Buckets:

hf-doc-build
/

doc

Files
xet
hf-doc-build/doc / diffusers /v0.43.0 /ja /modular_diffusers /quickstart.html
mishig
2 days ago
download
raw
57.8 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;クイックスタート&quot;,&quot;local&quot;:&quot;quickstart&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;パイプラインの実行&quot;,&quot;local&quot;:&quot;run-a-pipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;構造を理解する&quot;,&quot;local&quot;:&quot;understand-the-structure&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;ワークフロー&quot;,&quot;local&quot;:&quot;workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;サブブロック&quot;,&quot;local&quot;:&quot;sub-blocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;新しいワークフローを構築&quot;,&quot;local&quot;:&quot;compose-new-workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;次のステップ&quot;,&quot;local&quot;:&quot;next-steps&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/>
		<link href="/docs/diffusers/main/en/_app/immutable/entry/start.0S0_iEoG.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/I9eooN77.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/Bupkjr9f.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/entry/app.CPdOxCbw.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/CyvZ3FS5.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/l0QnnOED.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DnVIJfC5.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DsnmJJEf.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/0.DteTvmQZ.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/CoSrdQax.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/328.g2gShkQF.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DSPDl1gz.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/D5eijdnr.js" rel="modulepreload">
		<!--1x94atl--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Quickstart&quot;,&quot;local&quot;:&quot;quickstart&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Run a pipeline&quot;,&quot;local&quot;:&quot;run-a-pipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Understand the structure&quot;,&quot;local&quot;:&quot;understand-the-structure&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Workflows&quot;,&quot;local&quot;:&quot;workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Sub-blocks&quot;,&quot;local&quot;:&quot;sub-blocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Compose new workflows&quot;,&quot;local&quot;:&quot;compose-new-workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Next steps&quot;,&quot;local&quot;:&quot;next-steps&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/><!---->
		<link href="/docs/diffusers/main/en/_app/immutable/assets/0.tn0RQdqM.css" rel="modulepreload"> <!--[--><!--[0--><!--[--><!--[0--><!--[--><p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: inline-end; margin-inline-start: 10px; display: inline-flex; position: relative; z-index: 10;"><div dir="ltr" class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" 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><!----></span> <span dir="auto">Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg><!----></button></div> <!--[-1--><!--]--></div><!----> <!--[0--><h1 class="relative group"><a id="quickstart" 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="#quickstart"><span><svg 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><!---->クイックスタート<!----></span></h1><!--]--><!----> <p>Modular Diffusersは、柔軟でカスタマイズ可能なパイプラインを迅速に構築するためのフレームワークです。これらのパイプラインは、標準的な<code>DiffusionPipeline</code>の機能を超えた処理を行うことができます。Modular Diffusersの中核をなすのは<a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline_blocks#diffusers.ModularPipelineBlocks">ModularPipelineBlocks</a>であり、他のブロックと組み合わせることで新しいワークフローに対応できます。これらのブロックは、生成タスクを実行するためのユーザーフレンドリーなインターフェースである<a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline">ModularPipeline</a>に変換されます。</p> <p>このガイドでは、モジュラーパイプラインの実行方法、その構造の理解、そしてそれを構成するブロックを変更することによるカスタマイズ方法について説明します。<a href="/docs/diffusers/v0.43.0/ja/api/pipelines/overview#diffusers.DiffusionPipeline">DiffusionPipeline</a>に精通している場合は、動作の違いについて理解するために、まず<a href="./coming_from_diffusion_pipeline">DiffusionPipelineからの移行</a>をお読みください。</p> <!--[1--><h2 class="relative group"><a id="run-a-pipeline" 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="#run-a-pipeline"><span><svg 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><!---->パイプラインの実行<!----></span></h2><!--]--><!----> <p><a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline">ModularPipeline</a>は、モジュラーパイプラインの読み込み、実行、および管理を行うためのメインインターフェースです。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> ModularPipeline, ComponentsManager

<span class="hljs-comment"># Use ComponentsManager to enable auto CPU offloading for memory efficiency</span>
manager = ComponentsManager()
manager.enable_auto_cpu_offload(device=<span class="hljs-string">&quot;cuda:0&quot;</span>)

pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;Qwen/Qwen-Image&quot;</span>, components_manager=manager)
pipe.load_components(dtype=torch.bfloat16)

image = pipe(
    prompt=<span class="hljs-string">&quot;cat wizard with red hat, gandalf, lord of the rings, detailed, fantasy, cute, adorable, Pixar, Disney&quot;</span>,
).images[<span class="hljs-number">0</span>]
image<!----></pre></div><!----> <p><a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline.from_pretrained">from_pretrained()</a> は遅延読み込みを使用します。設定を読み取って各コンポーネントの読み込み先を把握しますが、<a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline.load_components">load_components()</a> を呼び出すまで実際のモデルの重みは読み込まれません。これにより、コンポーネントの読み込みタイミングと方法を制御できます。</p> <blockquote class="tip"><p><code>ComponentsManager</code>に<code>enable_auto_cpu_offload</code>を指定すると、必要に応じてモデルをCPUとGPUの間で自動的に移動し、Qwen-Imageのような大規模モデルのメモリ使用量を削減します。詳細は<a href="./components_manager">ComponentsManager</a>ガイドを参照してください。</p> <p>オフローディングが不要な場合は、<code>components_manager</code> 引数を削除し、<code>to("cuda")</code> を使用してパイプラインを手動でデバイスに移動してください。</p></blockquote> <p><a href="https://huggingface.co/docs/diffusers/modular_diffusers/modular_pipeline#creating-a-pipeline" rel="nofollow">パイプラインの作成</a>および<a href="https://huggingface.co/docs/diffusers/modular_diffusers/modular_pipeline#loading-components" rel="nofollow">コンポーネントの読み込み</a>のガイドで、パイプラインの作成と読み込みについて詳しく学んでください。</p> <!--[1--><h2 class="relative group"><a id="understand-the-structure" 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="#understand-the-structure"><span><svg 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><!---->構造を理解する<!----></span></h2><!--]--><!----> <p><a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline">ModularPipeline</a> は2つの部分から構成されています。<strong>定義</strong>（ブロック）と、<strong>状態</strong>（読み込まれたコンポーネントと設定）です。</p> <p>パイプラインを出力して、その状態を確認します。コンポーネントとその読み込みステータス、設定が表示されます。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-built_in">print</span>(pipe)<!----></pre></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 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> コピーしました</div><!----></button><!----></div> <pre class=" "><!----><span class="hljs-symbol">QwenImageModularPipeline</span> {
  <span class="hljs-string">&quot;_blocks_class_name&quot;</span>: <span class="hljs-string">&quot;QwenImageAutoBlocks&quot;</span>,
  <span class="hljs-string">&quot;_class_name&quot;</span>: <span class="hljs-string">&quot;QwenImageModularPipeline&quot;</span>,
  <span class="hljs-string">&quot;_diffusers_version&quot;</span>: <span class="hljs-string">&quot;0.37.0.dev0&quot;</span>,
  <span class="hljs-string">&quot;transformer&quot;</span>: [
    <span class="hljs-string">&quot;diffusers&quot;</span>,
    <span class="hljs-string">&quot;QwenImageTransformer2DModel&quot;</span>,
    {
      <span class="hljs-string">&quot;pretrained_model_name_or_path&quot;</span>: <span class="hljs-string">&quot;Qwen/Qwen-Image&quot;</span>,
      <span class="hljs-string">&quot;revision&quot;</span>: null,
      <span class="hljs-string">&quot;subfolder&quot;</span>: <span class="hljs-string">&quot;transformer&quot;</span>,
      <span class="hljs-string">&quot;type_hint&quot;</span>: [
        <span class="hljs-string">&quot;diffusers&quot;</span>,
        <span class="hljs-string">&quot;QwenImageTransformer2DModel&quot;</span>
      ],
      <span class="hljs-string">&quot;variant&quot;</span>: null
    }
  ],
  ...
}<!----></pre></div><!----> <p><code>pipe.blocks</code> を通じて定義にアクセスします。これは、パイプラインのワークフロー、入力、出力、および計算ロジックを定義する <a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline_blocks#diffusers.ModularPipelineBlocks">ModularPipelineBlocks</a> です。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-built_in">print</span>(pipe.blocks)<!----></pre></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 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> コピーしました</div><!----></button><!----></div> <pre class=" "><!---->QwenImageAutoBlocks(
  Class: SequentialPipelineBlocks

  Description: Auto Modular pipeline <span class="hljs-keyword">for</span> <span class="hljs-keyword">text</span>-to-<span class="hljs-keyword">image</span>, <span class="hljs-keyword">image</span>-to-<span class="hljs-keyword">image</span>, inpainting, and controlnet tasks using QwenImage.
      
      Supported workflows:
        - <span class="hljs-string">`text2image`</span>: <span class="hljs-keyword">requires</span> <span class="hljs-string">`prompt`</span>
        - <span class="hljs-string">`image2image`</span>: <span class="hljs-keyword">requires</span> <span class="hljs-string">`prompt`</span>, <span class="hljs-string">`image`</span>
        - <span class="hljs-string">`inpainting`</span>: <span class="hljs-keyword">requires</span> <span class="hljs-string">`prompt`</span>, <span class="hljs-string">`mask_image`</span>, <span class="hljs-string">`image`</span>
        - <span class="hljs-string">`controlnet_text2image`</span>: <span class="hljs-keyword">requires</span> <span class="hljs-string">`prompt`</span>, <span class="hljs-string">`control_image`</span>
        ...

  Components:
      text_encoder (<span class="hljs-string">`Qwen2_5_VLForConditionalGeneration`</span>)
      vae (<span class="hljs-string">`AutoencoderKLQwenImage`</span>)
      transformer (<span class="hljs-string">`QwenImageTransformer2DModel`</span>)
      ...

  Sub-Blocks:
    [<span class="hljs-number">0</span>] text_encoder (QwenImageAutoTextEncoderStep)
    [<span class="hljs-number">1</span>] vae_encoder (QwenImageAutoVaeEncoderStep)
    [<span class="hljs-number">2</span>] controlnet_vae_encoder (QwenImageOptionalControlNetVaeEncoderStep)
    [<span class="hljs-number">3</span>] denoise (QwenImageAutoCoreDenoiseStep)
    [<span class="hljs-number">4</span>] decode (QwenImageAutoDecodeStep)
)<!----></pre></div><!----> <p>出力は以下を返します:</p> <ul><li>サポートされているワークフロー（text2image、image2image、inpainting など）</li> <li>構成されるサブブロック（text_encoder、vae_encoder、denoise、decode）</li></ul> <!--[2--><h3 class="relative group"><a id="workflows" 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="#workflows"><span><svg 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><!---->ワークフロー<!----></span></h3><!--]--><!----> <p>このパイプラインは複数のワークフローをサポートし、提供される入力に基づいて動作を適応させます。例えば、パイプラインに<code>image</code>を渡すと、テキストから画像への生成ではなく、画像から画像へのワークフローが実行されます。内部でどのように動作しているかについては、<a href="https://huggingface.co/docs/diffusers/modular_diffusers/auto_pipeline_blocks" rel="nofollow">AutoPipelineBlocks</a>ガイドを参照してください。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> load_image

input_image = load_image(<span class="hljs-string">&quot;https://github.com/Trgtuan10/Image_storage/blob/main/cute_cat.png?raw=true&quot;</span>)

image = pipe(
    prompt=<span class="hljs-string">&quot;cat wizard with red hat, gandalf, lord of the rings, detailed, fantasy, cute, adorable, Pixar, Disney&quot;</span>,
    image=input_image,
).images[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p><code>get_workflow()</code> を使用して、特定のワークフローのブロックを抽出します。ワークフロー名（例: <code>"image2image"</code>、<code>"inpainting"</code>、<code>"controlnet_text2image"</code>）を渡すことで、そのワークフローに関連するブロックのみを取得できます。これは、特定のワークフローをカスタマイズしたりデバッグしたりする場合に便利です。<code>pipe.blocks.available_workflows</code> を確認すると、利用可能なすべてのワークフローを表示できます。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!---->img2img_blocks = pipe.blocks.get_workflow(<span class="hljs-string">&quot;image2image&quot;</span>)<!----></pre></div><!----> <!--[2--><h3 class="relative group"><a id="sub-blocks" 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="#sub-blocks"><span><svg 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><!---->サブブロック<!----></span></h3><!--]--><!----> <p>ブロックは他のブロックを含むことができます。<code>pipe.blocks</code> はトップレベルのブロック定義（ここでは <code>QwenImageAutoBlocks</code>）を返し、<code>sub_blocks</code> を使用すると、その内部にあるより小さなブロックにアクセスできます。</p> <p><code>QwenImageAutoBlocks</code> は、<code>text_encoder</code>、<code>vae_encoder</code>、<code>controlnet_vae_encoder</code>、<code>denoise</code>、<code>decode</code> で構成されています。</p> <p>これらのサブブロックは順次実行され、データは1つのブロックから次のブロックへ直線的に流れます。各ブロックの<code>intermediate_outputs</code>は、次のブロックの<code>inputs</code>として利用可能になります。<a href="./sequential_pipeline_blocks"><code>SequentialPipelineBlocks</code></a>は、この仕組みで動作します。</p> <p><code>sub_blocks</code> プロパティを介してアクセスできます。<code>doc</code> プロパティは、ブロックの完全なドキュメント、つまり入力、出力、コンポーネントを確認するのに便利です。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!---->vae_encoder_block = pipe.blocks.sub_blocks[<span class="hljs-string">&quot;vae_encoder&quot;</span>]
<span class="hljs-built_in">print</span>(vae_encoder_block.doc)<!----></pre></div><!----> <p>このブロックは、<a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline_blocks#diffusers.ModularPipelineBlocks.init_pipeline">init_pipeline()</a> を使用してパイプラインに変換することで、単独で実行できるようになります。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!---->vae_encoder_pipe = vae_encoder_block.init_pipeline()

<span class="hljs-comment"># Reuse the VAE we already loaded, we can reuse it with update_components() method</span>
vae_encoder_pipe.update_components(vae=pipe.vae)

<span class="hljs-comment"># Run just this block</span>
image_latents = vae_encoder_pipe(image=input_image).image_latents
<span class="hljs-built_in">print</span>(image_latents.shape)<!----></pre></div><!----> <p>元のパイプラインのVAEを再読み込みせずに再利用することで、メモリ使用量を効率的に保ちます。詳細は<a href="https://huggingface.co/docs/diffusers/modular_diffusers/modular_pipeline#loading-components" rel="nofollow">コンポーネントの読み込み</a>ガイドを参照してください。</p> <p>ブロックは合成可能であるため、ブロックの追加、削除、または交換によってパイプラインの定義を変更し、新しいワークフローを作成できます。次のセクションでは、ControlNet パイプラインに canny 辺検出ブロックを追加します。これにより、事前処理済みの canny 辺マップの代わりに通常の画像を渡すことができます。</p> <!--[1--><h2 class="relative group"><a id="compose-new-workflows" 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="#compose-new-workflows"><span><svg 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><!---->新しいワークフローを構築<!----></span></h2><!--]--><!----> <p>ControlNet パイプラインに canny エッジ検出ブロックを追加しましょう。まず、Hub から事前構築済みの canny ブロックを読み込みます（独自のブロックを作成するには、<a href="https://huggingface.co/docs/diffusers/modular_diffusers/custom_blocks" rel="nofollow">カスタムブロックの構築</a>を参照してください）。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">from</span> diffusers.modular_pipelines <span class="hljs-keyword">import</span> ModularPipelineBlocks

<span class="hljs-comment"># Load a canny block from the Hub</span>
canny_block = ModularPipelineBlocks.from_pretrained(
    <span class="hljs-string">&quot;diffusers-internal-dev/canny-filtering&quot;</span>,
    trust_remote_code=<span class="hljs-literal">True</span>,
)

<span class="hljs-built_in">print</span>(canny_block.doc)<!----></pre></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 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> コピーしました</div><!----></button><!----></div> <pre class=" "><!---->class CannyBlock

  Inputs:
      image (`Union[Image, ndarray]`):
          Image <span class="hljs-keyword">to</span> compute canny <span class="hljs-built_in">filter</span> <span class="hljs-keyword">on</span>
      low_threshold (`<span class="hljs-keyword">int</span>`, *optional*, defaults <span class="hljs-keyword">to</span> <span class="hljs-number">50</span>):
          Low threshold <span class="hljs-keyword">for</span> the canny <span class="hljs-built_in">filter</span>.
      high_threshold (`<span class="hljs-keyword">int</span>`, *optional*, defaults <span class="hljs-keyword">to</span> <span class="hljs-number">200</span>):
          High threshold <span class="hljs-keyword">for</span> the canny <span class="hljs-built_in">filter</span>.
      ...

  Outputs:
      control_image (`PIL.Image`):
          Canny <span class="hljs-keyword">map</span> <span class="hljs-keyword">for</span> <span class="hljs-built_in">input</span> image<!----></pre></div><!----> <p><code>get_workflow</code> を使用して、<code>QwenImageAutoBlocks</code> から ControlNet ワークフローを抽出します。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># Get the controlnet workflow that we want to work with</span>
blocks = pipe.blocks.get_workflow(<span class="hljs-string">&quot;controlnet_text2image&quot;</span>)
<span class="hljs-built_in">print</span>(blocks.doc)<!----></pre></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 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> コピーしました</div><!----></button><!----></div> <pre class=" "><!---->class <span class="hljs-symbol">SequentialPipelineBlocks</span>

  <span class="hljs-symbol">Inputs</span>:
      prompt (<span class="hljs-string">`str`</span>):
          <span class="hljs-symbol">The</span> prompt or prompts to guide image generation.
      control_image (<span class="hljs-string">`Image`</span>):
          <span class="hljs-symbol">Control</span> image for <span class="hljs-symbol">ControlNet</span> conditioning.
      ...<!----></pre></div><!----> <p>抽出されたワークフローは<a href="./sequential_pipeline_blocks"><code>SequentialPipelineBlocks</code></a>であり、現在<code>control_image</code>を入力として必要としています。パイプラインが通常の画像を受け入れられるように、canny ブロックを先頭に挿入してください。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># Insert canny at the beginning</span>
blocks.sub_blocks.insert(<span class="hljs-string">&quot;canny&quot;</span>, canny_block, <span class="hljs-number">0</span>)

<span class="hljs-comment"># Check the updated structure: CannyBlock is now listed as first sub-block</span>
<span class="hljs-built_in">print</span>(blocks)
<span class="hljs-comment"># Check the updated doc</span>
<span class="hljs-built_in">print</span>(blocks.doc)<!----></pre></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 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> コピーしました</div><!----></button><!----></div> <pre class=" "><!---->class SequentialPipelineBlocks

  Inputs:
      image (<span class="hljs-string">`Union[Image, ndarray]`</span>):
          Image <span class="hljs-keyword">to</span> compute canny filter <span class="hljs-keyword">on</span>
      low_threshold (<span class="hljs-string">`int`</span>, *optional*, defaults <span class="hljs-keyword">to</span> <span class="hljs-number">50</span>):
          Low threshold for the canny filter.
      high_threshold (<span class="hljs-string">`int`</span>, *optional*, defaults <span class="hljs-keyword">to</span> <span class="hljs-number">200</span>):
          High threshold for the canny filter.
      prompt (<span class="hljs-string">`str`</span>):
          The prompt <span class="hljs-literal">or</span> prompts <span class="hljs-keyword">to</span> guide image generation.
      <span class="hljs-params">...</span><!----></pre></div><!----> <p>これにより、パイプラインは<code>control_image</code>ではなく<code>image</code>を入力として受け取るようになります。シーケンス内のブロックは自動的にデータを共有するため、cannyブロックの出力（<code>control_image</code>）はそれを必要とするdenoiseブロックに流れます。また、cannyブロックの入力（<code>image</code>）は、それより前のブロックから提供されないため、パイプラインの入力となります。</p> <p>修正されたブロックからパイプラインを作成し、ControlNet モデルを読み込みます。ControlNet は元のモデルリポジトリの一部ではないため、個別に読み込み、<a href="/docs/diffusers/v0.43.0/ja/api/modular_diffusers/pipeline#diffusers.ModularPipeline.update_components">update_components()</a> を使用して追加します。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!---->pipeline = blocks.init_pipeline(<span class="hljs-string">&quot;Qwen/Qwen-Image&quot;</span>, components_manager=manager)

pipeline.load_components(dtype=torch.bfloat16)

<span class="hljs-comment"># Load the ControlNet model</span>
controlnet_spec = pipeline.get_component_spec(<span class="hljs-string">&quot;controlnet&quot;</span>)
controlnet_spec.pretrained_model_name_or_path = <span class="hljs-string">&quot;InstantX/Qwen-Image-ControlNet-Union&quot;</span>
controlnet = controlnet_spec.load(dtype=torch.bfloat16)
pipeline.update_components(controlnet=controlnet)<!----></pre></div><!----> <p>次にパイプラインを実行します。cannyブロックは、ControlNet用に画像を前処理します。</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 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> コピーしました</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> load_image

prompt = <span class="hljs-string">&quot;cat wizard with red hat, gandalf, lord of the rings, detailed, fantasy, cute, adorable, Pixar, Disney&quot;</span>
image = load_image(<span class="hljs-string">&quot;https://github.com/Trgtuan10/Image_storage/blob/main/cute_cat.png?raw=true&quot;</span>)

output = pipeline(
    prompt=prompt,
    image=image,
).images[<span class="hljs-number">0</span>]
output<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="next-steps" 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="#next-steps"><span><svg 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><!---->次のステップ<!----></span></h2><!--]--><!----> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><!--[--><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">基本を学ぶ</div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">カスタムブロックの構築</div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">コンポーネントの共有</div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">ビジュアルインターフェース</div><!--]--></div> <div class="language-select"><div><p>Modular Diffusersのコアとなる構成要素を理解します:</p> <ul><li><a href="./pipeline_block">ModularPipelineBlocks</a>: パイプラインのステップを定義するための基本単位です。</li> <li><a href="./sequential_pipeline_blocks">SequentialPipelineBlocks</a>: ブロックを連鎖させて、順番に実行します。</li> <li><a href="./auto_pipeline_blocks">AutoPipelineBlocks</a>: 複数のワークフローをサポートするパイプラインを作成します。</li> <li><a href="./modular_diffusers_states">状態</a>: ブロック間でデータがどのように共有されるか。</li></ul><!----></div><!----> <div hidden=""><p><a href="./custom_blocks">Building Custom Blocks</a> ガイドで、カスタムロジックを含む独自のブロックを作成する方法を学びましょう。</p><!----></div><!----> <div hidden=""><p><a href="./components_manager"><code>ComponentsManager</code></a> を使用して、複数のパイプライン間でモデルを共有し、メモリを効率的に管理します。</p><!----></div><!----> <div hidden=""><p>モジュラーパイプラインを、ワークフローを構築するためのビジュアルなノードベースインターフェースである<a href="https://github.com/cubiq/Mellon" rel="nofollow">Mellon</a>に接続できます。Modular Diffusersで構築したカスタムブロックは、Mellonでそのまま動作します。UIコードは不要です。詳細はMellonのガイドを参照してください。</p><!----></div><!----><!----></div><!----> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/diffusers/blob/main/docs/source/en/modular_diffusers/quickstart.md" target="_blank"><svg class="mr-1" 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="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg><!----> <span><span class="underline">Update</span> on GitHub</span></a><!----> <p></p><!--]--><!----><!--]--><!--]--><!--]--> <!--[-1--><!--]--><!--]-->
			
			<script>
				{
					__sveltekit_1qsdbwf = {
						base: "/docs/diffusers/main/en",
						assets: "/docs/diffusers/main/en"
					};
					const element = document.currentScript.parentElement;
					Promise.all([
						import("/docs/diffusers/main/en/_app/immutable/entry/start.0S0_iEoG.js"),
						import("/docs/diffusers/main/en/_app/immutable/entry/app.CPdOxCbw.js")
					]).then(([kit, app]) => {
						kit.start(app, element, {
							node_ids: [0, 328],
							data: [null,null],
							form: null,
							error: null
						});
					});
				}
			</script>
		
Xet Storage Details

Size:: 57.8 kB
Xet hash:: 86699dd76a01ed1b161c2c20b3094086cb925b3ef6b09678e8a5d0de07f31940
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.