diffusers/v0.20.0/en/using-diffusers/distilled_sd.html

<meta charset="utf-8" /><meta http-equiv="content-security-policy" content=""><meta name="hf:doc:metadata" content="{&quot;local&quot;:&quot;distilled-stable-diffusion-inference&quot;,&quot;sections&quot;:[{&quot;local&quot;:&quot;tiny-autoencoder&quot;,&quot;title&quot;:&quot;Tiny AutoEncoder&quot;}],&quot;title&quot;:&quot;Distilled Stable Diffusion inference&quot;}" data-svelte="svelte-1phssyn">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/assets/pages/__layout.svelte-hf-doc-builder.css">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/start-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/vendor-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/paths-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/pages/__layout.svelte-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/pages/using-diffusers/distilled_sd.mdx-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/Tip-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/IconCopyLink-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/CodeBlock-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/diffusers/v0.20.0/en/_app/chunks/DocNotebookDropdown-hf-doc-builder.js"> 





<h1 class="relative group"><a id="distilled-stable-diffusion-inference" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#distilled-stable-diffusion-inference"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a>
	<span>Distilled Stable Diffusion inference
	</span></h1>



<div class="flex space-x-1 absolute z-10 right-0 top-0">
	<div class="relative colab-dropdown ">
	<button class="  " type="button">
		
				<img alt="Open In Colab" class="!m-0" src="https://colab.research.google.com/assets/colab-badge.svg">
			</button>
	
	
	
	</div>
	<div class="relative colab-dropdown ">
	<button class="  " type="button">
		
				<img alt="Open In Studio Lab" class="!m-0" src="https://studiolab.sagemaker.aws/studiolab.svg">
			</button>
	
	
	
	</div></div>
<p>Stable Diffusion inference can be a computationally intensive process because it must iteratively denoise the latents to generate an image. To reduce the computational burden, you can use a <em>distilled</em> version of the Stable Diffusion model from <a href="https://huggingface.co/nota-ai" rel="nofollow">Nota AI</a>. The distilled version of their Stable Diffusion model eliminates some of the residual and attention blocks from the UNet, reducing the model size by 51% and improving latency on CPU/GPU by 43%.</p>


<div class="course-tip  bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p>Read this <a href="https://huggingface.co/blog/sd_distillation" rel="nofollow">blog post</a> to learn more about how knowledge distillation training works to produce a faster, smaller, and cheaper generative model.</p></div>
<p>Let’s load the distilled Stable Diffusion model and compare it against the original Stable Diffusion model:</p>

	<div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg>
	<div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div>
	Copied</div></button></div>
	<pre><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> StableDiffusionPipeline
<span class="hljs-keyword">import</span> torch

distilled = StableDiffusionPipeline.from_pretrained(
    <span class="hljs-string">&quot;nota-ai/bk-sdm-small&quot;</span>, torch_dtype=torch.float16, use_safetensors=<span class="hljs-literal">True</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)

original = StableDiffusionPipeline.from_pretrained(
    <span class="hljs-string">&quot;CompVis/stable-diffusion-v1-4&quot;</span>, torch_dtype=torch.float16, use_safetensors=<span class="hljs-literal">True</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)<!-- HTML_TAG_END --></pre></div>
<p>Given a prompt, get the inference time for the original model:</p>

	<div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg>
	<div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div>
	Copied</div></button></div>
	<pre><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> time

seed = <span class="hljs-number">2023</span>
generator = torch.manual_seed(seed)

NUM_ITERS_TO_RUN = <span class="hljs-number">3</span>
NUM_INFERENCE_STEPS = <span class="hljs-number">25</span>
NUM_IMAGES_PER_PROMPT = <span class="hljs-number">4</span>

prompt = <span class="hljs-string">&quot;a golden vase with different flowers&quot;</span>

start = time.time_ns()
<span class="hljs-keyword">for</span> _ <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(NUM_ITERS_TO_RUN):
    images = original(
        prompt,
        num_inference_steps=NUM_INFERENCE_STEPS,
        generator=generator,
        num_images_per_prompt=NUM_IMAGES_PER_PROMPT
    ).images
end = time.time_ns()
original_sd = <span class="hljs-string">f&quot;<span class="hljs-subst">{(end - start) / <span class="hljs-number">1e6</span>:<span class="hljs-number">.1</span>f}</span>&quot;</span>

<span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Execution time -- <span class="hljs-subst">{original_sd}</span> ms\n&quot;</span>)
<span class="hljs-string">&quot;Execution time -- 45781.5 ms&quot;</span><!-- HTML_TAG_END --></pre></div>
<p>Time the distilled model inference:</p>

	<div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg>
	<div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div>
	Copied</div></button></div>
	<pre><!-- HTML_TAG_START -->start = time.time_ns()
<span class="hljs-keyword">for</span> _ <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(NUM_ITERS_TO_RUN):
    images = distilled(
        prompt,
        num_inference_steps=NUM_INFERENCE_STEPS,
        generator=generator,
        num_images_per_prompt=NUM_IMAGES_PER_PROMPT
    ).images
end = time.time_ns()

distilled_sd = <span class="hljs-string">f&quot;<span class="hljs-subst">{(end - start) / <span class="hljs-number">1e6</span>:<span class="hljs-number">.1</span>f}</span>&quot;</span>
<span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Execution time -- <span class="hljs-subst">{distilled_sd}</span> ms\n&quot;</span>)
<span class="hljs-string">&quot;Execution time -- 29884.2 ms&quot;</span><!-- HTML_TAG_END --></pre></div>
<div class="flex gap-4"><div><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/original_sd.png">
    <figcaption class="mt-2 text-center text-sm text-gray-500">original Stable Diffusion (45781.5 ms)</figcaption></div>
  <div><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/distilled_sd.png">
    <figcaption class="mt-2 text-center text-sm text-gray-500">distilled Stable Diffusion (29884.2 ms)</figcaption></div></div>
<h2 class="relative group"><a id="tiny-autoencoder" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#tiny-autoencoder"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a>
	<span>Tiny AutoEncoder
	</span></h2>

<p>To speed inference up even more, use a tiny distilled version of the <a href="https://huggingface.co/sayakpaul/taesdxl-diffusers" rel="nofollow">Stable Diffusion VAE</a> to denoise the latents into images. Replace the VAE in the distilled Stable Diffusion model with the tiny VAE:</p>

	<div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg>
	<div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div>
	Copied</div></button></div>
	<pre><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> AutoencoderTiny

distilled.vae = AutoencoderTiny.from_pretrained(
    <span class="hljs-string">&quot;sayakpaul/taesd-diffusers&quot;</span>, torch_dtype=torch.float16, use_safetensors=<span class="hljs-literal">True</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)<!-- HTML_TAG_END --></pre></div>
<p>Time the distilled model and distilled VAE inference:</p>

	<div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg>
	<div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div>
	Copied</div></button></div>
	<pre><!-- HTML_TAG_START -->start = time.time_ns()
<span class="hljs-keyword">for</span> _ <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(NUM_ITERS_TO_RUN):
    images = distilled(
        prompt,
        num_inference_steps=NUM_INFERENCE_STEPS,
        generator=generator,
        num_images_per_prompt=NUM_IMAGES_PER_PROMPT
    ).images
end = time.time_ns()

distilled_tiny_sd = <span class="hljs-string">f&quot;<span class="hljs-subst">{(end - start) / <span class="hljs-number">1e6</span>:<span class="hljs-number">.1</span>f}</span>&quot;</span>
<span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Execution time -- <span class="hljs-subst">{distilled_tiny_sd}</span> ms\n&quot;</span>)
<span class="hljs-string">&quot;Execution time -- 27165.7 ms&quot;</span><!-- HTML_TAG_END --></pre></div>
<div class="flex justify-center"><div><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/distilled_sd_vae.png">
    <figcaption class="mt-2 text-center text-sm text-gray-500">distilled Stable Diffusion + Tiny AutoEncoder (27165.7 ms)</figcaption></div></div>


		<script type="module" data-hydrate="ijx7dr">
		import { start } from "/docs/diffusers/v0.20.0/en/_app/start-hf-doc-builder.js";
		start({
			target: document.querySelector('[data-hydrate="ijx7dr"]').parentNode,
			paths: {"base":"/docs/diffusers/v0.20.0/en","assets":"/docs/diffusers/v0.20.0/en"},
			session: {},
			route: false,
			spa: false,
			trailing_slash: "never",
			hydrate: {
				status: 200,
				error: null,
				nodes: [
					import("/docs/diffusers/v0.20.0/en/_app/pages/__layout.svelte-hf-doc-builder.js"),
						import("/docs/diffusers/v0.20.0/en/_app/pages/using-diffusers/distilled_sd.mdx-hf-doc-builder.js")
				],
				params: {}
			}
		});
	</script>