Hugging Face's logo

Buckets:
hf-doc-build
/
doc

Files
xet
hf-doc-build/doc / diffusers /v0.25.0 /ko /using-diffusers /weighted_prompts.html
rtrm's picture
rtrm
download
raw
48.5 kB
<!-- META HERE --><meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;local&quot;:&quot;prompt-weighting&quot;,&quot;sections&quot;:[{&quot;local&quot;:&quot;weighting&quot;,&quot;title&quot;:&quot;Weighting&quot;},{&quot;local&quot;:&quot;blending&quot;,&quot;title&quot;:&quot;Blending&quot;},{&quot;local&quot;:&quot;conjunction&quot;,&quot;title&quot;:&quot;Conjunction&quot;},{&quot;local&quot;:&quot;textual-inversion&quot;,&quot;title&quot;:&quot;Textual inversion&quot;},{&quot;local&quot;:&quot;dreambooth&quot;,&quot;title&quot;:&quot;DreamBooth&quot;},{&quot;local&quot;:&quot;stable-diffusion-xl&quot;,&quot;title&quot;:&quot;Stable Diffusion XL&quot;}],&quot;title&quot;:&quot;Prompt weighting&quot;}">
 <link href="/docs/diffusers/v0.25.0/ko/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/entry/start.739fcb44.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/scheduler.182ea377.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/singletons.5abecdb2.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/index.1f6d62f6.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/paths.497209e7.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/entry/app.a301e47d.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/index.008d68e4.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/nodes/0.a130e184.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/each.e59479a4.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/nodes/186.174ba819.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/Tip.4f096367.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/IconCopyLink.96bbb92b.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/CodeBlock.5ed6eb7b.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/DocNotebookDropdown.bb388256.js">
 <link rel="modulepreload" href="/docs/diffusers/v0.25.0/ko/_app/immutable/chunks/globals.7f7f1b26.js"><!-- HEAD_svelte-1phssyn_START --><!-- HEAD_svelte-1phssyn_END --> <h1 class="relative group"><a id="prompt-weighting" 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="#prompt-weighting"><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 data-svelte-h="svelte-ni1uk3">Prompt weighting</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 data-svelte-h="svelte-1tre73w">Prompt weighting provides a way to emphasize or de-emphasize certain parts of a prompt, allowing for more control over the generated image. A prompt can include several concepts, which gets turned into contextualized text embeddings. The embeddings are used by the model to condition its cross-attention layers to generate an image (read the Stable Diffusion <a href="https://huggingface.co/blog/stable_diffusion" rel="nofollow">blog post</a> to learn more about how it works).</p> <p data-svelte-h="svelte-1browmn">Prompt weighting works by increasing or decreasing the scale of the text embedding vector that corresponds to its concept in the prompt because you may not necessarily want the model to focus on all concepts equally. The easiest way to prepare the prompt-weighted embeddings is to use <a href="https://github.com/damian0815/compel" rel="nofollow">Compel</a>, a text prompt-weighting and blending library. Once you have the prompt-weighted embeddings, you can pass them to any pipeline that has a <a href="https://huggingface.co/docs/diffusers/en/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline.__call__.prompt_embeds" rel="nofollow"><code>prompt_embeds</code></a> (and optionally <a href="https://huggingface.co/docs/diffusers/en/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline.__call__.negative_prompt_embeds" rel="nofollow"><code>negative_prompt_embeds</code></a>) parameter, such as <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline">StableDiffusionPipeline</a>, <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/controlnet#diffusers.StableDiffusionControlNetPipeline">StableDiffusionControlNetPipeline</a>, and <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLPipeline">StableDiffusionXLPipeline</a>.</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 data-svelte-h="svelte-nya19q">If your favorite pipeline doesn’t have a <code>prompt_embeds</code> parameter, please open an <a href="https://github.com/huggingface/diffusers/issues/new/choose" rel="nofollow">issue</a> so we can add it!</p></div> <p data-svelte-h="svelte-xw5jy8">This guide will show you how to weight and blend your prompts with Compel in 🤗 Diffusers.</p> <p data-svelte-h="svelte-1tbon0u">Before you begin, make sure you have the latest version of Compel installed:</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-comment"># uncomment to install in Colab</span>
<span class="hljs-comment">#!pip install compel --upgrade</span><!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-138cbmq">For this guide, let’s generate an image with the prompt <code>&quot;a red cat playing with a ball&quot;</code> using the <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline">StableDiffusionPipeline</a>:</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, UniPCMultistepScheduler
<span class="hljs-keyword">import</span> torch
pipe = StableDiffusionPipeline.from_pretrained(<span class="hljs-string">&quot;CompVis/stable-diffusion-v1-4&quot;</span>, use_safetensors=<span class="hljs-literal">True</span>)
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
pipe.to(<span class="hljs-string">&quot;cuda&quot;</span>)
prompt = <span class="hljs-string">&quot;a red cat playing with a ball&quot;</span>
generator = torch.Generator(device=<span class="hljs-string">&quot;cpu&quot;</span>).manual_seed(<span class="hljs-number">33</span>)
image = pipe(prompt, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-1brza9b"><img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/compel/forest_0.png"></div> <h2 class="relative group"><a id="weighting" 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="#weighting"><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 data-svelte-h="svelte-1w0yhgb">Weighting</span></h2> <p data-svelte-h="svelte-137cdqc">You’ll notice there is no “ball” in the image! Let’s use compel to upweight the concept of “ball” in the prompt. Create a <a href="https://github.com/damian0815/compel/blob/main/doc/compel.md#compel-objects" rel="nofollow"><code>Compel</code></a> object, and pass it a tokenizer and text encoder:</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> compel <span class="hljs-keyword">import</span> Compel
compel_proc = Compel(tokenizer=pipe.tokenizer, text_encoder=pipe.text_encoder)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-18ay2r3">compel uses <code>+</code> or <code>-</code> to increase or decrease the weight of a word in the prompt. To increase the weight of “ball”:</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 data-svelte-h="svelte-1dn4vb7"><code>+</code> corresponds to the value <code>1.1</code>, <code>++</code> corresponds to <code>1.1^2</code>, and so on. Similarly, <code>-</code> corresponds to <code>0.9</code> and <code>--</code> corresponds to <code>0.9^2</code>. Feel free to experiment with adding more <code>+</code> or <code>-</code> in your prompt!</p></div> <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 -->prompt = <span class="hljs-string">&quot;a red cat playing with a ball++&quot;</span><!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-vulem2">Pass the prompt to <code>compel_proc</code> to create the new prompt embeddings which are passed to the pipeline:</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 -->prompt_embeds = compel_proc(prompt)
generator = torch.manual_seed(<span class="hljs-number">33</span>)
image = pipe(prompt_embeds=prompt_embeds, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-17zkpq4"><img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/compel/forest_1.png"></div> <p data-svelte-h="svelte-gxkdpi">To downweight parts of the prompt, use the <code>-</code> suffix:</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 -->prompt = <span class="hljs-string">&quot;a red------- cat playing with a ball&quot;</span>
prompt_embeds = compel_proc(prompt)
generator = torch.manual_seed(<span class="hljs-number">33</span>)
image = pipe(prompt_embeds=prompt_embeds, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-1d7lfen"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-neg.png"></div> <p data-svelte-h="svelte-1f6exw0">You can even up or downweight multiple concepts in the same prompt:</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 -->prompt = <span class="hljs-string">&quot;a red cat++ playing with a ball----&quot;</span>
prompt_embeds = compel_proc(prompt)
generator = torch.manual_seed(<span class="hljs-number">33</span>)
image = pipe(prompt_embeds=prompt_embeds, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-1qezu4q"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-pos-neg.png"></div> <h2 class="relative group"><a id="blending" 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="#blending"><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 data-svelte-h="svelte-10e85ai">Blending</span></h2> <p data-svelte-h="svelte-1oxm60">You can also create a weighted <em>blend</em> of prompts by adding <code>.blend()</code> to a list of prompts and passing it some weights. Your blend may not always produce the result you expect because it breaks some assumptions about how the text encoder functions, so just have fun and experiment with it!</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 -->prompt_embeds = compel_proc(<span class="hljs-string">&#x27;(&quot;a red cat playing with a ball&quot;, &quot;jungle&quot;).blend(0.7, 0.8)&#x27;</span>)
generator = torch.Generator(device=<span class="hljs-string">&quot;cuda&quot;</span>).manual_seed(<span class="hljs-number">33</span>)
image = pipe(prompt_embeds=prompt_embeds, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-8v41o8"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-blend.png"></div> <h2 class="relative group"><a id="conjunction" 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="#conjunction"><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 data-svelte-h="svelte-1yjkhmz">Conjunction</span></h2> <p data-svelte-h="svelte-tjkp7t">A conjunction diffuses each prompt independently and concatenates their results by their weighted sum. Add <code>.and()</code> to the end of a list of prompts to create a conjunction:</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 -->prompt_embeds = compel_proc(<span class="hljs-string">&#x27;[&quot;a red cat&quot;, &quot;playing with a&quot;, &quot;ball&quot;].and()&#x27;</span>)
generator = torch.Generator(device=<span class="hljs-string">&quot;cuda&quot;</span>).manual_seed(<span class="hljs-number">55</span>)
image = pipe(prompt_embeds=prompt_embeds, generator=generator, num_inference_steps=<span class="hljs-number">20</span>).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-1ycuwub"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-conj.png"></div> <h2 class="relative group"><a id="textual-inversion" 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="#textual-inversion"><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 data-svelte-h="svelte-ucw1ox">Textual inversion</span></h2> <p data-svelte-h="svelte-lcas44"><a href="../training/text_inversion">Textual inversion</a> is a technique for learning a specific concept from some images which you can use to generate new images conditioned on that concept.</p> <p data-svelte-h="svelte-5hel8g">Create a pipeline and use the <a href="/docs/diffusers/v0.25.0/ko/api/loaders/textual_inversion#diffusers.loaders.TextualInversionLoaderMixin.load_textual_inversion">load_textual_inversion()</a> function to load the textual inversion embeddings (feel free to browse the <a href="https://huggingface.co/spaces/sd-concepts-library/stable-diffusion-conceptualizer" rel="nofollow">Stable Diffusion Conceptualizer</a> for 100+ trained concepts):</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> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> StableDiffusionPipeline
<span class="hljs-keyword">from</span> compel <span class="hljs-keyword">import</span> Compel, DiffusersTextualInversionManager
pipe = StableDiffusionPipeline.from_pretrained(
 <span class="hljs-string">&quot;runwayml/stable-diffusion-v1-5&quot;</span>, torch_dtype=torch.float16,
 use_safetensors=<span class="hljs-literal">True</span>, variant=<span class="hljs-string">&quot;fp16&quot;</span>).to(<span class="hljs-string">&quot;cuda&quot;</span>)
pipe.load_textual_inversion(<span class="hljs-string">&quot;sd-concepts-library/midjourney-style&quot;</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-h0g1y9">Compel provides a <code>DiffusersTextualInversionManager</code> class to simplify prompt weighting with textual inversion. Instantiate <code>DiffusersTextualInversionManager</code> and pass it to the <code>Compel</code> class:</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 -->textual_inversion_manager = DiffusersTextualInversionManager(pipe)
compel_proc = Compel(
 tokenizer=pipe.tokenizer,
 text_encoder=pipe.text_encoder,
 textual_inversion_manager=textual_inversion_manager)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1fkpura">Incorporate the concept to condition a prompt with using the <code>&lt;concept&gt;</code> syntax:</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 -->prompt_embeds = compel_proc(<span class="hljs-string">&#x27;(&quot;A red cat++ playing with a ball &lt;midjourney-style&gt;&quot;)&#x27;</span>)
image = pipe(prompt_embeds=prompt_embeds).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-1awripq"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-text-inversion.png"></div> <h2 class="relative group"><a id="dreambooth" 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="#dreambooth"><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 data-svelte-h="svelte-13nggkq">DreamBooth</span></h2> <p data-svelte-h="svelte-1mms94e"><a href="../training/dreambooth">DreamBooth</a> is a technique for generating contextualized images of a subject given just a few images of the subject to train on. It is similar to textual inversion, but DreamBooth trains the full model whereas textual inversion only fine-tunes the text embeddings. This means you should use <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/overview#diffusers.DiffusionPipeline.from_pretrained">from_pretrained()</a> to load the DreamBooth model (feel free to browse the <a href="https://huggingface.co/sd-dreambooth-library" rel="nofollow">Stable Diffusion Dreambooth Concepts Library</a> for 100+ trained models):</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> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline, UniPCMultistepScheduler
<span class="hljs-keyword">from</span> compel <span class="hljs-keyword">import</span> Compel
pipe = DiffusionPipeline.from_pretrained(<span class="hljs-string">&quot;sd-dreambooth-library/dndcoverart-v1&quot;</span>, torch_dtype=torch.float16).to(<span class="hljs-string">&quot;cuda&quot;</span>)
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1le33cx">Create a <code>Compel</code> class with a tokenizer and text encoder, and pass your prompt to it. Depending on the model you use, you’ll need to incorporate the model’s unique identifier into your prompt. For example, the <code>dndcoverart-v1</code> model uses the identifier <code>dndcoverart</code>:</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 -->compel_proc = Compel(tokenizer=pipe.tokenizer, text_encoder=pipe.text_encoder)
prompt_embeds = compel_proc(<span class="hljs-string">&#x27;(&quot;magazine cover of a dndcoverart dragon, high quality, intricate details, larry elmore art style&quot;).and()&#x27;</span>)
image = pipe(prompt_embeds=prompt_embeds).images[<span class="hljs-number">0</span>]
image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-zvft88"><img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/compel-dreambooth.png"></div> <h2 class="relative group"><a id="stable-diffusion-xl" 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="#stable-diffusion-xl"><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 data-svelte-h="svelte-1g3g87b">Stable Diffusion XL</span></h2> <p data-svelte-h="svelte-xo93eu">Stable Diffusion XL (SDXL) has two tokenizers and text encoders so it’s usage is a bit different. To address this, you should pass both tokenizers and encoders to the <code>Compel</code> class:</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> compel <span class="hljs-keyword">import</span> Compel, ReturnedEmbeddingsType
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> make_image_grid
<span class="hljs-keyword">import</span> torch
pipeline = DiffusionPipeline.from_pretrained(
 <span class="hljs-string">&quot;stabilityai/stable-diffusion-xl-base-1.0&quot;</span>,
 variant=<span class="hljs-string">&quot;fp16&quot;</span>,
 use_safetensors=<span class="hljs-literal">True</span>,
 torch_dtype=torch.float16
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
compel = Compel(
 tokenizer=[pipeline.tokenizer, pipeline.tokenizer_2] ,
 text_encoder=[pipeline.text_encoder, pipeline.text_encoder_2],
 returned_embeddings_type=ReturnedEmbeddingsType.PENULTIMATE_HIDDEN_STATES_NON_NORMALIZED,
 requires_pooled=[<span class="hljs-literal">False</span>, <span class="hljs-literal">True</span>]
)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-fhwvqd">This time, let’s upweight “ball” by a factor of 1.5 for the first prompt, and downweight “ball” by 0.6 for the second prompt. The <a href="/docs/diffusers/v0.25.0/ko/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLPipeline">StableDiffusionXLPipeline</a> also requires <a href="https://huggingface.co/docs/diffusers/en/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLInpaintPipeline.__call__.pooled_prompt_embeds" rel="nofollow"><code>pooled_prompt_embeds</code></a> (and optionally <a href="https://huggingface.co/docs/diffusers/en/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLInpaintPipeline.__call__.negative_pooled_prompt_embeds" rel="nofollow"><code>negative_pooled_prompt_embeds</code></a>) so you should pass those to the pipeline along with the conditioning tensors:</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-comment"># apply weights</span>
prompt = [<span class="hljs-string">&quot;a red cat playing with a (ball)1.5&quot;</span>, <span class="hljs-string">&quot;a red cat playing with a (ball)0.6&quot;</span>]
conditioning, pooled = compel(prompt)
<span class="hljs-comment"># generate image</span>
generator = [torch.Generator().manual_seed(<span class="hljs-number">33</span>) <span class="hljs-keyword">for</span> _ <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(<span class="hljs-built_in">len</span>(prompt))]
images = pipeline(prompt_embeds=conditioning, pooled_prompt_embeds=pooled, generator=generator, num_inference_steps=<span class="hljs-number">30</span>).images
make_image_grid(images, rows=<span class="hljs-number">1</span>, cols=<span class="hljs-number">2</span>)<!-- HTML_TAG_END --></pre></div> <div class="flex gap-4" data-svelte-h="svelte-idvlgw"><div><img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/compel/sdxl_ball1.png"> <figcaption class="mt-2 text-center text-sm text-gray-500">&quot;a red cat playing with a (ball)1.5&quot;</figcaption></div> <div><img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/compel/sdxl_ball2.png"> <figcaption class="mt-2 text-center text-sm text-gray-500">&quot;a red cat playing with a (ball)0.6&quot;</figcaption></div></div>
<script>
 {
 __sveltekit_160bpzs = {
 assets: "/docs/diffusers/v0.25.0/ko",
 base: "/docs/diffusers/v0.25.0/ko",
 env: {}
 };
 const element = document.currentScript.parentElement;
 const data = [null,null];
 Promise.all([
 import("/docs/diffusers/v0.25.0/ko/_app/immutable/entry/start.739fcb44.js"),
 import("/docs/diffusers/v0.25.0/ko/_app/immutable/entry/app.a301e47d.js")
 ]).then(([kit, app]) => {
 kit.start(app, element, {
 node_ids: [0, 186],
 data,
 form: null,
 error: null
 });
 });
 }
 </script>

Xet Storage Details

Size:
48.5 kB
·
Xet hash:
a5b907988c7bfd6d92a940e879484710cf46dff09d67fd10c2436ad61d4d4edf

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.