Buckets:

hf-doc-build
/

doc-dev

Files

xet

hf-doc-build/doc-dev / diffusers /pr_12242 /en /using-diffusers /models.html

rtrm

17 days ago

download

raw

28.3 kB

	<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Models","local":"models","sections":[{"title":"Loading a model","local":"loading-a-model","sections":[],"depth":2},{"title":"AutoModel","local":"automodel","sections":[],"depth":2},{"title":"Model data types","local":"model-data-types","sections":[],"depth":2},{"title":"Device placement","local":"device-placement","sections":[],"depth":2},{"title":"Saving models","local":"saving-models","sections":[],"depth":2}],"depth":1}">
	<link href="/docs/diffusers/pr_12242/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/entry/start.f08d0bba.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/scheduler.8c3d61f6.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/singletons.4845f8b6.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/index.0997d446.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/paths.5de71654.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/entry/app.121c7723.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/index.da70eac4.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/nodes/0.8d75f0af.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/each.e59479a4.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/nodes/311.9fcfc2a6.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/CodeBlock.a9c4becf.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/DocNotebookDropdown.48852948.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/globals.7f7f1b26.js">
	<link rel="modulepreload" href="/docs/diffusers/pr_12242/en/_app/immutable/chunks/getInferenceSnippets.8d281f31.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Models","local":"models","sections":[{"title":"Loading a model","local":"loading-a-model","sections":[],"depth":2},{"title":"AutoModel","local":"automodel","sections":[],"depth":2},{"title":"Model data types","local":"model-data-types","sections":[],"depth":2},{"title":"Device placement","local":"device-placement","sections":[],"depth":2},{"title":"Saving models","local":"saving-models","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <div class="flex space-x-1 absolute z-10 right-0 top-0"> <div class="relative colab-dropdown "> <button class=" " type="button"> <img alt="Open In Colab" class="!m-0" src="https://colab.research.google.com/assets/colab-badge.svg"> </button> </div> <div class="relative colab-dropdown "> <button class=" " type="button"> <img alt="Open In Studio Lab" class="!m-0" src="https://studiolab.sagemaker.aws/studiolab.svg"> </button> </div></div> <h1 class="relative group"><a id="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="#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>Models</span></h1> <p data-svelte-h="svelte-12lg8hb">A diffusion model relies on a few individual models working together to generate an output. These models are responsible for denoising, encoding inputs, and decoding latents into the actual outputs.</p> <p data-svelte-h="svelte-174dtxl">This guide will show you how to load models.</p> <h2 class="relative group"><a id="loading-a-model" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#loading-a-model"><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>Loading a model</span></h2> <p data-svelte-h="svelte-2zzfof">All models are loaded with the <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> method, which downloads and caches the latest model version. If the latest files are available in the local cache, <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> reuses files in the cache.</p> <p data-svelte-h="svelte-1pzmt0z">Pass the <code>subfolder</code> argument to <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> to specify where to load the model weights from. Omit the <code>subfolder</code> argument if the repository doesn’t have a subfolder structure or if you’re loading a standalone model.</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">from</span> diffusers <span class="hljs-keyword">import</span> QwenImageTransformer2DModel

	model = QwenImageTransformer2DModel.from_pretrained(<span class="hljs-string">"Qwen/Qwen-Image"</span>, subfolder=<span class="hljs-string">"transformer"</span>)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="automodel" 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="#automodel"><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>AutoModel</span></h2> <p data-svelte-h="svelte-19a0opd"><a href="/docs/diffusers/pr_12242/en/api/models/auto_model#diffusers.AutoModel">AutoModel</a> detects the model class from a <code>model_index.json</code> file or a model’s <code>config.json</code> file. It fetches the correct model class from these files and delegates the actual loading to the model class. <a href="/docs/diffusers/pr_12242/en/api/models/auto_model#diffusers.AutoModel">AutoModel</a> is useful for automatic model type detection without needing to know the exact model class beforehand.</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">from</span> diffusers <span class="hljs-keyword">import</span> AutoModel

	model = AutoModel.from_pretrained(
	<span class="hljs-string">"Qwen/Qwen-Image"</span>, subfolder=<span class="hljs-string">"transformer"</span>
	)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="model-data-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="#model-data-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>Model data types</span></h2> <p data-svelte-h="svelte-39ettl">Use the <code>torch_dtype</code> argument in <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> to load a model with a specific data type. This allows you to load a model in a lower precision to reduce memory usage.</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> QwenImageTransformer2DModel

	model = QwenImageTransformer2DModel.from_pretrained(
	<span class="hljs-string">"Qwen/Qwen-Image"</span>,
	subfolder=<span class="hljs-string">"transformer"</span>,
	torch_dtype=torch.bfloat16
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-urfcp0"><a href="https://docs.pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.to" rel="nofollow">nn.Module.to</a> can also convert to a specific data type on the fly. However, it converts <em>all</em> weights to the requested data type unlike <code>torch_dtype</code> which respects <code>_keep_in_fp32_modules</code>. This argument preserves layers in <code>torch.float32</code> for numerical stability and best generation quality (see example <a href="https://github.com/huggingface/diffusers/blob/f864a9a352fa4a220d860bfdd1782e3e5af96382/src/diffusers/models/transformers/transformer_wan.py#L374" rel="nofollow">_keep_in_fp32_modules</a>)</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">from</span> diffusers <span class="hljs-keyword">import</span> QwenImageTransformer2DModel

	model = QwenImageTransformer2DModel.from_pretrained(
	<span class="hljs-string">"Qwen/Qwen-Image"</span>, subfolder=<span class="hljs-string">"transformer"</span>
	)
	model = model.to(dtype=torch.float16) <!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="device-placement" 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="#device-placement"><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>Device placement</span></h2> <p data-svelte-h="svelte-1761eke">Use the <code>device_map</code> argument in <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> to place a model on an accelerator like a GPU. It is especially helpful where there are multiple GPUs.</p> <p data-svelte-h="svelte-1ef9vko">Diffusers currently provides three options to <code>device_map</code> for individual models, <code>"cuda"</code>, <code>"balanced"</code> and <code>"auto"</code>. Refer to the table below to compare the three placement strategies.</p> <table data-svelte-h="svelte-27jvvd"><thead><tr><th>parameter</th> <th>description</th></tr></thead> <tbody><tr><td><code>"cuda"</code></td> <td>places pipeline on a supported accelerator (CUDA)</td></tr> <tr><td><code>"balanced"</code></td> <td>evenly distributes pipeline on all GPUs</td></tr> <tr><td><code>"auto"</code></td> <td>distribute model from fastest device first to slowest</td></tr></tbody></table> <p data-svelte-h="svelte-mgdx4k">Use the <code>max_memory</code> argument in <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.from_pretrained">from_pretrained()</a> to allocate a maximum amount of memory to use on each device. By default, Diffusers uses the maximum amount available.</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> QwenImagePipeline

	max_memory = {<span class="hljs-number">0</span>: <span class="hljs-string">"16GB"</span>, <span class="hljs-number">1</span>: <span class="hljs-string">"16GB"</span>}
	pipeline = QwenImagePipeline.from_pretrained(
	<span class="hljs-string">"Qwen/Qwen-Image"</span>,
	torch_dtype=torch.bfloat16,
	device_map=<span class="hljs-string">"cuda"</span>,
	max_memory=max_memory
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-jt67uv">The <code>hf_device_map</code> attribute allows you to access and view the <code>device_map</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-built_in">print</span>(transformer.hf_device_map)
	<span class="hljs-comment"># {'': device(type='cuda')}</span><!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="saving-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-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 models</span></h2> <p data-svelte-h="svelte-8weehm">Save a model with the <a href="/docs/diffusers/pr_12242/en/api/models/overview#diffusers.ModelMixin.save_pretrained">save_pretrained()</a> method.</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">from</span> diffusers <span class="hljs-keyword">import</span> QwenImageTransformer2DModel

	model = QwenImageTransformer2DModel.from_pretrained(<span class="hljs-string">"Qwen/Qwen-Image"</span>, subfolder=<span class="hljs-string">"transformer"</span>)
	model.save_pretrained(<span class="hljs-string">"./local/model"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-v59muj">For large models, it is helpful to use <code>max_shard_size</code> to save a model as multiple shards. A shard can be loaded faster and save memory (refer to the <a href="./loading#parallel-loading">parallel loading</a> docs for more details), especially if there is more than one 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 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 -->model.save_pretrained(<span class="hljs-string">"./local/model"</span>, max_shard_size=<span class="hljs-string">"5GB"</span>)<!-- HTML_TAG_END --></pre></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/using-diffusers/models.md" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span data-svelte-h="svelte-x0xyl0">></span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a> <p></p>

	<script>
	{
	__sveltekit_6g2lna = {
	assets: "/docs/diffusers/pr_12242/en",
	base: "/docs/diffusers/pr_12242/en",
	env: {}
	};

	const element = document.currentScript.parentElement;

	const data = [null,null];

	Promise.all([
	import("/docs/diffusers/pr_12242/en/_app/immutable/entry/start.f08d0bba.js"),
	import("/docs/diffusers/pr_12242/en/_app/immutable/entry/app.121c7723.js")
	]).then(([kit, app]) => {
	kit.start(app, element, {
	node_ids: [0, 311],
	data,
	form: null,
	error: null
	});
	});
	}
	</script>

Xet Storage Details

Size:: 28.3 kB
Xet hash:: d4a7e3065055b920d748f61724432f9fb8a83d4bd086eaf5d60349e6bf34e3d6

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