Hugging Face's logo

Buckets:
hf-doc-build
/
doc

Files
xet
hf-doc-build/doc / diffusers /v0.43.0 /pt /quantization /overview.html
mishig's picture
mishig
download
raw
34.8 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Começando&quot;,&quot;local&quot;:&quot;getting-started&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Quantização em nível de pipeline&quot;,&quot;local&quot;:&quot;pipeline-level-quantization&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Quantização básica&quot;,&quot;local&quot;:&quot;basic-quantization&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Quantização avançada&quot;,&quot;local&quot;:&quot;advanced-quantization&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Carregando checkpoints pré-quantizados&quot;,&quot;local&quot;:&quot;loading-prequantized-checkpoints&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Recursos&quot;,&quot;local&quot;:&quot;resources&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/>
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/entry/start.1gaXe03k.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/BObs6oQO.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/DPnyVmXs.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/entry/app.BwdYRbCn.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/DSGUT_OW.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/C045MWWx.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/6tuJ3nyi.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/DsnmJJEf.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/nodes/0.CornNeQR.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/IaDvMxLQ.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/nodes/312._S_peb4w.js" rel="modulepreload">
<link href="/docs/diffusers/v0.43.0/en/_app/immutable/chunks/bKdwVYpU.js" rel="modulepreload">
<!--cubdu--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Getting started&quot;,&quot;local&quot;:&quot;getting-started&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Pipeline-level quantization&quot;,&quot;local&quot;:&quot;pipeline-level-quantization&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Basic quantization&quot;,&quot;local&quot;:&quot;basic-quantization&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Advanced quantization&quot;,&quot;local&quot;:&quot;advanced-quantization&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Loading prequantized checkpoints&quot;,&quot;local&quot;:&quot;loading-prequantized-checkpoints&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Resources&quot;,&quot;local&quot;:&quot;resources&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/><!---->
<link href="/docs/diffusers/v0.43.0/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="getting-started" 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="#getting-started"><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><!---->Começando<!----></span></h1><!--]--><!----> <p>A quantização foca em representar dados com menos bits, ao mesmo tempo em que tenta preservar a precisão dos dados originais. Isso geralmente significa converter um tipo de dados para representar a mesma informação com menos bits. Por exemplo, se os pesos do seu modelo forem armazenados como números de ponto flutuante de 32 bits e forem quantizados para números de ponto flutuante de 16 bits, o tamanho do modelo é reduzido pela metade, o que facilita o armazenamento e diminui o uso de memória. A menor precisão também pode acelerar a inferência, pois leva menos tempo para realizar cálculos com menos bits.</p> <p>O Diffusers suporta vários backends de quantização para tornar grandes modelos de difusão, como o <a href="../api/pipelines/flux">Flux</a>, mais acessíveis. Este guia mostra como usar a classe <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.PipelineQuantizationConfig">PipelineQuantizationConfig</a> para quantizar um pipeline durante sua inicialização a partir de um checkpoint pré-treinado ou não quantizado.</p> <!--[1--><h2 class="relative group"><a id="pipeline-level-quantization" 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="#pipeline-level-quantization"><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><!---->Quantização em nível de pipeline<!----></span></h2><!--]--><!----> <p>Existem duas formas de usar <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.PipelineQuantizationConfig">PipelineQuantizationConfig</a>, dependendo do nível de personalização que você deseja aplicar à configuração de quantização.</p> <ul><li>para casos de uso básicos, defina os argumentos <code>quant_backend</code>, <code>quant_kwargs</code> e <code>components_to_quantize</code></li> <li>para controle granular de quantização, defina um <code>quant_mapping</code> que fornece a configuração de quantização para componentes individuais do modelo</li></ul> <!--[2--><h3 class="relative group"><a id="basic-quantization" 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="#basic-quantization"><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><!---->Quantização básica<!----></span></h3><!--]--><!----> <p>Inicialize <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.PipelineQuantizationConfig">PipelineQuantizationConfig</a> com os seguintes parâmetros.</p> <ul><li><code>quant_backend</code> especifica qual backend de quantização usar. Os backends atualmente suportados incluem: <code>bitsandbytes_4bit</code>, <code>bitsandbytes_8bit</code>, <code>gguf</code>, <code>quanto</code> e <code>torchao</code>.</li> <li><code>quant_kwargs</code> especifica os argumentos de quantização a serem usados.</li></ul> <blockquote class="tip"><p>Esses argumentos de <code>quant_kwargs</code> são diferentes para cada backend. Consulte a documentação da <a href="../api/quantization">Quantization API</a> para ver os argumentos de cada backend.</p></blockquote> <ul><li><p><code>components_to_quantize</code> especifica qual(is) componente(s) do pipeline devem ser quantizados. Normalmente, você deve quantizar os componentes mais intensivos em computação, como o transformer. O codificador de texto é outro componente a ser considerado para quantização se o pipeline tiver mais de um, como no caso de <a href="/docs/diffusers/v0.43.0/pt/api/pipelines/flux#diffusers.FluxPipeline">FluxPipeline</a>. O exemplo abaixo quantiza o codificador de texto T5 em <a href="/docs/diffusers/v0.43.0/pt/api/pipelines/flux#diffusers.FluxPipeline">FluxPipeline</a>, mantendo o modelo CLIP intacto.</p> <p><code>components_to_quantize</code> aceita uma lista para vários modelos ou uma string para um único modelo.</p></li></ul> <p>O exemplo abaixo carrega o backend bitsandbytes com os seguintes argumentos de <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.BitsAndBytesConfig">BitsAndBytesConfig</a>, <code>load_in_4bit</code>, <code>bnb_4bit_quant_type</code> e <code>bnb_4bit_compute_dtype</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-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline
<span class="hljs-keyword">from</span> diffusers.quantizers <span class="hljs-keyword">import</span> PipelineQuantizationConfig
pipeline_quant_config = PipelineQuantizationConfig(
quant_backend=<span class="hljs-string">&quot;bitsandbytes_4bit&quot;</span>,
quant_kwargs={<span class="hljs-string">&quot;load_in_4bit&quot;</span>: <span class="hljs-literal">True</span>, <span class="hljs-string">&quot;bnb_4bit_quant_type&quot;</span>: <span class="hljs-string">&quot;nf4&quot;</span>, <span class="hljs-string">&quot;bnb_4bit_compute_dtype&quot;</span>: torch.bfloat16},
components_to_quantize=[<span class="hljs-string">&quot;transformer&quot;</span>, <span class="hljs-string">&quot;text_encoder_2&quot;</span>],
)<!----></pre></div><!----> <p>Passe o <code>pipeline_quant_config</code> para <a href="/docs/diffusers/v0.43.0/pt/api/pipelines/overview#diffusers.DiffusionPipeline.from_pretrained">from_pretrained()</a> para quantizar o pipeline.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg 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 "><!---->pipe = DiffusionPipeline.from_pretrained(
<span class="hljs-string">&quot;black-forest-labs/FLUX.1-dev&quot;</span>,
quantization_config=pipeline_quant_config,
dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>) <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>
image = pipe(<span class="hljs-string">&quot;photo of a cute dog&quot;</span>).images[<span class="hljs-number">0</span>]<!----></pre></div><!----> <!--[2--><h3 class="relative group"><a id="advanced-quantization" 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="#advanced-quantization"><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><!---->Quantização avançada<!----></span></h3><!--]--><!----> <p>O argumento <code>quant_mapping</code> oferece mais opções sobre como quantizar cada componente individual em um pipeline, como combinar diferentes backends de quantização.</p> <p>Inicialize <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.PipelineQuantizationConfig">PipelineQuantizationConfig</a> e passe um <code>quant_mapping</code> para ele. O <code>quant_mapping</code> permite especificar as opções de quantização para cada componente do pipeline, como o transformer e o text encoder.</p> <p>O exemplo abaixo usa dois backends de quantização, <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.QuantoConfig">QuantoConfig</a> e <a href="https://huggingface.co/docs/transformers/v5.22.2/en/main_classes/quantization#transformers.BitsAndBytesConfig" rel="nofollow">transformers.BitsAndBytesConfig</a>, para o transformer e o codificador 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-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> BitsAndBytesConfig <span class="hljs-keyword">as</span> DiffusersBitsAndBytesConfig
<span class="hljs-keyword">from</span> diffusers.quantizers.quantization_config <span class="hljs-keyword">import</span> QuantoConfig
<span class="hljs-keyword">from</span> diffusers.quantizers <span class="hljs-keyword">import</span> PipelineQuantizationConfig
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> BitsAndBytesConfig <span class="hljs-keyword">as</span> TransformersBitsAndBytesConfig
pipeline_quant_config = PipelineQuantizationConfig(
quant_mapping={
<span class="hljs-string">&quot;transformer&quot;</span>: QuantoConfig(weights_dtype=<span class="hljs-string">&quot;int8&quot;</span>),
<span class="hljs-string">&quot;text_encoder_2&quot;</span>: TransformersBitsAndBytesConfig(
load_in_4bit=<span class="hljs-literal">True</span>, compute_dtype=torch.bfloat16
),
}
)<!----></pre></div><!----> <p>Existe um backend bitsandbytes separado no <a href="https://huggingface.co/docs/transformers/main_classes/quantization#transformers.BitsAndBytesConfig" rel="nofollow">Transformers</a>. Você precisa importar e usar <a href="https://huggingface.co/docs/transformers/v5.22.2/en/main_classes/quantization#transformers.BitsAndBytesConfig" rel="nofollow">transformers.BitsAndBytesConfig</a> para componentes que vêm do Transformers. Por exemplo, <code>text_encoder_2</code> em <a href="/docs/diffusers/v0.43.0/pt/api/pipelines/flux#diffusers.FluxPipeline">FluxPipeline</a> é um <a href="https://huggingface.co/docs/transformers/v5.22.2/en/model_doc/t5#transformers.T5EncoderModel" rel="nofollow">T5EncoderModel</a> do Transformers, então você precisa usar <a href="https://huggingface.co/docs/transformers/v5.22.2/en/main_classes/quantization#transformers.BitsAndBytesConfig" rel="nofollow">transformers.BitsAndBytesConfig</a> em vez de <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.BitsAndBytesConfig">diffusers.BitsAndBytesConfig</a>.</p> <blockquote class="tip"><p>Use o método de <a href="#basic-quantization">quantização básica</a> acima se você não quiser gerenciar esses imports distintos ou não tiver certeza de onde cada componente do pipeline vem.</p></blockquote> <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> DiffusionPipeline
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> BitsAndBytesConfig <span class="hljs-keyword">as</span> DiffusersBitsAndBytesConfig
<span class="hljs-keyword">from</span> diffusers.quantizers <span class="hljs-keyword">import</span> PipelineQuantizationConfig
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> BitsAndBytesConfig <span class="hljs-keyword">as</span> TransformersBitsAndBytesConfig
pipeline_quant_config = PipelineQuantizationConfig(
quant_mapping={
<span class="hljs-string">&quot;transformer&quot;</span>: DiffusersBitsAndBytesConfig(load_in_4bit=<span class="hljs-literal">True</span>, bnb_4bit_compute_dtype=torch.bfloat16),
<span class="hljs-string">&quot;text_encoder_2&quot;</span>: TransformersBitsAndBytesConfig(
load_in_4bit=<span class="hljs-literal">True</span>, compute_dtype=torch.bfloat16
),
}
)<!----></pre></div><!----> <p>Passe o <code>pipeline_quant_config</code> para <a href="/docs/diffusers/v0.43.0/pt/api/pipelines/overview#diffusers.DiffusionPipeline.from_pretrained">from_pretrained()</a> para quantizar o pipeline.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg 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 "><!---->pipe = DiffusionPipeline.from_pretrained(
<span class="hljs-string">&quot;black-forest-labs/FLUX.1-dev&quot;</span>,
quantization_config=pipeline_quant_config,
dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>) <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>
image = pipe(<span class="hljs-string">&quot;photo of a cute dog&quot;</span>).images[<span class="hljs-number">0</span>]<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="loading-prequantized-checkpoints" 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="#loading-prequantized-checkpoints"><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><!---->Carregando checkpoints pré-quantizados<!----></span></h2><!--]--><!----> <p>Alguns backends de quantização suportam o carregamento de checkpoints que já foram quantizados e enviados ao Hub. Nesse caso, a configuração de quantização é armazenada no <code>config.json</code> do modelo e é lida automaticamente — você não precisa criar um <a href="/docs/diffusers/v0.43.0/pt/api/quantization#diffusers.PipelineQuantizationConfig">PipelineQuantizationConfig</a> nem passar um argumento <code>quantization_config</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-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline
pipe = DiffusionPipeline.from_pretrained(
<span class="hljs-string">&quot;rootonchair/ERNIE-Image-Turbo-nunchaku-lite-nvfp4&quot;</span>,
dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>) <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span><!----></pre></div><!----> <p>Os seguintes backends suportam o carregamento de checkpoints pré-quantizados nativamente.</p> <table><thead><tr><th>Backend</th><th>Observações</th></tr></thead><tbody><tr><td><a href="./bitsandbytes">bitsandbytes</a></td><td>Config é salvo em <code>config.json</code>; nenhum argumento extra é necessário. Suporta quantização em tempo de execução, bem como o carregamento de checkpoints pré-quantizados.</td></tr><tr><td><a href="./torchao">torchao</a></td><td>Igual ao acima.</td></tr><tr><td><a href="./gguf">GGUF</a></td><td>Usa <code>from_single_file</code> com classes de Model; o carregamento em nível de pipeline não é suportado.</td></tr><tr><td><a href="./bfl">BFL</a></td><td>Carrega checkpoints de arquivo único <code>fp8r</code> e <code>nvfp4</code> da Black Forest Labs com <code>from_single_file</code>; o carregamento em nível de pipeline não é suportado.</td></tr><tr><td><a href="./autoround">AutoRound</a></td><td>Apenas o carregamento é suportado; quantize primeiro com a CLI ou a API Python do AutoRound.</td></tr><tr><td><a href="./nunchaku">Nunchaku Lite</a></td><td>Config é salvo em <code>config.json</code>; requer o pacote <code>kernels</code>. Apenas o carregamento é suportado.</td></tr><tr><td><a href="./modelopt">ModelOpt</a></td><td>Apenas o carregamento de modelos pré-quantizados é suportado.</td></tr><tr><td><a href="./sdnq">SDNQ</a></td><td>Suporta tanto a quantização em tempo de execução quanto o carregamento de modelos pré-quantizados; requer o pacote <code>sdnq</code>.</td></tr></tbody></table> <!--[1--><h2 class="relative group"><a id="resources" 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="#resources"><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><!---->Recursos<!----></span></h2><!--]--><!----> <p>Confira os recursos abaixo para saber mais sobre quantização.</p> <ul><li><p>Se você é novo em quantização, recomendamos conferir os seguintes cursos introdutórios, desenvolvidos em colaboração com a DeepLearning.AI.</p> <ul><li><a href="https://www.deeplearning.ai/short-courses/quantization-fundamentals-with-hugging-face/" rel="nofollow">Fundamentos de Quantização com Hugging Face</a></li> <li><a href="https://www.deeplearning.ai/short-courses/quantization-in-depth/" rel="nofollow">Quantização em profundidade</a></li></ul></li> <li><p>Consulte o <a href="https://huggingface.co/docs/transformers/main/en/quantization/contribute" rel="nofollow">Guia para contribuir com um novo método de quantização</a> se você estiver interessado em adicionar um novo método de quantização.</p></li> <li><p>A <a href="https://huggingface.co/docs/transformers/quantization/overview#when-to-use-what" rel="nofollow">Visão geral</a> da quantização do Transformers fornece uma visão geral das vantagens e desvantagens dos diferentes backends de quantização.</p></li> <li><p>Leia o post no blog <a href="https://huggingface.co/blog/diffusers-quantization" rel="nofollow">Exploring Quantization Backends in Diffusers</a> para uma breve introdução a cada backend de quantização, como escolher um backend e como combinar a quantização com outras otimizações de memória.</p></li></ul> <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/quantization/overview.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_ywq9g6 = {
base: "/docs/diffusers/v0.43.0/en",
assets: "/docs/diffusers/v0.43.0/en"
};
const element = document.currentScript.parentElement;
Promise.all([
import("/docs/diffusers/v0.43.0/en/_app/immutable/entry/start.1gaXe03k.js"),
import("/docs/diffusers/v0.43.0/en/_app/immutable/entry/app.BwdYRbCn.js")
]).then(([kit, app]) => {
kit.start(app, element, {
node_ids: [0, 312],
data: [null,null],
form: null,
error: null
});
});
}
</script>

Xet Storage Details

Size:
34.8 kB
·
Xet hash:
1ba5f2333db4309f64376acdbbadfed5c0ac672caba30fa209a1484b349ce383

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