Buckets:

hf-doc-build/doc / diffusers /main /pt /api /pipelines /minimax_h3_image.html
mishig's picture
download
raw
43 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;MiniMax-H3-Image&quot;,&quot;local&quot;:&quot;minimax-h3-image&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Componentes&quot;,&quot;local&quot;:&quot;components&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;Melhoria de prompt&quot;,&quot;local&quot;:&quot;prompt-enhancement&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Tamanho da imagem&quot;,&quot;local&quot;:&quot;image-size&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Amostragem&quot;,&quot;local&quot;:&quot;sampling&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;MiniMaxH3ImageModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.MiniMaxH3ImageModularPipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;MiniMaxH3ImageBlocks&quot;,&quot;local&quot;:&quot;diffusers.MiniMaxH3ImageBlocks&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/9.BHf5c0WW.js" rel="modulepreload">
<link href="/docs/diffusers/main/en/_app/immutable/chunks/Dk44y1gU.js" rel="modulepreload">
<link href="/docs/diffusers/main/en/_app/immutable/chunks/C4N0BX8L.js" rel="modulepreload">
<!--mp1ecb--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;MiniMax-H3-Image&quot;,&quot;local&quot;:&quot;minimax-h3-image&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Components&quot;,&quot;local&quot;:&quot;components&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;Prompt enhancement&quot;,&quot;local&quot;:&quot;prompt-enhancement&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Image size&quot;,&quot;local&quot;:&quot;image-size&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Sampling&quot;,&quot;local&quot;:&quot;sampling&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;MiniMaxH3ImageModularPipeline&quot;,&quot;local&quot;:&quot;diffusers.MiniMaxH3ImageModularPipeline&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;MiniMaxH3ImageBlocks&quot;,&quot;local&quot;:&quot;diffusers.MiniMaxH3ImageBlocks&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/><!---->
<link href="/docs/diffusers/main/en/_app/immutable/assets/0.tn0RQdqM.css" rel="modulepreload"> <!--[--><!--[0--><!--[--><!--[0--><!--[--><p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: inline-end; margin-inline-start: 10px; display: inline-flex; position: relative; z-index: 10;"><div dir="ltr" class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----></span> <span dir="auto">Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg><!----></button></div> <!--[-1--><!--]--></div><!----> <!--[0--><h1 class="relative group"><a id="minimax-h3-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="#minimax-h3-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><!---->MiniMax-H3-Image<!----></span></h1><!--]--><!----> <p>MiniMax-H3-Image é um modelo de imagem da MiniMax baseado no <a href="./minimax_h3">MiniMax-H3</a>, para geração de imagem a partir de texto e edição de imagem de referência. Você pode encontrar o checkpoint em <a href="https://huggingface.co/MiniMaxAI/MiniMax-H3-Image" rel="nofollow">MiniMaxAI/MiniMax-H3-Image</a>.</p> <p>Assim como o MiniMax-H3, ele é integrado apenas como blocos de <a href="../../modular_diffusers/overview">Modular Diffusers</a>: <a href="/docs/diffusers/main/pt/api/pipelines/minimax_h3_image#diffusers.MiniMaxH3ImageBlocks">MiniMaxH3ImageBlocks</a> e seu <a href="/docs/diffusers/main/pt/api/pipelines/minimax_h3_image#diffusers.MiniMaxH3ImageModularPipeline">MiniMaxH3ImageModularPipeline</a> constituem toda a integração.</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-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> ComponentsManager, ModularPipeline
manager = ComponentsManager()
manager.enable_auto_cpu_offload(device=<span class="hljs-string">&quot;cuda&quot;</span>)
pipe = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;MiniMaxAI/MiniMax-H3-Image&quot;</span>, components_manager=manager)
pipe.load_components(dtype=torch.bfloat16)<!----></pre></div><!----> <p>O transformer e o codificador de texto têm cerca de 62 GB cada em bfloat16 e não cabem juntos em uma GPU de 80 GB, portanto, os exemplos usam auto offloading: o <a href="/docs/diffusers/main/pt/api/modular_diffusers/pipeline_components#diffusers.ComponentsManager">ComponentsManager</a> move cada componente para o acelerador quando um bloco precisa dele. Uma imagem de 4736x3552 então atinge um pico de cerca de 65 GB.</p> <p>Texto para 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-py "><!---->prompt = (
<span class="hljs-string">&quot;An overhead photograph of a rustic farm-to-table spread on a weathered wooden table: a round loaf of sourdough &quot;</span>
<span class="hljs-string">&quot;on burlap, wedges of cheese on a dark board, a copper bowl of vegetables, vine tomatoes, olives and walnuts.&quot;</span>
)
image = pipe(
prompt=prompt, height=<span class="hljs-number">3552</span>, width=<span class="hljs-number">4736</span>, generator=torch.Generator().manual_seed(<span class="hljs-number">42</span>), output=<span class="hljs-string">&quot;images&quot;</span>
)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <p>Editando uma imagem de referência:</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-py "><!----><span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> load_image
reference = load_image(<span class="hljs-string">&quot;https://huggingface.co/MiniMaxAI/MiniMax-H3-Image/resolve/main/examples/reference1.png&quot;</span>)
image = pipe(
prompt=<span class="hljs-string">&quot;Keep everything in &lt;Picture 1&gt; as it is, but remove the headphones and have the two hosts fist bump.&quot;</span>,
image=reference,
height=<span class="hljs-number">4096</span>,
width=<span class="hljs-number">4096</span>,
generator=torch.Generator().manual_seed(<span class="hljs-number">1234</span>),
output=<span class="hljs-string">&quot;images&quot;</span>,
)[<span class="hljs-number">0</span>]<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="components" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#components"><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><!---->Componentes<!----></span></h2><!--]--><!----> <table><thead><tr><th>Componente</th><th>Classe</th></tr></thead><tbody><tr><td><code>transformer</code></td><td><a href="/docs/diffusers/main/pt/api/models/minimax_h3_image_transformer2d#diffusers.MiniMaxH3ImageTransformer2DModel">MiniMaxH3ImageTransformer2DModel</a></td></tr><tr><td><code>text_encoder</code></td><td><code>Qwen3VLForConditionalGeneration</code></td></tr><tr><td><code>vae</code></td><td><a href="/docs/diffusers/main/pt/api/models/autoencoderkl_minimax_h3_image#diffusers.AutoencoderKLMiniMaxH3Image">AutoencoderKLMiniMaxH3Image</a></td></tr><tr><td><code>tokenizer</code>, <code>processor</code></td><td>O tokenizer do codificador de texto e um <code>Qwen3VLProcessor</code></td></tr><tr><td><code>scheduler</code></td><td><a href="/docs/diffusers/main/pt/api/schedulers/minimax_h3#diffusers.MiniMaxH3Scheduler">MiniMaxH3Scheduler</a></td></tr></tbody></table> <p>O codificador de texto é o mesmo utilizado pelo <a href="./minimax_h3">MiniMax-H3</a>. O checkpoint mantém uma cópia própria para que o repositório seja autocontido; o Hub deduplica arquivos idênticos entre repositórios, portanto a cópia não ocupa armazenamento extra lá.</p> <!--[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 workflow é selecionado na entrada <code>image</code>:</p> <ul><li><strong><code>text2image</code></strong> gera a partir de <code>prompt</code>.</li> <li><strong><code>image_conditioned</code></strong> gera a partir de <code>prompt</code> e uma ou mais imagens de referência passadas como <code>image</code>. O prompt se refere a elas em ordem como <code>&lt;Picture 1></code>, <code>&lt;Picture 2></code>, e assim por diante. As referências mantêm seu próprio tamanho: cada uma é redimensionada apenas quando sua borda curta está fora de 1024 a 2048 pixels, e nunca para o tamanho da imagem gerada.</li></ul> <p>O prompt é usado literalmente, sem template de chat, e descrições longas e completas funcionam melhor. Um prompt de edição geralmente nomeia o que preservar, depois o que alterar e, em seguida, descreve a imagem de destino por completo.</p> <!--[1--><h2 class="relative group"><a id="prompt-enhancement" 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-enhancement"><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><!---->Melhoria de prompt<!----></span></h2><!--]--><!----> <p><a href="https://huggingface.co/MiniMaxAI/MiniMax-H3-Image-PE" rel="nofollow">MiniMax-H3-Image-PE</a> é o modelo de expansão de prompt da MiniMax: ele reescreve uma solicitação curta no tipo de descrição detalhada com a qual o modelo de imagem funciona melhor e recomenda uma proporção de aspecto. Os blocos personalizados <code>diffusers/minimax-h3-image-pe-blocks</code> o executam como um pipeline próprio. Eles precisam de <code>xgrammar</code> e de cerca de 18 GB de memória de GPU.</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-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> ModularPipeline
enhancer = ModularPipeline.from_pretrained(<span class="hljs-string">&quot;diffusers/minimax-h3-image-pe-blocks&quot;</span>, trust_remote_code=<span class="hljs-literal">True</span>)
enhancer.load_components(dtype=torch.bfloat16, device_map=<span class="hljs-string">&quot;cuda&quot;</span>)
state = enhancer(prompt=<span class="hljs-string">&quot;a cozy cabin in the snow&quot;</span>)
prompt, aspect_ratio = state.get(<span class="hljs-string">&quot;prompt&quot;</span>), state.get(<span class="hljs-string">&quot;aspect_ratio&quot;</span>)<!----></pre></div><!----> <p>Passe as imagens de referência como <code>image</code> para expandir uma instrução de edição. Em seguida, passe <code>prompt</code> para o pipeline de imagem, com <code>height</code> e <code>width</code> que correspondam a <code>aspect_ratio</code>.</p> <!--[1--><h2 class="relative group"><a id="image-size" 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-size"><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><!---->Tamanho da imagem<!----></span></h2><!--]--><!----> <p><code>height</code> e <code>width</code> são o tamanho da imagem retornada e, por padrão, são 2048. O modelo gera em 4 megapixels ou mais: uma solicitação menor é gerada em um canvas maior com a mesma proporção de aspecto e redimensionada de volta, portanto, uma solicitação de 1024x1024 custa o mesmo que uma de 2048x2048. <code>min_generation_area</code> define esse limite em pixels; passe <code>1024</code> para sempre gerar no tamanho solicitado. De qualquer forma, os lados do canvas são arredondados para múltiplos de 32.</p> <p>A MiniMax recomenda 4096x4096 para 1:1, 4736x3552 para 4:3, 5024x3360 para 3:2, 5472x3072 para 16:9 e 6272x2688 para 21:9, ou os mesmos tamanhos com altura e largura invertidos para formatos retrato.</p> <!--[1--><h2 class="relative group"><a id="sampling" 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="#sampling"><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><!---->Amostragem<!----></span></h2><!--]--><!----> <ul><li>O checkpoint amostra em 6 etapas sobre seu <code>sample_sigmas</code>, e o transformer possui uma cabeça de saída por etapa, portanto não há <code>num_inference_steps</code>.</li> <li>O modelo é destilado por orientação, portanto não há um guia, <code>negative_prompt</code> ou <code>guidance_scale</code>.</li></ul> <!--[1--><h2 class="relative group"><a id="diffusers.MiniMaxH3ImageModularPipeline" 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.MiniMaxH3ImageModularPipeline"><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><!---->MiniMaxH3ImageModularPipeline<!----></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.MiniMaxH3ImageModularPipeline"><!----><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">MiniMaxH3ImageModularPipeline</span></span></h3><!----> <a id="diffusers.MiniMaxH3ImageModularPipeline" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.MiniMaxH3ImageModularPipeline"><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/minimax_h3_image/modular_pipeline.py#L27" 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 MiniMax-H3-Image: texto para imagem (<code>text2image</code>) e edição de imagem de referência
(<code>image_conditioned</code>).</p> <p>MiniMax-H3-Image executa o transformer MiniMax-H3 sem áudio sobre uma sequência empacotada de linhas de texto, as linhas das imagens de referência e as linhas da imagem gerada, que é um único frame latente. O checkpoint é destilado duas vezes: o guidance é incorporado nos pesos, portanto não há guider, nem <code>negative_prompt</code> e nem <code>guidance_scale</code>, e ele amostra em uma grade fixa de seis passos com uma cabeça de saída do transformer por passo, portanto também não há <code>num_inference_steps</code>.</p> <div class="relative group rounded-md"><a id="diffusers.MiniMaxH3ImageModularPipeline.example" class="header-link block pr-0.5 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.MiniMaxH3ImageModularPipeline.example"><span><svg class="text-smd" 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> <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-py "><!----><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;MiniMaxAI/MiniMax-H3-Image&quot;</span>)
pipe.load_components(dtype=torch.bfloat16)<!----></pre></div><!----></div><!----></div> <!--[1--><h2 class="relative group"><a id="diffusers.MiniMaxH3ImageBlocks" 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.MiniMaxH3ImageBlocks"><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><!---->MiniMaxH3ImageBlocks<!----></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.MiniMaxH3ImageBlocks"><!----><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">MiniMaxH3ImageBlocks</span></span></h3><!----> <a id="diffusers.MiniMaxH3ImageBlocks" class="header-link invisible with-hover:group-hover:visible pr-2" href="#diffusers.MiniMaxH3ImageBlocks"><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/minimax_h3_image/modular_blocks_minimax_h3_image.py#L371" 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>Blocos de pipeline modulares para MiniMax-H3-Image: <code>text2image</code> e edição de imagem de referência (<code>image_conditioned</code>) quando <code>image</code> é passado.</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:
image_processor (<code>VaeImageProcessor</code>) text_encoder (<code>Qwen3VLForConditionalGeneration</code>) tokenizer
(<code>Qwen2Tokenizer</code>) processor (<code>Qwen3VLProcessor</code>) vae (<code>AutoencoderKLMiniMaxH3Image</code>) scheduler
(<code>MiniMaxH3Scheduler</code>) transformer (<code>MiniMaxH3ImageTransformer2DModel</code>)</p> <p>Configs:
reference_min_short_edge (padrão: 1024) reference_max_short_edge (padrão: 2048) vision_min_short_edge
(padrão: 1024) vision_max_short_edge (padrão: 1024) sample_sigmas (padrão: [1.0, 0.9675236940383911,
0.8875485062599182, 0.75, 0.5327739119529724, 0.23200887441635132, 0.0]): A grade de amostragem na qual as
cabeças de saída por etapa do checkpoint foram destiladas.</p> <p>Entradas:
image (<code>Image | list</code>, <em>opcional</em>):
A imagem de referência da edição, ou uma lista delas. O prompt se refere a elas em ordem como <code>&lt;Picture 1></code>, <code>&lt;Picture 2></code>, …
prompt (<code>str</code>):
O prompt para orientar a geração, uma única string.
vision_images (<code>list</code>, <em>opcional</em>):
As imagens de referência redimensionadas para o codificador de texto.
reference_images (<code>list</code>, <em>opcional</em>):
As imagens de referência redimensionadas para o VAE.
generator (<code>Generator</code>, <em>opcional</em>):
O gerador da solicitação. Uma amostra posterior por referência é extraída dele.
height (<code>int</code>):
Altura da imagem retornada em pixels.
width (<code>int</code>):
Largura da imagem retornada em pixels.
min_generation_area (<code>int</code>, <em>opcional</em>, padrão 4194304):
A menor área, em pixels, na qual uma imagem é gerada. Um <code>height x width</code> menor é gerado em um
canvas ampliado para essa área e redimensionado de volta para <code>height x width</code> após a decodificação. Passe 1024 para sempre
gerar no tamanho solicitado.
reference_latents (<code>list</code>, <em>opcional</em>):
Os latentes de condicionamento das imagens de referência. Sua forma é de onde a geometria de cada referência vem.
latents (<code>Tensor</code>):
Ruído pré-gerado de forma <code>(1, latent_channels, latent_height, latent_width)</code>, usado no lugar da
extração.
reference_noise (<code>list</code>, <em>opcional</em>):
Ruído pré-gerado para as imagens de referência, um tensor da forma de cada entrada de <code>reference_latents</code>, usado no lugar das extrações.
*<em>denoiser_input_fields (<code>None</code>,</em>opcional<em>):
A descrição estrutural da sequência empacotada que o transformer lê por nome: <code>token_tags</code>, <code>position_ids</code> e os dois tensores de índice de linha.
output_type (<code>str</code>,</em>opcional*, padrão pil):
Formato de saída: ‘pil’, ‘np’, ‘pt’.</p> <p>Saídas:
imagens (<code>list</code>):
Imagens geradas.</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/minimax_h3_image.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, 9],
data: [null,null],
form: null,
error: null
});
});
}
</script>

Xet Storage Details

Size:
43 kB
·
Xet hash:
55d75c12c2f1fb0eca3370fdb3b7922cce890935e5fed878525188f6ab1bf61f

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