Buckets:

hf-doc-build
/

doc

Files
xet
hf-doc-build/doc / diffusers /main /pt /api /pipelines /flux3.html
mishig
about 3 hours ago
download
raw
96.9 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;FLUX 3 Image Klein&quot;,&quot;local&quot;:&quot;flux-3-image-klein&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Ampliação de prompt&quot;,&quot;local&quot;:&quot;prompt-upsampling&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Imagem para imagem&quot;,&quot;local&quot;:&quot;image-to-image&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Controle de layout&quot;,&quot;local&quot;:&quot;layout-control&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Escala de orientação&quot;,&quot;local&quot;:&quot;guidance-scale&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Modelo destilado por etapa&quot;,&quot;local&quot;:&quot;step-distilled-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Geração reproduzível&quot;,&quot;local&quot;:&quot;reproducible-generation&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Fluxos de trabalho&quot;,&quot;local&quot;:&quot;workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Callbacks por etapa&quot;,&quot;local&quot;:&quot;per-step-callbacks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Memória&quot;,&quot;local&quot;:&quot;memory&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Carregamento de arquivo único&quot;,&quot;local&quot;:&quot;single-file-loading&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3ModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.Flux3ModularPipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3AutoBlocks&quot;,&quot;local&quot;:&quot;diffusers.Flux3AutoBlocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3DistilledAutoBlocks&quot;,&quot;local&quot;:&quot;diffusers.Flux3DistilledAutoBlocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3DistilledModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.Flux3DistilledModularPipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/>
		<link href="/docs/diffusers/main/en/_app/immutable/entry/start.U6TUchwJ.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/yxSHiGDF.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/CPfOwUO_.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/entry/app.3qKhPoX-.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/BYe-f74j.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/chunks/D32hXVzh.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/0.DqpuBF3n.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/Dyp9xKq8.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/8.AKHyGgGH.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/Dk44y1gU.js" rel="modulepreload">
		<!--hchutw--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;FLUX 3 Image Klein&quot;,&quot;local&quot;:&quot;flux-3-image-klein&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Prompt upsampling&quot;,&quot;local&quot;:&quot;prompt-upsampling&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Image-to-image&quot;,&quot;local&quot;:&quot;image-to-image&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Layout control&quot;,&quot;local&quot;:&quot;layout-control&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Guidance scale&quot;,&quot;local&quot;:&quot;guidance-scale&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Step-distilled model&quot;,&quot;local&quot;:&quot;step-distilled-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Reproducible generation&quot;,&quot;local&quot;:&quot;reproducible-generation&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Workflows&quot;,&quot;local&quot;:&quot;workflows&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Per-step callbacks&quot;,&quot;local&quot;:&quot;per-step-callbacks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Memory&quot;,&quot;local&quot;:&quot;memory&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Single-file loading&quot;,&quot;local&quot;:&quot;single-file-loading&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3ModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.Flux3ModularPipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3AutoBlocks&quot;,&quot;local&quot;:&quot;diffusers.Flux3AutoBlocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3DistilledAutoBlocks&quot;,&quot;local&quot;:&quot;diffusers.Flux3DistilledAutoBlocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Flux3DistilledModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.Flux3DistilledModularPipeline&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> <!--[0--><h1 class="relative group"><a id="flux-3-image-klein" 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="#flux-3-image-klein"><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><!---->FLUX 3 Image Klein<!----></span></h1><!--]--><!----> <p>FLUX 3 Image Klein é um modelo de geração de imagens de 5,6B da Black Forest Labs: um único transformer (<a href="/docs/diffusers/main/pt/api/models/flux3_transformer#diffusers.Flux3Transformer2DModel">Flux3Transformer2DModel</a>) gera imagens a partir de um prompt de texto, sozinho ou junto com uma ou mais imagens de referência. A condicionamento de texto vem do Qwen3-VL, e as imagens usam o autoencoder empacotado FLUX.2 (<code>AutoencoderKLFlux2</code>).</p> <p>FLUX 3 Image Klein é distribuído como um <a href="../../modular_diffusers/overview">pipeline modular</a>. Comece com uma geração de imagem a partir de texto:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> ModularPipeline

pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>)
pipe.load_components(dtype={<span class="hljs-string">&quot;vae&quot;</span>: torch.float32, <span class="hljs-string">&quot;default&quot;</span>: torch.bfloat16})
pipe.to(<span class="hljs-string">&quot;cuda&quot;</span>)

image = pipe(
    prompt=<span class="hljs-string">&quot;A red fox in a snow-covered forest clearing at golden hour&quot;</span>,
    generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>),
    output=<span class="hljs-string">&quot;images&quot;</span>,
)[<span class="hljs-number">0</span>]
image.save(<span class="hljs-string">&quot;fox.png&quot;</span>)<!----></pre></div><!----> <p><code>prompt</code> é a única entrada obrigatória: o pipeline carrega as configurações de amostragem recomendadas (50 etapas, escala de orientação 3.5, deslocamento do agendador 6.93) e gera imagens de 1024 x 1024 por padrão. Ele aceita um prompt por chamada; use <code>num_images_per_prompt</code> para várias imagens do mesmo prompt.</p> <p>Como você pode ver no exemplo acima, o carregamento e a execução diferem um pouco de um <code>DiffusionPipeline</code>. O carregamento é preguiçoso: <a href="/docs/diffusers/main/pt/api/modular_diffusers/pipeline#diffusers.ModularPipeline.from_pretrained">from_pretrained()</a> apenas constrói o pipeline, e você chama <a href="/docs/diffusers/main/pt/api/modular_diffusers/pipeline#diffusers.ModularPipeline.load_components">load_components()</a> para carregar os pesos. Uma chamada retorna o estado do pipeline, e você solicita o que precisa com <code>output</code>: <code>output="images"</code> aqui, ou <code>output=["images", "prompt"]</code> para também obter a legenda que o prompt upsampler escreveu.</p> <blockquote class="tip"><p>Se você está trabalhando com <code>ModularPipeline</code> pela primeira vez, leia o <a href="../../modular_diffusers/quickstart">quickstart</a> e <a href="../../modular_diffusers/coming_from_diffusion_pipeline">Vindo do DiffusionPipeline</a> se você conhece os pipelines padrão. Além do carregamento e das saídas, algumas outras coisas diferem: <a href="#guidance-scale">guidance</a>, <a href="#per-step-callbacks">callbacks</a> e <a href="#memory">offloading</a>.</p></blockquote> <!--[1--><h2 class="relative group"><a id="prompt-upsampling" 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-upsampling"><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><!---->Ampliação de prompt<!----></span></h2><!--]--><!----> <p>O prompt importa: o FLUX 3 Image Klein aceita legendas completas de cena que descrevem o sujeito, a composição, a iluminação e a atmosfera, e não <code>prompt</code> curtos; uma solicitação enviada como digitada gera uma saída genérica. É por isso que o pipeline executa o upsampler de prompt primeiro em todas as solicitações, tanto para texto-para-imagem quanto para edição. Trata-se de um modelo separado de 9B da Black Forest Labs (<a href="https://huggingface.co/black-forest-labs/flux-3-image-klein-pup" rel="nofollow">black-forest-labs/flux-3-image-klein-pup</a>) que expande <code>prompt</code> para uma legenda assim antes do codificador de texto, e um prompt curto é tudo o que é necessário.</p> <p>O prompt upsampler transforma uma solicitação curta, junto com as imagens de referência durante a edição, na legenda completa da cena que o modelo recebe: meio e estilo, cena, iluminação e câmera, cada sujeito com suas cores, materiais e posição no quadro. Um único modelo cobre texto-para-imagem, edição com imagens de referência e layout com caixas. A legenda que ele escreveu retorna como a saída <code>prompt</code>. No exemplo acima, solicite-a junto com as imagens para ver o que o modelo de imagem recebeu:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->out = pipe(
    prompt=<span class="hljs-string">&quot;A red fox in a snow-covered forest clearing at golden hour&quot;</span>,
    generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>),
    output=[<span class="hljs-string">&quot;images&quot;</span>, <span class="hljs-string">&quot;prompt&quot;</span>],
)
<span class="hljs-built_in">print</span>(out[<span class="hljs-string">&quot;prompt&quot;</span>])<!----></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> Copiado</div><!----></button><!----></div> <pre class=" "><!---->Full-color outdoor wildlife photograph taken <span class="hljs-built_in">from</span> <span class="hljs-keyword">a</span> low eye-level perspective <span class="hljs-keyword">in</span> <span class="hljs-keyword">a</span> snow-covered forest clearing <span class="hljs-keyword">at</span> golden hour. In <span class="hljs-keyword">the</span> center foreground, <span class="hljs-keyword">a</span> red fox stands <span class="hljs-keyword">in</span> profile facing left, its body turned toward <span class="hljs-keyword">the</span> viewer <span class="hljs-keyword">and</span> its head raised <span class="hljs-keyword">as</span> <span class="hljs-keyword">it</span> looks across <span class="hljs-keyword">the</span> clearing. The fox has dense rust-red winter fur, <span class="hljs-keyword">a</span> white throat <span class="hljs-keyword">and</span> chest, <span class="hljs-keyword">black</span> <span class="hljs-built_in">lower</span> legs, pointed ears <span class="hljs-keyword">with</span> dark tips, <span class="hljs-keyword">and</span> <span class="hljs-keyword">a</span> <span class="hljs-keyword">long</span> bushy tail ending <span class="hljs-keyword">in</span> white. ...<!----></pre></div><!----> <p>A legenda completa tem cerca de 240 palavras; <code>max_new_tokens</code> limita seu comprimento (3584 por padrão).</p> <p>Você pode imprimir o <code>doc</code> da etapa para ver o que ela recebe e o que ela grava:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-built_in">print</span>(pipe.blocks.sub_blocks[<span class="hljs-string">&quot;prompt_upsampler&quot;</span>].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> Copiado</div><!----></button><!----></div> <pre class=" "><!---->  <span class="hljs-symbol">Inputs</span>:
      prompt (<span class="hljs-string">`str`</span>):
          <span class="hljs-symbol">The</span> edit instruction to expand into the full caption (one per call).
      image (<span class="hljs-string">`Image | list`</span>, *optional*):
          <span class="hljs-symbol">The</span> reference image(s) of the edit, at most <span class="hljs-number">10</span>; the same ones the pipeline receives.
      boxes (<span class="hljs-string">`list`</span>, *optional*):
          <span class="hljs-symbol">Layout</span> boxes, each {<span class="hljs-string">&#x27;id&#x27;</span>, <span class="hljs-string">&#x27;desc&#x27;</span>, <span class="hljs-string">&#x27;from&#x27;</span>, <span class="hljs-string">&#x27;src_bbox&#x27;</span>, <span class="hljs-string">&#x27;tgt_bbox&#x27;</span>} with the boxes as [y_min, x_min, y_max, x_max] in
          <span class="hljs-number">0</span><span class="hljs-number">-1</span>: ...
      max_new_tokens (<span class="hljs-string">`int`</span>, *optional*, defaults to <span class="hljs-number">3584</span>):
          <span class="hljs-symbol">The</span> most tokens the upsampler may generate for the caption.

  <span class="hljs-symbol">Outputs</span>:
      prompt (<span class="hljs-string">`str`</span>):
          <span class="hljs-symbol">The</span> full caption, replacing the request.<!----></pre></div><!----> <p>Assim como os outros blocos do pipeline, o bloco <code>prompt_upsampler</code> também é executado de forma independente. A maneira recomendada é removê-lo dos blocos: o bloco removido se torna um upsampler autônomo, e os blocos restantes formam um pipeline que recebe a legenda expandida como <code>prompt</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 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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->blocks = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>).blocks
pup_block = blocks.sub_blocks.pop(<span class="hljs-string">&quot;prompt_upsampler&quot;</span>)

pup = pup_block.init_pipeline(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>)
pup.load_components(dtype=torch.bfloat16)
pup.to(<span class="hljs-string">&quot;cuda&quot;</span>)
caption = pup(prompt=<span class="hljs-string">&quot;A red fox in a snow-covered forest clearing at golden hour&quot;</span>, output=<span class="hljs-string">&quot;prompt&quot;</span>)

pipe = blocks.init_pipeline(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>)
pipe.load_components(dtype={<span class="hljs-string">&quot;vae&quot;</span>: torch.float32, <span class="hljs-string">&quot;default&quot;</span>: torch.bfloat16})
pipe.to(<span class="hljs-string">&quot;cuda&quot;</span>)
image = pipe(prompt=caption, generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>), output=<span class="hljs-string">&quot;images&quot;</span>)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p>Dessa forma, ele pode ser servido como um estágio separado, em sua própria GPU ou atrás de seu próprio endpoint, e uma interface pode exibir a legenda que ele escreveu e permitir que o usuário a edite antes de gerar. Qualquer outro LLM pode assumir o lugar do upsampler, com os mesmos system prompts: forneça <code>system_prompts["t2i"]</code> do repositório do upsampler (ou <code>"i2i"</code>, com as referências primeiro na mensagem do usuário) como o system prompt e a solicitação como a mensagem do usuário, e passe a legenda que ele retorna como <code>prompt</code> para <code>pipe</code>.</p> <!--[1--><h2 class="relative group"><a id="image-to-image" 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="#image-to-image"><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><!---->Imagem para imagem<!----></span></h2><!--]--><!----> <p>Passe uma ou mais imagens de referência com <code>image</code>. Uma única referência é mantida inteira: ela é redimensionada apenas acima de 4 megapixels, preenchida até a grade de 16 pixels e a geração usa seu tamanho por padrão, de modo que a edição retorna no tamanho da entrada. Várias referências são ajustadas individualmente para menos de 1 megapixel e recortadas para a grade, e a geração usa o tamanho da primeira por padrão. Um <code>height</code>/<code>width</code> específico substitui o padrão. Descreva a imagem <em>alvo</em> como uma legenda completa da referência com sua edição incorporada, em vez de uma instrução de uma linha, ou deixe que o <a href="#prompt-upsampling">prompt upsampler</a> a escreva a partir da instrução e das referências:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->night = pipe(
    prompt=<span class="hljs-string">&quot;the same scene at night under a full moon&quot;</span>,
    image=image,
    generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>),
    output=[<span class="hljs-string">&quot;images&quot;</span>, <span class="hljs-string">&quot;prompt&quot;</span>],
)<!----></pre></div><!----> <p>Com várias referências, refira-se a elas como a primeira imagem, a segunda imagem, … na ordem em que são passadas para <code>image</code>; a legenda as nomeia como <code>&lt;ref_image_0></code>, <code>&lt;ref_image_1></code>, … para o modelo de imagem:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->combined = pipe(
    prompt=<span class="hljs-string">&quot;the fox from the first image sitting on the wooden bench from the second image&quot;</span>,
    image=[image, bench],
    height=<span class="hljs-number">768</span>,
    width=<span class="hljs-number">1360</span>,
    output=<span class="hljs-string">&quot;images&quot;</span>,
)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p>Pass <code>height</code> e/ou <code>width</code> para substituir o tamanho derivado.</p> <!--[1--><h2 class="relative group"><a id="layout-control" 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="#layout-control"><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><!---->Controle de layout<!----></span></h2><!--]--><!----> <p>Para posicionar o conteúdo explicitamente, passe <code>boxes</code>: para cada elemento, um <code>id</code>, uma <code>desc</code> curta e um <code>bbox</code> de <code>[y_min, x_min, y_max, x_max]</code> em 0-1 (origem no canto superior esquerdo), para que seja independente do tamanho da saída. O upsampler escreve a legenda ao redor deles, citando cada elemento como <code>&lt;id></code>, e anexa a tabela de boxes, mantendo os seus exatamente:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->boxes = [
    {<span class="hljs-string">&quot;id&quot;</span>: <span class="hljs-string">&quot;balloon_1&quot;</span>, <span class="hljs-string">&quot;desc&quot;</span>: <span class="hljs-string">&quot;a red and yellow striped hot-air balloon&quot;</span>, <span class="hljs-string">&quot;bbox&quot;</span>: [<span class="hljs-number">0.08</span>, <span class="hljs-number">0.06</span>, <span class="hljs-number">0.48</span>, <span class="hljs-number">0.38</span>]},
    {<span class="hljs-string">&quot;id&quot;</span>: <span class="hljs-string">&quot;balloon_2&quot;</span>, <span class="hljs-string">&quot;desc&quot;</span>: <span class="hljs-string">&quot;a small blue hot-air balloon&quot;</span>, <span class="hljs-string">&quot;bbox&quot;</span>: [<span class="hljs-number">0.20</span>, <span class="hljs-number">0.62</span>, <span class="hljs-number">0.42</span>, <span class="hljs-number">0.80</span>]},
]
out = pipe(prompt=<span class="hljs-string">&quot;hot-air balloons over a green valley at sunrise&quot;</span>, boxes=boxes, output=[<span class="hljs-string">&quot;images&quot;</span>, <span class="hljs-string">&quot;prompt&quot;</span>])<!----></pre></div><!----> <p>Ao editar, uma caixa nomeia sua origem com <code>from</code>: <code>"target"</code> com um <code>tgt_bbox</code> adiciona um novo elemento ali, <code>"ref_image_0"</code> com um <code>src_bbox</code> e um <code>tgt_bbox</code> move um elemento daquela referência (caixas iguais o mantêm no lugar):</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->boxes = [{<span class="hljs-string">&quot;id&quot;</span>: <span class="hljs-string">&quot;cactus&quot;</span>, <span class="hljs-string">&quot;desc&quot;</span>: <span class="hljs-string">&quot;a small potted cactus&quot;</span>, <span class="hljs-string">&quot;from&quot;</span>: <span class="hljs-string">&quot;target&quot;</span>, <span class="hljs-string">&quot;src_bbox&quot;</span>: <span class="hljs-literal">None</span>, <span class="hljs-string">&quot;tgt_bbox&quot;</span>: [<span class="hljs-number">0.55</span>, <span class="hljs-number">0.62</span>, <span class="hljs-number">0.85</span>, <span class="hljs-number">0.85</span>]}]
out = pipe(prompt=<span class="hljs-string">&quot;add a small potted cactus in the marked spot&quot;</span>, image=mug, boxes=boxes, output=[<span class="hljs-string">&quot;images&quot;</span>, <span class="hljs-string">&quot;prompt&quot;</span>])

boxes = [{<span class="hljs-string">&quot;id&quot;</span>: <span class="hljs-string">&quot;knight&quot;</span>, <span class="hljs-string">&quot;desc&quot;</span>: <span class="hljs-string">&quot;the knight figure&quot;</span>, <span class="hljs-string">&quot;from&quot;</span>: <span class="hljs-string">&quot;ref_image_0&quot;</span>, <span class="hljs-string">&quot;src_bbox&quot;</span>: [<span class="hljs-number">0.50</span>, <span class="hljs-number">0.15</span>, <span class="hljs-number">0.85</span>, <span class="hljs-number">0.35</span>], <span class="hljs-string">&quot;tgt_bbox&quot;</span>: [<span class="hljs-number">0.19</span>, <span class="hljs-number">0.06</span>, <span class="hljs-number">0.54</span>, <span class="hljs-number">0.26</span>]}]
out = pipe(prompt=<span class="hljs-string">&quot;move the knight up and to the left, keep the rest unchanged&quot;</span>, image=knight, boxes=boxes, output=[<span class="hljs-string">&quot;images&quot;</span>, <span class="hljs-string">&quot;prompt&quot;</span>])<!----></pre></div><!----> <p>A legenda que o modelo de imagem lê tem os elementos marcados com <code>&lt;id></code> no texto e a tabela anexada como JSON, com as caixas em um canvas de 0 a 1000; imprima a saída de <code>prompt</code> para vê-la. Uma legenda escrita manualmente nesse formato vai para o pipeline sem a etapa, como em <a href="#prompt-upsampling">Prompt upsampling</a>.</p> <!--[1--><h2 class="relative group"><a id="guidance-scale" 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="#guidance-scale"><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><!---->Escala de orientação<!----></span></h2><!--]--><!----> <p>O pipeline aplica a orientação sem classificador por meio de seu componente <code>guider</code> e não aceita um argumento <code>guidance_scale</code> no momento da chamada. O guider padrão é <a href="/docs/diffusers/main/pt/api/modular_diffusers/guiders#diffusers.ClassifierFreeGuidance">ClassifierFreeGuidance</a> com <code>guidance_scale=3.5</code>, a configuração recomendada para ambos os fluxos de trabalho. Para alterá-lo, crie uma cópia atualizada do guider:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->pipe.guider = pipe.guider.new(guidance_scale=<span class="hljs-number">5.0</span>)<!----></pre></div><!----> <p>Consulte o <a href="../../using-diffusers/guiders">guia de guias</a> para alternar para um método de orientação diferente.</p> <!--[1--><h2 class="relative group"><a id="step-distilled-model" 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="#step-distilled-model"><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><!---->Modelo destilado por etapa<!----></span></h2><!--]--><!----> <p>O checkpoint destilado por etapas gera em 16 passos sem classifier-free guidance. Seu transformer é um mapa de fluxo: em cada passo, ele recebe o timestep atual e aquele para o qual o passo salta. Ele é carregado como o modelo base (o <code>modular_model_index.json</code> do repositório resolve para <a href="/docs/diffusers/main/pt/api/pipelines/flux3#diffusers.Flux3DistilledModularPipeline">Flux3DistilledModularPipeline</a> e <a href="/docs/diffusers/main/pt/api/pipelines/flux3#diffusers.Flux3DistilledAutoBlocks">Flux3DistilledAutoBlocks</a>); o pipeline não possui <code>guider</code> e não aceita <code>negative_prompt</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 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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein&quot;</span>)
pipe.load_components(dtype=torch.bfloat16)
pipe.to(<span class="hljs-string">&quot;cuda&quot;</span>)

image = pipe(prompt=prompt, generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>), output=<span class="hljs-string">&quot;images&quot;</span>)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p>Ele atende aos mesmos dois fluxos de trabalho que o modelo base. <code>num_inference_steps</code> tem o valor padrão de 16; o modelo é treinado para 4 a 16 etapas.</p> <!--[1--><h2 class="relative group"><a id="reproducible-generation" 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="#reproducible-generation"><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><!---->Geração reproduzível<!----></span></h2><!--]--><!----> <p>Um <code>generator</code> fixa o ruído que o pipeline sorteia (veja <a href="../../using-diffusers/reusing_seeds">Reprodutibilidade</a>). Para reutilizar um ruído sorteado em outro lugar, passe-o como <code>latents</code> com a forma <code>(batch_size, 128, height // 16, width // 16)</code>. Por exemplo, a <a href="https://github.com/black-forest-labs/flux-3-image-klein" rel="nofollow">implementação da BFL</a> sorteia seu ruído como tokens de imagem na CPU e o converte para bf16; isso reproduz sua amostra para uma semente dada:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->height, width = <span class="hljs-number">1024</span>, <span class="hljs-number">1024</span>
h, w = height // <span class="hljs-number">16</span>, width // <span class="hljs-number">16</span>

noise = torch.randn((h * w, <span class="hljs-number">128</span>), generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>)).to(torch.bfloat16)
latents = noise.T.reshape(<span class="hljs-number">1</span>, <span class="hljs-number">128</span>, h, w)

image = pipe(prompt=prompt, latents=latents, height=height, width=width, output=<span class="hljs-string">&quot;images&quot;</span>)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <!--[1--><h2 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><!---->Fluxos de trabalho<!----></span></h2><!--]--><!----> <p>O pipeline atende a dois fluxos de trabalho, <code>text2image</code> e <code>image_conditioned</code>; o que é executado é selecionado pelas entradas (<code>image</code> seleciona image-to-image). <a href="/docs/diffusers/main/pt/api/modular_diffusers/pipeline#diffusers.ModularPipeline.from_pretrained">from_pretrained()</a> também pode construir um pipeline para um fluxo de trabalho com o argumento <code>workflow</code>, que então espera e carrega apenas os componentes necessários para esse fluxo de trabalho:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">from</span> diffusers.modular_pipelines <span class="hljs-keyword">import</span> Flux3AutoBlocks

<span class="hljs-built_in">print</span>(Flux3AutoBlocks().available_workflows)
<span class="hljs-comment"># [&quot;text2image&quot;, &quot;image_conditioned&quot;]</span>

t2i = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>, workflow=<span class="hljs-string">&quot;text2image&quot;</span>)
<span class="hljs-built_in">print</span>(t2i.blocks.doc)<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="per-step-callbacks" 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="#per-step-callbacks"><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><!---->Callbacks por etapa<!----></span></h2><!--]--><!----> <p>Para usar uma função de callback existente, insira um bloco no loop de desruidificação. O <a href="../../using-diffusers/callback#modular-pipelines">guia de callbacks</a> tem um exemplo simples com o bloco <code>CallbackOnStepEnd</code> e o auxiliar <code>walk_denoise_loops</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 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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->blocks = Flux3AutoBlocks()  <span class="hljs-comment"># or Flux3DistilledAutoBlocks</span>
<span class="hljs-keyword">for</span> loop <span class="hljs-keyword">in</span> walk_denoise_loops(blocks):
    loop.sub_blocks[<span class="hljs-string">&quot;callback&quot;</span>] = CallbackOnStepEnd()
pipe = blocks.init_pipeline(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>)

image = pipe(prompt=caption, callback_on_step_end=my_callback, output=<span class="hljs-string">&quot;images&quot;</span>)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p>Para interromper precocemente, lance uma exceção a partir do callback; não existe a flag <code>interrupt</code>.</p> <!--[1--><h2 class="relative group"><a id="memory" 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="#memory"><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><!---->Memória<!----></span></h2><!--]--><!----> <p>Carregar tudo em bf16 consome cerca de 39 GB de pesos: o transformer tem 5,5B parâmetros, o codificador de texto Qwen3-VL-4B tem 4B e o <a href="#prompt-upsampling">prompt upsampler</a> tem 9B (cerca de 19 GB por si só), com o VAE de imagem em fp32. Um pipeline construído sem a etapa de upsampler (veja lá) carrega cerca de 20 GB. Para GPUs com menos memória, habilite o offloading de componentes; o gerenciador move os componentes entre a CPU e a GPU conforme cada etapa os necessita, incluindo o upsampler:</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> ComponentsManager, ModularPipeline

manager = ComponentsManager()
pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>, components_manager=manager)
pipe.load_components(dtype={<span class="hljs-string">&quot;vae&quot;</span>: torch.float32, <span class="hljs-string">&quot;default&quot;</span>: torch.bfloat16})
manager.enable_auto_cpu_offload(device=<span class="hljs-string">&quot;cuda&quot;</span>)<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="single-file-loading" 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="#single-file-loading"><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><!---->Carregamento de arquivo único<!----></span></h2><!--]--><!----> <p><a href="/docs/diffusers/main/pt/api/models/flux3_transformer#diffusers.Flux3Transformer2DModel">Flux3Transformer2DModel</a> carrega o transformer diretamente de um pacote de modelo da Black Forest Labs com <a href="/docs/diffusers/main/pt/api/loaders/single_file#diffusers.loaders.FromOriginalModelMixin.from_single_file">from_single_file()</a>. O pacote <code>config.json</code> não é lido; passe a configuração do diffusers com <code>config</code> e <code>subfolder</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 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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> Flux3Transformer2DModel, ModularPipeline

transformer = Flux3Transformer2DModel.from_single_file(
    <span class="hljs-string">&quot;https://huggingface.co/black-forest-labs/flux-3-image-klein-base/blob/main/model.safetensors&quot;</span>,
    config=<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>,
    subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
    dtype=torch.bfloat16,
)

pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>)
pipe.load_components(dtype={<span class="hljs-string">&quot;vae&quot;</span>: torch.float32, <span class="hljs-string">&quot;default&quot;</span>: torch.bfloat16})
pipe.update_components(transformer=transformer)
pipe.to(<span class="hljs-string">&quot;cuda&quot;</span>)<!----></pre></div><!----> <p>Carregue primeiro os outros componentes pelo nome: <a href="/docs/diffusers/main/pt/api/modular_diffusers/pipeline#diffusers.ModularPipeline.load_components">load_components()</a> sem <code>names</code> também carregaria o transformer do repositório do pipeline.</p> <p>O carregamento dos checkpoints de arquivo único FP8r e NVFP4 também é suportado. Ambos são carregados por meio da mesma chamada com um <a href="/docs/diffusers/main/pt/api/quantization#diffusers.BFLQuantizationConfig">BFLQuantizationConfig</a>. Consulte o <a href="../../quantization/bfl">guia de quantização BFL</a> para os requisitos de kernel de cada esquema.</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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> BFLQuantizationConfig

transformer = Flux3Transformer2DModel.from_single_file(
    <span class="hljs-string">&quot;https://huggingface.co/black-forest-labs/flux-3-image-klein-base/blob/main/variants/nvfp4/model.safetensors&quot;</span>,
    config=<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein-base&quot;</span>,
    subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
    quantization_config=BFLQuantizationConfig(compute_dtype=torch.bfloat16),
    dtype=torch.bfloat16,
)<!----></pre></div><!----> <p>Ao carregar os checkpoints destilados por etapa, precisamos de seus
configs (<code>flow_map=True</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 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> Copiado</div><!----></button><!----></div> <pre class="language-python "><!---->transformer = Flux3Transformer2DModel.from_single_file(
    <span class="hljs-string">&quot;https://huggingface.co/black-forest-labs/flux-3-image-klein/blob/main/variants/nvfp4/model.safetensors&quot;</span>,
    config=<span class="hljs-string">&quot;black-forest-labs/flux-3-image-klein&quot;</span>,
    subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
    quantization_config=BFLQuantizationConfig(compute_dtype=torch.bfloat16),
    dtype=torch.bfloat16,
)<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="diffusers.Flux3ModularPipeline" 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="#diffusers.Flux3ModularPipeline"><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><!---->Flux3ModularPipeline<!----></span></h2><!--]--><!----> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"><div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="diffusers.Flux3ModularPipeline"><!----><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">classe</span> <span class="font-medium">diffusers.</span><span class="font-semibold">Flux3ModularPipeline</span></span></h3><!----> <a id="diffusers.Flux3ModularPipeline" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.Flux3ModularPipeline"><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><!----></a> <!--[0--><a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/diffusers/blob/main/src/diffusers/modular_pipelines/flux3/modular_pipeline.py#L24" target="_blank"><span>&lt;</span> <span class="hidden md:block mx-0.5 hover:!underline">source</span> <span>></span></a><!--]--></span> <!--[0--><p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span>(</span> <!--[--><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">blocks<span class="opacity-60">: diffusers.modular_pipelines.modular_pipeline.ModularPipelineBlocks | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">pretrained_model_name_or_path<span class="opacity-60">: str | os.PathLike | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">components_manager<span class="opacity-60">: diffusers.modular_pipelines.components_manager.ComponentsManager | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">collection<span class="opacity-60">: str | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">workflow<span class="opacity-60">: str | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">modular_config_dict<span class="opacity-60">: dict[str, typing.Any] | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">config_dict<span class="opacity-60">: dict[str, typing.Any] | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">**kwargs<span class="opacity-60"></span></span></span><!--]--> <span>)</span> <!--[-1--><!--]--></p><!--]--> <div class="!mb-10 relative docstring-details "><!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--></div></div><!----> <p>Uma ModularPipeline para FLUX.3.</p></div> <!--[1--><h2 class="relative group"><a id="diffusers.Flux3AutoBlocks" 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="#diffusers.Flux3AutoBlocks"><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><!---->Flux3AutoBlocks<!----></span></h2><!--]--><!----> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"><div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="diffusers.Flux3AutoBlocks"><!----><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">classe</span> <span class="font-medium">diffusers.</span><span class="font-semibold">Flux3AutoBlocks</span></span></h3><!----> <a id="diffusers.Flux3AutoBlocks" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.Flux3AutoBlocks"><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><!----></a> <!--[0--><a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/diffusers/blob/main/src/diffusers/modular_pipelines/flux3/modular_blocks_flux3.py#L339" target="_blank"><span>&lt;</span> <span class="hidden md:block mx-0.5 hover:!underline">source</span> <span>></span></a><!--]--></span> <!--[0--><p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span>(</span> <!--[--><!--]--> <span>)</span> <!--[-1--><!--]--></p><!--]--> <div class="!mb-10 relative docstring-details "><!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--></div></div><!----> <p>Auto blocks que executam a geração de texto para imagem e imagem para imagem usando FLUX.3. O <code>prompt</code> passa primeiro pelo prompt upsampler; remova a etapa <code>prompt_upsampler</code> para codificá-lo como fornecido.</p> <p>Fluxos de trabalho suportados:</p> <ul><li><code>text2image</code>: requer <code>prompt</code></li> <li><code>image_conditioned</code>: requer <code>prompt</code>, <code>image</code></li></ul> <p>Componentes:
prompt_upsampler (<code>Qwen3_5ForConditionalGeneration</code>): O prompt upsampler. prompt_upsampler_processor
(<code>AutoProcessor</code>): Seu processor. text_encoder (<code>Qwen3VLForConditionalGeneration</code>) tokenizer
(<code>AutoProcessor</code>) image_processor (<code>Flux3ImageProcessor</code>) vae (<code>AutoencoderKLFlux2</code>) transformer
(<code>Flux3Transformer2DModel</code>) scheduler (<code>FlowMatchEulerDiscreteScheduler</code>) guider (<code>ClassifierFreeGuidance</code>)</p> <p>Entradas:
prompt (<code>str</code>):
A instrução de edição a ser expandida para a legenda completa (uma por chamada).
image (<code>Image | list</code>, <em>opcional</em>):
A(s) imagem(ns) de referência da edição, no máximo 10; as mesmas que o pipeline recebe.
boxes (<code>list</code>, <em>opcional</em>):
Caixas de layout, cada uma {‘id’, ‘desc’, ‘from’, ‘src_bbox’, ‘tgt_bbox’} com as caixas como [y_min, x_min, y_max,
x_max] em 0-1: ‘from’: ‘target’ com <code>tgt_bbox</code> adiciona um elemento ali, ‘from’: ‘ref_image_0’, … com <code>src_bbox</code> e <code>tgt_bbox</code> move um de uma imagem de referência.
max_new_tokens (<code>int</code>, <em>opcional</em>, padrão 3584):
O número máximo de tokens que o upsampler pode gerar para a legenda.
negative_prompt (<code>str</code>, <em>opcional</em>):
O prompt que não deve guiar a geração
max_sequence_length (<code>int</code>, <em>opcional</em>, padrão 8192):
Número máximo de tokens do prompt; prompts mais longos são truncados.
text_encoder_out_layers (<code>tuple</code>, <em>opcional</em>, padrão (4, 8, 12, 16, 20, 24, 28, 32)):
Índices das camadas ocultas do text-encoder cujos estados são empilhados para formar os embeddings de texto
num_images_per_prompt (<code>int</code>, <em>opcional</em>, padrão 1):
O número de imagens a gerar por prompt.
image_cond_latents (<code>list</code>, <em>opcional</em>):
Latents empacotados normalizados das imagens de referência, um tensor de forma (B, C, height_i//16, width_i//16)
por imagem
height (<code>int</code>, <em>opcional</em>, padrão None ou 1024, dependendo do fluxo de trabalho):
A altura em pixels da imagem gerada; padrão para a própria altura da primeira imagem de referência
width (<code>int</code>, <em>opcional</em>, padrão None ou 1024, dependendo do fluxo de trabalho):
A largura em pixels da imagem gerada; padrão para a própria largura da primeira imagem de referência
reference_crop (<code>list | NoneType</code>, <em>opcional</em>):
Por referência, a caixa da referência dentro do seu canvas com padding ou None, da etapa de pré-processamento
latents (<code>Tensor</code>):
Latents ruidosos pré-gerados para a geração de imagem.
generator (<code>Generator</code>, <em>opcional</em>):
Gerador do Torch para geração determinística.
num_inference_steps (<code>int</code>):
O número de etapas de remoção de ruído.
timesteps (<code>Tensor</code>):
Timesteps para o processo de remoção de ruído. (sobrepõe <code>num_inference_steps</code> quando fornecido)
sigmas (<code>list</code>, <em>opcional</em>):
Sigmas personalizados para o processo de remoção de ruído. (sobrepõe <code>num_inference_steps</code> quando fornecido)
*<em>denoiser_input_fields (<code>None</code>,</em>opcional<em>):
entradas condicionais do modelo para o denoiser: por exemplo, prompt_embeds, negative_prompt_embeds, etc.
attention_kwargs (<code>dict</code>,</em>opcional<em>):
Kwargs adicionais para os processadores de atenção.
output_type (<code>str</code>,</em>opcional*, padrão pil):
Formato de saída: ‘pil’, ‘np’, ‘pt’.</p> <p>Saídas:
images (<code>list</code>):
Imagens geradas.
latents (<code>Tensor</code>):
Latentes dessonificados.</p></div> <!--[1--><h2 class="relative group"><a id="diffusers.Flux3DistilledAutoBlocks" 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="#diffusers.Flux3DistilledAutoBlocks"><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><!---->Flux3DistilledAutoBlocks<!----></span></h2><!--]--><!----> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"><div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="diffusers.Flux3DistilledAutoBlocks"><!----><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">classe</span> <span class="font-medium">diffusers.</span><span class="font-semibold">Flux3DistilledAutoBlocks</span></span></h3><!----> <a id="diffusers.Flux3DistilledAutoBlocks" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.Flux3DistilledAutoBlocks"><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><!----></a> <!--[0--><a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/diffusers/blob/main/src/diffusers/modular_pipelines/flux3/modular_blocks_flux3_distilled.py#L258" target="_blank"><span>&lt;</span> <span class="hidden md:block mx-0.5 hover:!underline">source</span> <span>></span></a><!--]--></span> <!--[0--><p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span>(</span> <!--[--><!--]--> <span>)</span> <!--[-1--><!--]--></p><!--]--> <div class="!mb-10 relative docstring-details "><!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--></div></div><!----> <p>Auto blocks que executam a geração de texto para imagem e imagem para imagem usando o modelo FLUX.3 destilado por etapas (16 etapas, sem classifier-free guidance). O <code>prompt</code> passa primeiro pelo prompt upsampler; remova a etapa <code>prompt_upsampler</code> para codificá-lo como fornecido.</p> <p>Fluxos de trabalho suportados:</p> <ul><li><code>text2image</code>: requer <code>prompt</code></li> <li><code>image_conditioned</code>: requer <code>prompt</code>, <code>image</code></li></ul> <p>Componentes:
prompt_upsampler (<code>Qwen3_5ForConditionalGeneration</code>): O prompt upsampler. prompt_upsampler_processor
(<code>AutoProcessor</code>): Seu processador. text_encoder (<code>Qwen3VLForConditionalGeneration</code>) tokenizer
(<code>AutoProcessor</code>) image_processor (<code>Flux3ImageProcessor</code>) vae (<code>AutoencoderKLFlux2</code>) transformer
(<code>Flux3Transformer2DModel</code>) scheduler (<code>FlowMatchEulerDiscreteScheduler</code>)</p> <p>Entradas:
prompt (<code>str</code>):
A instrução de edição a ser expandida para a legenda completa (uma por chamada).
image (<code>Image | list</code>, <em>opcional</em>):
A(s) imagem(ns) de referência da edição, no máximo 10; as mesmas que o pipeline recebe.
boxes (<code>list</code>, <em>opcional</em>):
Caixas de layout, cada uma {‘id’, ‘desc’, ‘from’, ‘src_bbox’, ‘tgt_bbox’} com as caixas como [y_min, x_min, y_max,
x_max] em 0-1: ‘from’: ‘target’ com <code>tgt_bbox</code> adiciona um elemento ali, ‘from’: ‘ref_image_0’, … com <code>src_bbox</code> e <code>tgt_bbox</code> move um de uma imagem de referência.
max_new_tokens (<code>int</code>, <em>opcional</em>, padrão 3584):
O número máximo de tokens que o upsampler pode gerar para a legenda.
max_sequence_length (<code>int</code>, <em>opcional</em>, padrão 8192):
Número máximo de tokens do prompt; prompts mais longos são truncados.
text_encoder_out_layers (<code>tuple</code>, <em>opcional</em>, padrão (4, 8, 12, 16, 20, 24, 28, 32)):
Índices das camadas ocultas do text-encoder cujos estados são empilhados para formar os text embeddings
num_images_per_prompt (<code>int</code>, <em>opcional</em>, padrão 1):
O número de imagens a serem geradas por prompt.
image_cond_latents (<code>list</code>, <em>opcional</em>):
Latents empacotados normalizados das imagens de referência, um tensor de forma (B, C, height_i//16, width_i//16)
por imagem
height (<code>int</code>, <em>opcional</em>, padrão None ou 1024, dependendo do workflow):
A altura em pixels da imagem gerada; padrão para a própria altura da primeira imagem de referência
width (<code>int</code>, <em>opcional</em>, padrão None ou 1024, dependendo do workflow):
A largura em pixels da imagem gerada; padrão para a própria largura da primeira imagem de referência
reference_crop (<code>list | NoneType</code>, <em>opcional</em>):
Por referência, a caixa da referência dentro do seu canvas com padding ou None, da etapa de pré-processamento
latents (<code>Tensor</code>):
Latents ruidosos pré-gerados para a geração de imagem.
generator (<code>Generator</code>, <em>opcional</em>):
Gerador do Torch para geração determinística.
num_inference_steps (<code>int</code>):
O número de etapas de remoção de ruído.
timesteps (<code>Tensor</code>):
Timesteps para o processo de remoção de ruído. (sobrepõe <code>num_inference_steps</code> quando fornecido)
sigmas (<code>list</code>, <em>opcional</em>):
Sigmas personalizados para o processo de remoção de ruído. (sobrepõe <code>num_inference_steps</code> quando fornecido)
*<em>denoiser_input_fields (<code>None</code>,</em>opcional<em>):
entradas condicionais do modelo para o denoiser: por exemplo, prompt_embeds, negative_prompt_embeds, etc.
output_type (<code>str</code>,</em>opcional*, padrão pil):
Formato de saída: ‘pil’, ‘np’, ‘pt’.</p> <p>Saídas:
images (<code>list</code>):
Imagens geradas.
latents (<code>Tensor</code>):
Latentes dessonificados.</p></div> <!--[1--><h2 class="relative group"><a id="diffusers.Flux3DistilledModularPipeline" 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="#diffusers.Flux3DistilledModularPipeline"><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><!---->Flux3DistilledModularPipeline<!----></span></h2><!--]--><!----> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"><div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="diffusers.Flux3DistilledModularPipeline"><!----><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">classe</span> <span class="font-medium">diffusers.</span><span class="font-semibold">Flux3DistilledModularPipeline</span></span></h3><!----> <a id="diffusers.Flux3DistilledModularPipeline" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.Flux3DistilledModularPipeline"><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><!----></a> <!--[0--><a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/diffusers/blob/main/src/diffusers/modular_pipelines/flux3/modular_pipeline.py#L52" target="_blank"><span>&lt;</span> <span class="hidden md:block mx-0.5 hover:!underline">source</span> <span>></span></a><!--]--></span> <!--[0--><p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span>(</span> <!--[--><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">blocks<span class="opacity-60">: diffusers.modular_pipelines.modular_pipeline.ModularPipelineBlocks | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">pretrained_model_name_or_path<span class="opacity-60">: str | os.PathLike | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">components_manager<span class="opacity-60">: diffusers.modular_pipelines.components_manager.ComponentsManager | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">collection<span class="opacity-60">: str | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">workflow<span class="opacity-60">: str | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">modular_config_dict<span class="opacity-60">: dict[str, typing.Any] | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">config_dict<span class="opacity-60">: dict[str, typing.Any] | None = None</span></span></span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">**kwargs<span class="opacity-60"></span></span></span><!--]--> <span>)</span> <!--[-1--><!--]--></p><!--]--> <div class="!mb-10 relative docstring-details "><!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--> <!--[-1--><!--]--></div></div><!----> <p>Uma ModularPipeline para o modelo FLUX.3 com destilação de etapas (16 etapas, sem classifier-free guidance).</p></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/api/pipelines/flux3.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_gxjg5b = {
						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.U6TUchwJ.js"),
						import("/docs/diffusers/main/en/_app/immutable/entry/app.3qKhPoX-.js")
					]).then(([kit, app]) => {
						kit.start(app, element, {
							node_ids: [0, 8],
							data: [null,null],
							form: null,
							error: null
						});
					});
				}
			</script>
		
Xet Storage Details

Size:: 96.9 kB
Xet hash:: 7e87b1ab5a441ccfef268c335e99f09cefe691d75c12928bea904e7049036c95
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.