALLE-MINI model card](https://huggingface.co/dalle-mini/dalle-mini), but applies in the same way to Stable Diffusion v1_.


The model should not be used to intentionally create or disseminate images that create hostile or alienating environments for people. This includes generating images that people would foreseeably find disturbing, distressing, or offensive; or content that propagates historical or current stereotypes.

#### 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.

#### Misuse and Malicious Use
Using the model to generate content that is cruel to individuals is a misuse of this model. This includes, but is not limited to:

- Generating demeaning, dehumanizing, or otherwise harmful representations of people or their environments, cultures, religions, etc.
- Intentionally promoting or propagating discriminatory content or harmful stereotypes.
- Impersonating individuals without their consent.
- Sexual content without consent of the people who might see it.
- Mis- and disinformation
- Representations of egregious violence and gore
- Sharing of copyrighted or licensed material in violation of its terms of use.
- Sharing content that is an alteration of copyrighted or licensed material in violation of its terms of use.

## Limitations and Bias

### Limitations

- The model does not achieve perfect photorealism
- The model cannot render legible text
- The model does not perform well on 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 model was trained mainly with English captions and will not work as well in other languages.
- The autoencoding part of the model is lossy
- The model was trained on a large-scale dataset
  [LAION-5B](https://laion.ai/blog/laion-5b/) which contains adult material
  and is not fit for product use without additional safety mechanisms and
  considerations.
- No additional measures were used to deduplicate the dataset. As a result, we observe some degree of memorization for images that are duplicated in the training data.
  The training data can be searched at [https://rom1504.github.io/clip-retrieval/](https://rom1504.github.io/clip-retrieval/) to possibly assist in the detection of memorized images.

### Bias

While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases. 
Stable Diffusion v1 was trained on subsets of [LAION-2B(en)](https://laion.ai/blog/laion-5b/), 
which consists of images that are primarily limited to English descriptions. 
Texts and images from communities and cultures that use other languages are likely to be insufficiently accounted for. 
This affects the overall output of the model, as white and western cultures are often set as the default. Further, the 
ability of the model to generate content with non-English prompts is significantly worse than with English-language prompts.

### Safety Module

The intended use of this model is with the [Safety Checker](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/stable_diffusion/safety_checker.py) in Diffusers. 
This checker works by checking model outputs against known hard-coded NSFW concepts.
The concepts are intentionally hidden to reduce the likelihood of reverse-engineering this filter.
Specifically, the checker compares the class probability of harmful concepts in the embedding space of the `CLIPTextModel` *after generation* of the images. 
The concepts are passed into the model with the generated image and compared to a hand-engineered weight for each NSFW concept.


## Training

**Training Data**
The model developers used the following dataset for training the model:

- LAION-2B (en) and subsets thereof (see next section)

**Training Procedure**
Stable Diffusion v1-4 is a latent diffusion model which combines an autoencoder with a diffusion model that is trained in the latent space of the autoencoder. During training, 

- Images are encoded through an encoder, which turns images into latent representations. The autoencoder uses a relative downsampling factor of 8 and maps images of shape H x W x 3 to latents of shape H/f x W/f x 4
- Text prompts are encoded through a ViT-L/14 text-encoder.
- The non-pooled output of the text encoder is fed into the UNet backbone of the latent diffusion model via cross-attention.
- The loss is a reconstruction objective between the noise that was added to the latent and the prediction made by the UNet.

We currently provide four checkpoints, which were trained as follows.
- [`stable-diffusion-v1-1`](https://huggingface.co/CompVis/stable-diffusion-v1-1): 237,000 steps at resolution `256x256` on [laion2B-en](https://huggingface.co/datasets/laion/laion2B-en).
  194,000 steps at resolution `512x512` on [laion-high-resolution](https://huggingface.co/datasets/laion/laion-high-resolution) (170M examples from LAION-5B with resolution `>= 1024x1024`).
- [`stable-diffusion-v1-2`](https://huggingface.co/CompVis/stable-diffusion-v1-2): Resumed from `stable-diffusion-v1-1`.
  515,000 steps at resolution `512x512` on "laion-improved-aesthetics" (a subset of laion2B-en,
filtered to images with an original size `>= 512x512`, estimated aesthetics score `> 5.0`, and an estimated watermark probability `< 0.5`. The watermark estimate is from the LAION-5B metadata, the aesthetics score is estimated using an [improved aesthetics estimator](https://github.com/christophschuhmann/improved-aesthetic-predictor)).
- [`stable-diffusion-v1-3`](https://huggingface.co/CompVis/stable-diffusion-v1-3): Resumed from `stable-diffusion-v1-2`. 195,000 steps at resolution `512x512` on "laion-improved-aesthetics" and 10 % dropping of the text-conditioning to improve [classifier-free guidance sampling](https://arxiv.org/abs/2207.12598).
- [`stable-diffusion-v1-4`](https://huggingface.co/CompVis/stable-diffusion-v1-4) Resumed from `stable-diffusion-v1-2`.225,000 steps at resolution `512x512` on "laion-aesthetics v2 5+"  and 10 % dropping of the text-conditioning to improve [classifier-free guidance sampling](https://arxiv.org/abs/2207.12598).

- **Hardware:** 32 x 8 x A100 GPUs
- **Optimizer:** AdamW
- **Gradient Accumulations**: 2
- **Batch:** 32 x 8 x 2 x 4 = 2048
- **Learning rate:** warmup to 0.0001 for 10,000 steps and then kept constant

## Evaluation Results 
Evaluations with different classifier-free guidance scales (1.5, 2.0, 3.0, 4.0,
5.0, 6.0, 7.0, 8.0) and 50 PLMS sampling
steps show the relative improvements of the checkpoints:

![pareto](https://huggingface.co/CompVis/stable-diffusion/resolve/main/v1-variants-scores.jpg)

Evaluated using 50 PLMS steps and 10000 random prompts from the COCO2017 validation set, evaluated at 512x512 resolution.  Not optimized for FID scores.
## Environmental Impact

**Stable Diffusion v1** **Estimated Emissions**
Based on that information, we estimate the following CO2 emissions using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). The hardware, runtime, cloud provider, and compute region were utilized to estimate the carbon impact.

- **Hardware Type:** A100 PCIe 40GB
- **Hours used:** 150000
- **Cloud Provider:** AWS
- **Compute Region:** US-east
- **Carbon Emitted (Power consumption x Time x Carbon produced based on location of power grid):** 11250 kg CO2 eq.


## Citation

```bibtex
    @InProceedings{Rombach_2022_CVPR,
        author    = {Rombach, Robin and Blattmann, Andreas and Lorenz, Dominik and Esser, Patrick and Ommer, Bj\"orn},
        title     = {High-Resolution Image Synthesis With Latent Diffusion Models},
        booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
        month     = {June},
        year      = {2022},
        pages     = {10684-10695}
    }
```

*This model card was written by: Robin Rombach and Patrick Esser and is based on the [DALL-E Mini model card](https://huggingface.co/dalle-mini/dalle-mini).*