Skip to content

GitLab

"...composable_kernel-1.git" did not exist on "cac014f17355d6504b618f5945c6326a285db7e9"
Unverified Commit 2533f925 authored by Patrick von Platen's avatar Patrick von Platen Committed by GitHub
Browse files

[Stable Diffusion Guide] 101 Stable Diffusion Guide directly into the docs (#1927) 



* finish

* improve

* Apply suggestions from code review
Co-authored-by: default avatarPedro Cuenca <pedro@huggingface.co>
Co-authored-by: default avatarPedro Cuenca <pedro@huggingface.co>
parent f6af0d1f
Hide whitespace changes
Inline Side-by-side
Showing with 328 additions and 10 deletions
docs/source/en/stable_diffusion.mdx
View file @ 2533f925
<!--Copyright 2022 The HuggingFace Team. All rights reserved. <!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
-->
# The Stable Diffusion Guide 🎨
<a target="_blank" href="https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_101_guide.ipynb">
<img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
</a>
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with ## Intro
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0 Stable Diffusion is a [Latent Diffusion model](https://github.com/CompVis/latent-diffusion) developed by researchers from the Machine Vision and Learning group at LMU Munich, *a.k.a* CompVis.
Model checkpoints were publicly released at the end of August 2022 by a collaboration of Stability AI, CompVis, and Runway with support from EleutherAI and LAION. For more information, you can check out [the official blog post](https://stability.ai/blog/stable-diffusion-public-release).
Since its public release the community has done an incredible job at working together to make the stable diffusion checkpoints **faster**, **more memory efficient**, and **more performant**.
🧨 Diffusers offers a simple API to run stable diffusion with all memory, computing, and quality improvements.
This notebook walks you through the improvements one-by-one so you can best leverage [`StableDiffusionPipeline`] for **inference**.
## Prompt Engineering 🎨
When running *Stable Diffusion* in inference, we usually want to generate a certain type, or style of image and then improve upon it. Improving upon a previously generated image means running inference over and over again with a different prompt and potentially a different seed until we are happy with our generation.
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on So to begin with, it is most important to speed up stable diffusion as much as possible to generate as many pictures as possible in a given amount of time.
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
-->
# Stable Diffusion This can be done by both improving the **computational efficiency** (speed) and the **memory efficiency** (GPU RAM).
Please visit this [very in-detail blog post](https://huggingface.co/blog/stable_diffusion) on Stable Diffusion! Let's start by looking into computational efficiency first.
Throughout the notebook, we will focus on [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5):
``` python
model_id = "runwayml/stable-diffusion-v1-5"
```
Let's load the pipeline.
## Speed Optimization
``` python
from diffusers import StableDiffusionPipeline
pipe = StableDiffusionPipeline.from_pretrained(model_id)
```
We aim at generating a beautiful photograph of an *old warrior chief* and will later try to find the best prompt to generate such a photograph. For now, let's keep the prompt simple:
``` python
prompt = "portrait photo of a old warrior chief"
```
To begin with, we should make sure we run inference on GPU, so let's move the pipeline to GPU, just like you would with any PyTorch module.
``` python
pipe = pipe.to("cuda")
```
To generate an image, you should use the [~`StableDiffusionPipeline.__call__`] method.
To make sure we can reproduce more or less the same image in every call, let's make use of the generator. See the documentation on reproducibility [here](./conceptual/reproducibility) for more information.
``` python
generator = torch.Generator("cuda").manual_seed(0)
```
Now, let's take a spin on it.
``` python
image = pipe(prompt, generator=generator).images[0]
image
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_1.png)
Cool, this now took roughly 30 seconds on a T4 GPU (you might see faster inference if your allocated GPU is better than a T4).
The default run we did above used full float32 precision and ran the default number of inference steps (50). The easiest speed-ups come from switching to float16 (or half) precision and simply running fewer inference steps. Let's load the model now in float16 instead.
``` python
import torch
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
pipe = pipe.to("cuda")
```
And we can again call the pipeline to generate an image.
``` python
generator = torch.Generator("cuda").manual_seed(0)
image = pipe(prompt, generator=generator).images[0]
image
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_2.png)
Cool, this is almost three times as fast for arguably the same image quality.
We strongly suggest always running your pipelines in float16 as so far we have very rarely seen degradations in quality because of it.
Next, let's see if we need to use 50 inference steps or whether we could use significantly fewer. The number of inference steps is associated with the denoising scheduler we use. Choosing a more efficient scheduler could help us decrease the number of steps.
Let's have a look at all the schedulers the stable diffusion pipeline is compatible with.
``` python
pipe.scheduler.compatibles
```
```
[diffusers.schedulers.scheduling_dpmsolver_singlestep.DPMSolverSinglestepScheduler,
diffusers.schedulers.scheduling_lms_discrete.LMSDiscreteScheduler,
diffusers.schedulers.scheduling_heun_discrete.HeunDiscreteScheduler,
diffusers.schedulers.scheduling_pndm.PNDMScheduler,
diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler,
diffusers.schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteScheduler,
diffusers.schedulers.scheduling_dpmsolver_multistep.DPMSolverMultistepScheduler,
diffusers.schedulers.scheduling_ddpm.DDPMScheduler,
diffusers.schedulers.scheduling_ddim.DDIMScheduler]
```
Cool, that's a lot of schedulers.
🧨 Diffusers is constantly adding a bunch of novel schedulers/samplers that can be used with Stable Diffusion. For more information, we recommend taking a look at the official documentation [here](https://huggingface.co/docs/diffusers/main/en/api/schedulers/overview).
Alright, right now Stable Diffusion is using the `PNDMScheduler` which usually requires around 50 inference steps. However, other schedulers such as `DPMSolverMultistepScheduler` or `DPMSolverSinglestepScheduler` seem to get away with just 20 to 25 inference steps. Let's try them out.
You can set a new scheduler by making use of the [from_config](https://huggingface.co/docs/diffusers/main/en/api/configuration#diffusers.ConfigMixin.from_config) function.
``` python
from diffusers import DPMSolverMultistepScheduler
pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)
```
Now, let's try to reduce the number of inference steps to just 20.
``` python
generator = torch.Generator("cuda").manual_seed(0)
image = pipe(prompt, generator=generator, num_inference_steps=20).images[0]
image
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_3.png)
The image now does look a little different, but it's arguably still of equally high quality. We now cut inference time to just 4 seconds though 😍.
## Memory Optimization
Less memory used in generation indirectly implies more speed, since we're often trying to maximize how many images we can generate per second. Usually, the more images per inference run, the more images per second too.
The easiest way to see how many images we can generate at once is to simply try it out, and see when we get a *"Out-of-memory (OOM)"* error.
We can run batched inference by simply passing a list of prompts and generators. Let's define a quick function that generates a batch for us.
``` python
def get_inputs(batch_size=1):
generator = [torch.Generator("cuda").manual_seed(i) for i in range(batch_size)]
prompts = batch_size * [prompt]
num_inference_steps = 20
return {"prompt": prompts, "generator": generator, "num_inference_steps": num_inference_steps}
```
This function returns a list of prompts and a list of generators, so we can reuse the generator that produced a result we like.
We also need a method that allows us to easily display a batch of images.
``` python
from PIL import Image
def image_grid(imgs, rows=2, cols=2):
w, h = imgs[0].size
grid = Image.new('RGB', size=(cols*w, rows*h))
for i, img in enumerate(imgs):
grid.paste(img, box=(i%cols*w, i//cols*h))
return grid
```
Cool, let's see how much memory we can use starting with `batch_size=4`.
``` python
images = pipe(**get_inputs(batch_size=4)).images
image_grid(images)
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_4.png)
Going over a batch_size of 4 will error out in this notebook (assuming we are running it on a T4 GPU). Also, we can see we only generate slightly more images per second (3.75s/image) compared to 4s/image previously.
However, the community has found some nice tricks to improve the memory constraints further. After stable diffusion was released, the community found improvements within days and shared them freely over GitHub - open-source at its finest! I believe the original idea came from [this](https://github.com/basujindal/stable-diffusion/pull/117) GitHub thread.
By far most of the memory is taken up by the cross-attention layers. Instead of running this operation in batch, one can run it sequentially to save a significant amount of memory.
It can easily be enabled by calling `enable_attention_slicing` as is documented [here](https://huggingface.co/docs/diffusers/main/en/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline.enable_attention_slicing).
``` python
pipe.enable_attention_slicing()
```
Great, now that attention slicing is enabled, let's try to double the batch size again, going for `batch_size=8`.
``` python
images = pipe(**get_inputs(batch_size=8)).images
image_grid(images, rows=2, cols=4)
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_5.png)
Nice, it works. However, the speed gain is again not very big (it might however be much more significant on other GPUs).
We're at roughly 3.5 seconds per image 🔥 which is probably the fastest we can be with a simple T4 without sacrificing quality.
Next, let's look into how to improve the quality!
## Quality Improvements
Now that our image generation pipeline is blazing fast, let's try to get maximum image quality.
First of all, image quality is extremely subjective, so it's difficult to make general claims here.
The most obvious step to take to improve quality is to use *better checkpoints*. Since the release of Stable Diffusion, many improved versions have been released, which are summarized here:
- *Official Release - 22 Aug 2022*: [Stable-Diffusion 1.4](https://huggingface.co/CompVis/stable-diffusion-v1-4)
- *20 October 2022*: [Stable-Diffusion 1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5)
- *24 Nov 2022*: [Stable-Diffusion 2.0](https://huggingface.co/stabilityai/stable-diffusion-2-0)
- *7 Dec 2022*: [Stable-Diffusion 2.1](https://huggingface.co/stabilityai/stable-diffusion-2-1)
Newer versions don't necessarily mean better image quality with the same parameters. People mentioned that *2.0* is slightly worse than *1.5* for certain prompts, but given the right prompt engineering *2.0* and *2.1* seem to be better.
Overall, we strongly recommend just trying the models out and reading up on advice online (e.g. it has been shown that using negative prompts is very important for 2.0 and 2.1 to get the highest possible quality. See for example [this nice blog post](https://minimaxir.com/2022/11/stable-diffusion-negative-prompt/).
Additionally, the community has started fine-tuning many of the above versions on certain styles with some of them having an extremely high quality and gaining a lot of traction.
We recommend having a look at all [diffusers checkpoints sorted by downloads and trying out the different checkpoints](https://huggingface.co/models?library=diffusers).
For the following, we will stick to v1.5 for simplicity.
Next, we can also try to optimize single components of the pipeline, e.g. switching out the latent decoder. For more details on how the whole Stable Diffusion pipeline works, please have a look at [this blog post](https://huggingface.co/blog/stable_diffusion).
Let's load [stabilityai's newest auto-decoder](https://huggingface.co/stabilityai/stable-diffusion-2-1).
``` python
from diffusers import AutoencoderKL
vae = AutoencoderKL.from_pretrained("stabilityai/sd-vae-ft-mse", torch_dtype=torch.float16).to("cuda")
```
Now we can set it to the vae of the pipeline to use it.
``` python
pipe.vae = vae
```
Let's run the same prompt as before to compare quality.
``` python
images = pipe(**get_inputs(batch_size=8)).images
image_grid(images, rows=2, cols=4)
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_6.png)
Seems like the difference is only very minor, but the new generations are arguably a bit *sharper*.
Cool, finally, let's look a bit into prompt engineering.
Our goal was to generate a photo of an old warrior chief. Let's now try to bring a bit more color into the photos and make the look more impressive.
Originally our prompt was "*portrait photo of an old warrior chief*".
To improve the prompt, it often helps to add cues that could have been used online to save high-quality photos, as well as add more details.
Essentially, when doing prompt engineering, one has to think:
- How was the photo or similar photos of the one I want probably stored on the internet?
- What additional detail can I give that steers the models into the style that I want?
Cool, let's add more details.
``` python
prompt += ", tribal panther make up, blue on red, side profile, looking away, serious eyes"
```
and let's also add some cues that usually help to generate higher quality images.
``` python
prompt += " 50mm portrait photography, hard rim lighting photography--beta --ar 2:3 --beta --upbeta"
prompt
```
Cool, let's now try this prompt.
``` python
images = pipe(**get_inputs(batch_size=8)).images
image_grid(images, rows=2, cols=4)
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_7.png)
Pretty impressive! We got some very high-quality image generations there. The 2nd image is my personal favorite, so I'll re-use this seed and see whether I can tweak the prompts slightly by using "oldest warrior", "old", "", and "young" instead of "old".
``` python
prompts = [
"portrait photo of the oldest warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes 50mm portrait photography, hard rim lighting photography--beta --ar 2:3 --beta --upbeta",
"portrait photo of a old warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes 50mm portrait photography, hard rim lighting photography--beta --ar 2:3 --beta --upbeta",
"portrait photo of a warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes 50mm portrait photography, hard rim lighting photography--beta --ar 2:3 --beta --upbeta",
"portrait photo of a young warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes 50mm portrait photography, hard rim lighting photography--beta --ar 2:3 --beta --upbeta",
]
generator = [torch.Generator("cuda").manual_seed(1) for _ in range(len(prompts))] # 1 because we want the 2nd image
images = pipe(prompt=prompts, generator=generator, num_inference_steps=25).images
image_grid(images)
```
![img](https://huggingface.co/datasets/diffusers/docs-images/resolve/main/stable_diffusion_101/sd_101_8.png)
The first picture looks nice! The eye movement slightly changed and looks nice. This finished up our 101-guide on how to use Stable Diffusion 🤗.
For more information on optimization or other guides, I recommend taking a look at the following:
- [Blog post about Stable Diffusion](https://huggingface.co/blog/stable_diffusion): In-detail blog post explaining Stable Diffusion.
- [FlashAttention](https://huggingface.co/docs/diffusers/optimization/xformers): XFormers flash attention can optimize your model even further with more speed and memory improvements.
- [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth) - Quickly customize the model by fine-tuning it.
- [General info on Stable Diffusion](https://huggingface.co/docs/diffusers/main/en/api/pipelines/stable_diffusion/overview) - Info on other tasks that are powered by Stable Diffusion.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment