Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Quanto","local":"quanto","sections":[{"title":"Skipping Quantization on specific modules","local":"skipping-quantization-on-specific-modules","sections":[],"depth":2},{"title":"Using from_single_file with the Quanto Backend","local":"using-fromsinglefile-with-the-quanto-backend","sections":[],"depth":2},{"title":"Saving Quantized models","local":"saving-quantized-models","sections":[],"depth":2},{"title":"Using torch.compile with Quanto","local":"using-torchcompile-with-quanto","sections":[],"depth":2},{"title":"Supported Quantization Types","local":"supported-quantization-types","sections":[{"title":"Weights","local":"weights","sections":[],"depth":3}],"depth":2}],"depth":1}"> | |
| <link href="/docs/diffusers/pr_12448/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/entry/start.77018737.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/scheduler.53228c21.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/singletons.5ccd8670.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/index.e93d0901.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/paths.e3119367.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/entry/app.30e12370.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/preload-helper.7a497b25.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/index.100fac89.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/nodes/0.62116836.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/each.e59479a4.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/nodes/296.9fd54b92.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/CopyLLMTxtMenu.88008e00.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/globals.7f7f1b26.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/IconCopy.38cf8f56.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.afa087fa.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/pr_12448/en/_app/immutable/chunks/CodeBlock.d30a6509.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Quanto","local":"quanto","sections":[{"title":"Skipping Quantization on specific modules","local":"skipping-quantization-on-specific-modules","sections":[],"depth":2},{"title":"Using from_single_file with the Quanto Backend","local":"using-fromsinglefile-with-the-quanto-backend","sections":[],"depth":2},{"title":"Saving Quantized models","local":"saving-quantized-models","sections":[],"depth":2},{"title":"Using torch.compile with Quanto","local":"using-torchcompile-with-quanto","sections":[],"depth":2},{"title":"Supported Quantization Types","local":"supported-quantization-types","sections":[{"title":"Weights","local":"weights","sections":[],"depth":3}],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"><div class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 max-sm:gap-0.5 h-6 max-sm:h-5 px-2 max-sm:px-1.5 text-[11px] max-sm:text-[9px] 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"><svg class="w-3 h-3 max-sm:w-2.5 max-sm:h-2.5" 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>Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-6 max-sm:h-5 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 w-3 h-3 max-sm:w-2.5 max-sm:h-2.5 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> </div> <h1 class="relative group"><a id="quanto" 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="#quanto"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Quanto</span></h1> <p data-svelte-h="svelte-wedqkd"><a href="https://github.com/huggingface/optimum-quanto" rel="nofollow">Quanto</a> is a PyTorch quantization backend for <a href="https://huggingface.co/docs/optimum/en/index" rel="nofollow">Optimum</a>. It has been designed with versatility and simplicity in mind:</p> <ul data-svelte-h="svelte-yhikdz"><li>All features are available in eager mode (works with non-traceable models)</li> <li>Supports quantization aware training</li> <li>Quantized models are compatible with <code>torch.compile</code></li> <li>Quantized models are Device agnostic (e.g CUDA,XPU,MPS,CPU)</li></ul> <p data-svelte-h="svelte-18uawxu">In order to use the Quanto backend, you will first need to install <code>optimum-quanto>=0.2.6</code> and <code>accelerate</code></p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->pip install optimum-quanto accelerate<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-il6k3t">Now you can quantize a model by passing the <code>QuantoConfig</code> object to the <code>from_pretrained()</code> method. Although the Quanto library does allow quantizing <code>nn.Conv2d</code> and <code>nn.LayerNorm</code> modules, currently, Diffusers only supports quantizing the weights in the <code>nn.Linear</code> layers of a model. The following snippet demonstrates how to apply <code>float8</code> quantization with Quanto.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxTransformer2DModel, QuantoConfig | |
| model_id = <span class="hljs-string">"black-forest-labs/FLUX.1-dev"</span> | |
| quantization_config = QuantoConfig(weights_dtype=<span class="hljs-string">"float8"</span>) | |
| transformer = FluxTransformer2DModel.from_pretrained( | |
| model_id, | |
| subfolder=<span class="hljs-string">"transformer"</span>, | |
| quantization_config=quantization_config, | |
| torch_dtype=torch.bfloat16, | |
| ) | |
| pipe = FluxPipeline.from_pretrained(model_id, transformer=transformer, torch_dtype=torch_dtype) | |
| pipe.to(<span class="hljs-string">"cuda"</span>) | |
| prompt = <span class="hljs-string">"A cat holding a sign that says hello world"</span> | |
| image = pipe( | |
| prompt, num_inference_steps=<span class="hljs-number">50</span>, guidance_scale=<span class="hljs-number">4.5</span>, max_sequence_length=<span class="hljs-number">512</span> | |
| ).images[<span class="hljs-number">0</span>] | |
| image.save(<span class="hljs-string">"output.png"</span>)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="skipping-quantization-on-specific-modules" 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="#skipping-quantization-on-specific-modules"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Skipping Quantization on specific modules</span></h2> <p data-svelte-h="svelte-4moh48">It is possible to skip applying quantization on certain modules using the <code>modules_to_not_convert</code> argument in the <code>QuantoConfig</code>. Please ensure that the modules passed in to this argument match the keys of the modules in the <code>state_dict</code></p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxTransformer2DModel, QuantoConfig | |
| model_id = <span class="hljs-string">"black-forest-labs/FLUX.1-dev"</span> | |
| quantization_config = QuantoConfig(weights_dtype=<span class="hljs-string">"float8"</span>, modules_to_not_convert=[<span class="hljs-string">"proj_out"</span>]) | |
| transformer = FluxTransformer2DModel.from_pretrained( | |
| model_id, | |
| subfolder=<span class="hljs-string">"transformer"</span>, | |
| quantization_config=quantization_config, | |
| torch_dtype=torch.bfloat16, | |
| )<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="using-fromsinglefile-with-the-quanto-backend" 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="#using-fromsinglefile-with-the-quanto-backend"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Using from_single_file with the Quanto Backend</span></h2> <p data-svelte-h="svelte-80xlz7"><code>QuantoConfig</code> is compatible with <code>~FromOriginalModelMixin.from_single_file</code>.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxTransformer2DModel, QuantoConfig | |
| ckpt_path = <span class="hljs-string">"https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/flux1-dev.safetensors"</span> | |
| quantization_config = QuantoConfig(weights_dtype=<span class="hljs-string">"float8"</span>) | |
| transformer = FluxTransformer2DModel.from_single_file(ckpt_path, quantization_config=quantization_config, torch_dtype=torch.bfloat16)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="saving-quantized-models" 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="#saving-quantized-models"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Saving Quantized models</span></h2> <p data-svelte-h="svelte-1bdnm93">Diffusers supports serializing Quanto models using the <code>~ModelMixin.save_pretrained</code> method.</p> <p data-svelte-h="svelte-18ura5r">The serialization and loading requirements are different for models quantized directly with the Quanto library and models quantized | |
| with Diffusers using Quanto as the backend. It is currently not possible to load models quantized directly with Quanto into Diffusers using <code>~ModelMixin.from_pretrained</code></p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxTransformer2DModel, QuantoConfig | |
| model_id = <span class="hljs-string">"black-forest-labs/FLUX.1-dev"</span> | |
| quantization_config = QuantoConfig(weights_dtype=<span class="hljs-string">"float8"</span>) | |
| transformer = FluxTransformer2DModel.from_pretrained( | |
| model_id, | |
| subfolder=<span class="hljs-string">"transformer"</span>, | |
| quantization_config=quantization_config, | |
| torch_dtype=torch.bfloat16, | |
| ) | |
| <span class="hljs-comment"># save quantized model to reuse</span> | |
| transformer.save_pretrained(<span class="hljs-string">"<your quantized model save path>"</span>) | |
| <span class="hljs-comment"># you can reload your quantized model with</span> | |
| model = FluxTransformer2DModel.from_pretrained(<span class="hljs-string">"<your quantized model save path>"</span>)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="using-torchcompile-with-quanto" 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="#using-torchcompile-with-quanto"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Using torch.compile with Quanto</span></h2> <p data-svelte-h="svelte-1sftzs6">Currently the Quanto backend supports <code>torch.compile</code> for the following quantization types:</p> <ul data-svelte-h="svelte-4fwo66"><li><code>int8</code> weights</li></ul> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxPipeline, FluxTransformer2DModel, QuantoConfig | |
| model_id = <span class="hljs-string">"black-forest-labs/FLUX.1-dev"</span> | |
| quantization_config = QuantoConfig(weights_dtype=<span class="hljs-string">"int8"</span>) | |
| transformer = FluxTransformer2DModel.from_pretrained( | |
| model_id, | |
| subfolder=<span class="hljs-string">"transformer"</span>, | |
| quantization_config=quantization_config, | |
| torch_dtype=torch.bfloat16, | |
| ) | |
| transformer = torch.<span class="hljs-built_in">compile</span>(transformer, mode=<span class="hljs-string">"max-autotune"</span>, fullgraph=<span class="hljs-literal">True</span>) | |
| pipe = FluxPipeline.from_pretrained( | |
| model_id, transformer=transformer, torch_dtype=torch_dtype | |
| ) | |
| pipe.to(<span class="hljs-string">"cuda"</span>) | |
| images = pipe(<span class="hljs-string">"A cat holding a sign that says hello"</span>).images[<span class="hljs-number">0</span>] | |
| images.save(<span class="hljs-string">"flux-quanto-compile.png"</span>)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="supported-quantization-types" 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="#supported-quantization-types"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Supported Quantization Types</span></h2> <h3 class="relative group"><a id="weights" 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="#weights"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Weights</span></h3> <ul data-svelte-h="svelte-gqdhfh"><li>float8</li> <li>int8</li> <li>int4</li> <li>int2</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/quanto.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 data-svelte-h="svelte-zjs2n5"><span class="underline">Update</span> on GitHub</span></a> <p></p> | |
| <script> | |
| { | |
| __sveltekit_n9i9q2 = { | |
| assets: "/docs/diffusers/pr_12448/en", | |
| base: "/docs/diffusers/pr_12448/en", | |
| env: {} | |
| }; | |
| const element = document.currentScript.parentElement; | |
| const data = [null,null]; | |
| Promise.all([ | |
| import("/docs/diffusers/pr_12448/en/_app/immutable/entry/start.77018737.js"), | |
| import("/docs/diffusers/pr_12448/en/_app/immutable/entry/app.30e12370.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 296], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 29.6 kB
- Xet hash:
- c4f78059904cd24d098f18ad0c8572a7906e56dd69b542b3a76947294ef72399
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.