Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Load community pipelines and components","local":"load-community-pipelines-and-components","sections":[{"title":"Community pipelines","local":"community-pipelines","sections":[],"depth":2},{"title":"Community components","local":"community-components","sections":[],"depth":2}],"depth":1}"> | |
| <link href="/docs/diffusers/v0.26.2/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/entry/start.ca2a3496.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/scheduler.182ea377.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/singletons.8615e179.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/index.1f6d62f6.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/paths.b0157720.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/entry/app.76ddbcd8.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/index.abf12888.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/nodes/0.1ba30a4e.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/each.e59479a4.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/nodes/164.be8253f0.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/Tip.230e2334.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/CodeBlock.57fe6e13.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/DocNotebookDropdown.5fa27ace.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/globals.7f7f1b26.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/v0.26.2/en/_app/immutable/chunks/Heading.16916d63.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Load community pipelines and components","local":"load-community-pipelines-and-components","sections":[{"title":"Community pipelines","local":"community-pipelines","sections":[],"depth":2},{"title":"Community components","local":"community-components","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <h1 class="relative group"><a id="load-community-pipelines-and-components" 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="#load-community-pipelines-and-components"><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>Load community pipelines and components</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> <h2 class="relative group"><a id="community-pipelines" 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="#community-pipelines"><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>Community pipelines</span></h2> <p data-svelte-h="svelte-1poa8jr">Community pipelines are any <a href="/docs/diffusers/v0.26.2/en/api/pipelines/overview#diffusers.DiffusionPipeline">DiffusionPipeline</a> class that are different from the original implementation as specified in their paper (for example, the <a href="/docs/diffusers/v0.26.2/en/api/pipelines/controlnet#diffusers.StableDiffusionControlNetPipeline">StableDiffusionControlNetPipeline</a> corresponds to the <a href="https://arxiv.org/abs/2302.05543" rel="nofollow">Text-to-Image Generation with ControlNet Conditioning</a> paper). They provide additional functionality or extend the original implementation of a pipeline.</p> <p data-svelte-h="svelte-10en0rk">There are many cool community pipelines like <a href="https://github.com/huggingface/diffusers/tree/main/examples/community#speech-to-image" rel="nofollow">Speech to Image</a> or <a href="https://github.com/huggingface/diffusers/tree/main/examples/community#composable-stable-diffusion" rel="nofollow">Composable Stable Diffusion</a>, and you can find all the official community pipelines <a href="https://github.com/huggingface/diffusers/tree/main/examples/community" rel="nofollow">here</a>.</p> <p data-svelte-h="svelte-1mmnk2f">To load any community pipeline on the Hub, pass the repository id of the community pipeline to the <code>custom_pipeline</code> argument and the model repository where you’d like to load the pipeline weights and components from. For example, the example below loads a dummy pipeline from <a href="https://huggingface.co/hf-internal-testing/diffusers-dummy-pipeline/blob/main/pipeline.py" rel="nofollow"><code>hf-internal-testing/diffusers-dummy-pipeline</code></a> and the pipeline weights and components from <a href="https://huggingface.co/google/ddpm-cifar10-32" rel="nofollow"><code>google/ddpm-cifar10-32</code></a>:</p> <div class="course-tip course-tip-orange bg-gradient-to-br dark:bg-gradient-to-r before:border-orange-500 dark:before:border-orange-800 from-orange-50 dark:from-gray-900 to-white dark:to-gray-950 border border-orange-50 text-orange-700 dark:text-gray-400"><p data-svelte-h="svelte-154jbq5">🔒 By loading a community pipeline from the Hugging Face Hub, you are trusting that the code you are loading is safe. Make sure to inspect the code online before loading and running it automatically!</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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline | |
| pipeline = DiffusionPipeline.from_pretrained( | |
| <span class="hljs-string">"google/ddpm-cifar10-32"</span>, custom_pipeline=<span class="hljs-string">"hf-internal-testing/diffusers-dummy-pipeline"</span>, use_safetensors=<span class="hljs-literal">True</span> | |
| )<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1yyre8w">Loading an official community pipeline is similar, but you can mix loading weights from an official repository id and pass pipeline components directly. The example below loads the community <a href="https://github.com/huggingface/diffusers/tree/main/examples/community#clip-guided-stable-diffusion" rel="nofollow">CLIP Guided Stable Diffusion</a> pipeline, and you can pass the CLIP model components directly to 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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline | |
| <span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CLIPImageProcessor, CLIPModel | |
| clip_model_id = <span class="hljs-string">"laion/CLIP-ViT-B-32-laion2B-s34B-b79K"</span> | |
| feature_extractor = CLIPImageProcessor.from_pretrained(clip_model_id) | |
| clip_model = CLIPModel.from_pretrained(clip_model_id) | |
| pipeline = DiffusionPipeline.from_pretrained( | |
| <span class="hljs-string">"runwayml/stable-diffusion-v1-5"</span>, | |
| custom_pipeline=<span class="hljs-string">"clip_guided_stable_diffusion"</span>, | |
| clip_model=clip_model, | |
| feature_extractor=feature_extractor, | |
| use_safetensors=<span class="hljs-literal">True</span>, | |
| )<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1ho2g1z">For more information about community pipelines, take a look at the <a href="custom_pipeline_examples">Community pipelines</a> guide for how to use them and if you’re interested in adding a community pipeline check out the <a href="contribute_pipeline">How to contribute a community pipeline</a> guide!</p> <h2 class="relative group"><a id="community-components" 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="#community-components"><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>Community components</span></h2> <p data-svelte-h="svelte-1vh7uou">Community components allow users to build pipelines that may have customized components that are not a part of Diffusers. If your pipeline has custom components that Diffusers doesn’t already support, you need to provide their implementations as Python modules. These customized components could be a VAE, UNet, and scheduler. In most cases, the text encoder is imported from the Transformers library. The pipeline code itself can also be customized.</p> <p data-svelte-h="svelte-phpvnj">This section shows how users should use community components to build a community pipeline.</p> <p data-svelte-h="svelte-7sknw0">You’ll use the <a href="https://huggingface.co/showlab/show-1-base" rel="nofollow">showlab/show-1-base</a> pipeline checkpoint as an example. So, let’s start loading the components:</p> <ol data-svelte-h="svelte-8ccbl9"><li>Import and load the text encoder from Transformers:</li></ol> <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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> T5Tokenizer, T5EncoderModel | |
| pipe_id = <span class="hljs-string">"showlab/show-1-base"</span> | |
| tokenizer = T5Tokenizer.from_pretrained(pipe_id, subfolder=<span class="hljs-string">"tokenizer"</span>) | |
| text_encoder = T5EncoderModel.from_pretrained(pipe_id, subfolder=<span class="hljs-string">"text_encoder"</span>)<!-- HTML_TAG_END --></pre></div> <ol start="2" data-svelte-h="svelte-1rtkhjg"><li>Load a scheduler:</li></ol> <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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DPMSolverMultistepScheduler | |
| scheduler = DPMSolverMultistepScheduler.from_pretrained(pipe_id, subfolder=<span class="hljs-string">"scheduler"</span>)<!-- HTML_TAG_END --></pre></div> <ol start="3" data-svelte-h="svelte-o9t4x9"><li>Load an image processor:</li></ol> <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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CLIPFeatureExtractor | |
| feature_extractor = CLIPFeatureExtractor.from_pretrained(pipe_id, subfolder=<span class="hljs-string">"feature_extractor"</span>)<!-- HTML_TAG_END --></pre></div> <div class="course-tip course-tip-orange bg-gradient-to-br dark:bg-gradient-to-r before:border-orange-500 dark:before:border-orange-800 from-orange-50 dark:from-gray-900 to-white dark:to-gray-950 border border-orange-50 text-orange-700 dark:text-gray-400"><p data-svelte-h="svelte-ptu9pr">In steps 4 and 5, the custom <a href="https://github.com/showlab/Show-1/blob/main/showone/models/unet_3d_condition.py" rel="nofollow">UNet</a> and <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/unet/showone_unet_3d_condition.py" rel="nofollow">pipeline</a> implementation must match the format shown in their files for this example to work.</p></div> <ol start="4" data-svelte-h="svelte-3jez55"><li>Now you’ll load a <a href="https://github.com/showlab/Show-1/blob/main/showone/models/unet_3d_condition.py" rel="nofollow">custom UNet</a>, which in this example, has already been implemented in the <code>showone_unet_3d_condition.py</code> <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/unet/showone_unet_3d_condition.py" rel="nofollow">script</a> for your convenience. You’ll notice the <code>UNet3DConditionModel</code> class name is changed to <code>ShowOneUNet3DConditionModel</code> because <a href="/docs/diffusers/v0.26.2/en/api/models/unet3d-cond#diffusers.UNet3DConditionModel">UNet3DConditionModel</a> already exists in Diffusers. Any components needed for the <code>ShowOneUNet3DConditionModel</code> class should be placed in the <code>showone_unet_3d_condition.py</code> script.</li></ol> <p data-svelte-h="svelte-4yol19">Once this is done, you can initialize the UNet:</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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> showone_unet_3d_condition <span class="hljs-keyword">import</span> ShowOneUNet3DConditionModel | |
| unet = ShowOneUNet3DConditionModel.from_pretrained(pipe_id, subfolder=<span class="hljs-string">"unet"</span>)<!-- HTML_TAG_END --></pre></div> <ol start="5" data-svelte-h="svelte-1m9qhhm"><li>Finally, you’ll load the custom pipeline code. For this example, it has already been created for you in the <code>pipeline_t2v_base_pixel.py</code> <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/pipeline_t2v_base_pixel.py" rel="nofollow">script</a>. This script contains a custom <code>TextToVideoIFPipeline</code> class for generating videos from text. Just like the custom UNet, any code needed for the custom pipeline to work should go in the <code>pipeline_t2v_base_pixel.py</code> script.</li></ol> <p data-svelte-h="svelte-1jpprmd">Once everything is in place, you can initialize the <code>TextToVideoIFPipeline</code> with the <code>ShowOneUNet3DConditionModel</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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> pipeline_t2v_base_pixel <span class="hljs-keyword">import</span> TextToVideoIFPipeline | |
| <span class="hljs-keyword">import</span> torch | |
| pipeline = TextToVideoIFPipeline( | |
| unet=unet, | |
| text_encoder=text_encoder, | |
| tokenizer=tokenizer, | |
| scheduler=scheduler, | |
| feature_extractor=feature_extractor | |
| ) | |
| pipeline = pipeline.to(device=<span class="hljs-string">"cuda"</span>) | |
| pipeline.torch_dtype = torch.float16<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-14lz5yb">Push the pipeline to the Hub to share with the community!</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 class=""><!-- HTML_TAG_START -->pipeline.push_to_hub(<span class="hljs-string">"custom-t2v-pipeline"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-azdq2p">After the pipeline is successfully pushed, you need a couple of changes:</p> <ol data-svelte-h="svelte-7wsjmh"><li>Change the <code>_class_name</code> attribute in <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/model_index.json#L2" rel="nofollow"><code>model_index.json</code></a> to <code>"pipeline_t2v_base_pixel"</code> and <code>"TextToVideoIFPipeline"</code>.</li> <li>Upload <code>showone_unet_3d_condition.py</code> to the <code>unet</code> <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/unet/showone_unet_3d_condition.py" rel="nofollow">directory</a>.</li> <li>Upload <code>pipeline_t2v_base_pixel.py</code> to the pipeline base <a href="https://huggingface.co/sayakpaul/show-1-base-with-code/blob/main/unet/showone_unet_3d_condition.py" rel="nofollow">directory</a>.</li></ol> <p data-svelte-h="svelte-xpz58c">To run inference, simply add the <code>trust_remote_code</code> argument while initializing the pipeline to handle all the “magic” behind the scenes.</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 class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline | |
| <span class="hljs-keyword">import</span> torch | |
| pipeline = DiffusionPipeline.from_pretrained( | |
| <span class="hljs-string">"<change-username>/<change-id>"</span>, trust_remote_code=<span class="hljs-literal">True</span>, torch_dtype=torch.float16 | |
| ).to(<span class="hljs-string">"cuda"</span>) | |
| prompt = <span class="hljs-string">"hello"</span> | |
| <span class="hljs-comment"># Text embeds</span> | |
| prompt_embeds, negative_embeds = pipeline.encode_prompt(prompt) | |
| <span class="hljs-comment"># Keyframes generation (8x64x40, 2fps)</span> | |
| video_frames = pipeline( | |
| prompt_embeds=prompt_embeds, | |
| negative_prompt_embeds=negative_embeds, | |
| num_frames=<span class="hljs-number">8</span>, | |
| height=<span class="hljs-number">40</span>, | |
| width=<span class="hljs-number">64</span>, | |
| num_inference_steps=<span class="hljs-number">2</span>, | |
| guidance_scale=<span class="hljs-number">9.0</span>, | |
| output_type=<span class="hljs-string">"pt"</span> | |
| ).frames<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1cmjoh5">As an additional reference example, you can refer to the repository structure of <a href="https://huggingface.co/stabilityai/japanese-stable-diffusion-xl/" rel="nofollow">stabilityai/japanese-stable-diffusion-xl</a>, that makes use of the <code>trust_remote_code</code> feature:</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 class=""><!-- HTML_TAG_START --> | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline | |
| <span class="hljs-keyword">import</span> torch | |
| pipeline = DiffusionPipeline.from_pretrained( | |
| <span class="hljs-string">"stabilityai/japanese-stable-diffusion-xl"</span>, trust_remote_code=<span class="hljs-literal">True</span> | |
| ) | |
| pipeline.to(<span class="hljs-string">"cuda"</span>) | |
| <span class="hljs-comment"># if using torch < 2.0</span> | |
| <span class="hljs-comment"># pipeline.enable_xformers_memory_efficient_attention()</span> | |
| prompt = <span class="hljs-string">"柴犬、カラフルアート"</span> | |
| image = pipeline(prompt=prompt).images[<span class="hljs-number">0</span>] | |
| <!-- HTML_TAG_END --></pre></div> <p></p> | |
| <script> | |
| { | |
| __sveltekit_j7lbip = { | |
| assets: "/docs/diffusers/v0.26.2/en", | |
| base: "/docs/diffusers/v0.26.2/en", | |
| env: {} | |
| }; | |
| const element = document.currentScript.parentElement; | |
| const data = [null,null]; | |
| Promise.all([ | |
| import("/docs/diffusers/v0.26.2/en/_app/immutable/entry/start.ca2a3496.js"), | |
| import("/docs/diffusers/v0.26.2/en/_app/immutable/entry/app.76ddbcd8.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 164], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 32.5 kB
- Xet hash:
- 973bb785c285ba34dd3ab2e8d512ab4d90d6dff4908db81a2a5efb14f24fa42f
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.