Buckets:

hf-doc-build
/

doc-dev

Files

xet

hf-doc-build/doc-dev / optimum-neuron /pr_1097 /en /supported_architectures.html

rtrm

about 1 month ago

download

raw

20.8 kB

	<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Supported architectures","local":"supported-architectures","sections":[{"title":"Training","local":"training","sections":[],"depth":2},{"title":"Inference","local":"inference","sections":[{"title":"Transformers","local":"transformers","sections":[],"depth":3},{"title":"Diffusers","local":"diffusers","sections":[],"depth":3},{"title":"Sentence Transformers","local":"sentence-transformers","sections":[],"depth":3}],"depth":2}],"depth":1}">
	<link href="/docs/optimum.neuron/pr_1097/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/entry/start.82e2ff17.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/scheduler.56725da7.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/singletons.2080b4fc.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/paths.90dabf70.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/entry/app.9b0c0103.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/preload-helper.9dba61fb.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/index.18a26576.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/nodes/0.912aab06.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/each.e59479a4.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/nodes/45.44383935.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/Tip.5b941656.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/CopyLLMTxtMenu.fb3856d8.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/globals.7f7f1b26.js">
	<link rel="modulepreload" href="/docs/optimum.neuron/pr_1097/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.a16844e0.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Supported architectures","local":"supported-architectures","sections":[{"title":"Training","local":"training","sections":[],"depth":2},{"title":"Inference","local":"inference","sections":[{"title":"Transformers","local":"transformers","sections":[],"depth":3},{"title":"Diffusers","local":"diffusers","sections":[],"depth":3},{"title":"Sentence Transformers","local":"sentence-transformers","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 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>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> </div> <h1 class="relative group"><a id="supported-architectures" 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-architectures"><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 architectures</span></h1> <h2 class="relative group"><a id="training" 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="#training"><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>Training</span></h2> <p data-svelte-h="svelte-bs045o">Training on AWS Trainium instances (Trn1) enables large-scale model training with distributed parallelism strategies.</p> <p data-svelte-h="svelte-11lzvnb"><strong>Requirements:</strong></p> <ul data-svelte-h="svelte-86fnwf"><li>Model must be compatible with the Neuron SDK. If it small enough to fit within 16GB, training is supported for any architecture that can be successfully compiled.</li> <li><strong>Memory constraint:</strong> Each accelerator has 16GB of memory for model weights, gradients, optimizer states, and activations.</li> <li><strong>For large models:</strong> Custom modeling implementation with tensor parallelism and/or pipeline parallelism support is required.</li></ul> <p data-svelte-h="svelte-1x9hhfd">The following architectures have custom modeling implementations with distributed training support:</p> <table data-svelte-h="svelte-1r98o4n"><thead><tr><th>Architecture</th> <th>Task</th> <th>Tensor Parallelism</th> <th>Pipeline Parallelism</th></tr></thead> <tbody><tr><td>Llama, Llama 2, Llama 3</td> <td>text-generation</td> <td>✓</td> <td>✓</td></tr> <tr><td>Qwen3</td> <td>text-generation</td> <td>✓</td> <td>✓</td></tr> <tr><td>Granite</td> <td>text-generation</td> <td>✓</td> <td>✗</td></tr></tbody></table> <blockquote class="tip"><p data-svelte-h="svelte-1ottfjl">If you need to add support for a custom model not listed above, check out our <a href="./contribute/contribute_for_training">contribute for training guide</a> to learn how to implement custom modeling with distributed training support. You can also open an issue in the <a href="https://github.com/huggingface/optimum-neuron/issues" rel="nofollow">Optimum Neuron GitHub repository</a> to request support for it.</p></blockquote> <h2 class="relative group"><a id="inference" 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="#inference"><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>Inference</span></h2> <p data-svelte-h="svelte-11aox24">The following table lists the architectures and tasks that Optimum Neuron supports for inference on Amazon EC2 Inf2 instances.</p> <blockquote class="tip"><p data-svelte-h="svelte-v50c0e">If a LLM is listed, e.g. a model with a <code>text-generation</code> task, it means that there is also <a href="https://github.com/vllm-project/vllm" rel="nofollow">vLLM</a> support for it.</p></blockquote> <h3 class="relative group"><a id="transformers" 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="#transformers"><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>Transformers</span></h3> <table data-svelte-h="svelte-14vt7nu"><thead><tr><th>Architecture</th> <th>Task</th></tr></thead> <tbody><tr><td>ALBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>AST</td> <td>feature-extraction, audio-classification</td></tr> <tr><td>BERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>Beit</td> <td>feature-extraction, image-classification</td></tr> <tr><td>CamemBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>CLIP</td> <td>feature-extraction, image-classification</td></tr> <tr><td>ConvBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>ConvNext</td> <td>feature-extraction, image-classification</td></tr> <tr><td>ConvNextV2</td> <td>feature-extraction, image-classification</td></tr> <tr><td>CvT</td> <td>feature-extraction, image-classification</td></tr> <tr><td>DeBERTa (INF2 only)</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>DeBERTa-v2 (INF2 only)</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>Deit</td> <td>feature-extraction, image-classification</td></tr> <tr><td>DistilBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>DonutSwin</td> <td>feature-extraction</td></tr> <tr><td>Dpt</td> <td>feature-extraction</td></tr> <tr><td>ELECTRA</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>ESM</td> <td>feature-extraction, fill-mask, text-classification, token-classification</td></tr> <tr><td>FlauBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>Granite</td> <td>text-generation</td></tr> <tr><td>Hubert</td> <td>feature-extraction, automatic-speech-recognition, audio-classification</td></tr> <tr><td>Levit</td> <td>feature-extraction, image-classification</td></tr> <tr><td>Llama, Llama 2, Llama 3</td> <td>text-generation</td></tr> <tr><td>Llama 4</td> <td>text-generation</td></tr> <tr><td>Mixtral</td> <td>text-generation</td></tr> <tr><td>MobileBERT</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>MobileNetV2</td> <td>feature-extraction, image-classification, semantic-segmentation</td></tr> <tr><td>MobileViT</td> <td>feature-extraction, image-classification, semantic-segmentation</td></tr> <tr><td>ModernBERT</td> <td>feature-extraction, fill-mask, text-classification, token-classification</td></tr> <tr><td>MPNet</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>Phi3</td> <td>text-generation</td></tr> <tr><td>Phi</td> <td>feature-extraction, text-classification, token-classification</td></tr> <tr><td>Qwen2</td> <td>text-generation</td></tr> <tr><td>Qwen3</td> <td>feature-extraction, text-generation</td></tr> <tr><td>Qwen3Moe</td> <td>text-generation</td></tr> <tr><td>RoBERTa</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>RoFormer</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>SmolLM3</td> <td>text-generation</td></tr> <tr><td>Swin</td> <td>feature-extraction, image-classification</td></tr> <tr><td>T5</td> <td>text2text-generation</td></tr> <tr><td>UniSpeech</td> <td>feature-extraction, automatic-speech-recognition, audio-classification</td></tr> <tr><td>UniSpeech-SAT</td> <td>feature-extraction, automatic-speech-recognition, audio-classification, audio-frame-classification, audio-xvector</td></tr> <tr><td>ViT</td> <td>feature-extraction, image-classification</td></tr> <tr><td>Wav2Vec2</td> <td>feature-extraction, automatic-speech-recognition, audio-classification, audio-frame-classification, audio-xvector</td></tr> <tr><td>WavLM</td> <td>feature-extraction, automatic-speech-recognition, audio-classification, audio-frame-classification, audio-xvector</td></tr> <tr><td>Whisper</td> <td>automatic-speech-recognition</td></tr> <tr><td>XLM</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>XLM-RoBERTa</td> <td>feature-extraction, fill-mask, multiple-choice, question-answering, text-classification, token-classification</td></tr> <tr><td>Yolos</td> <td>feature-extraction, object-detection</td></tr></tbody></table> <h3 class="relative group"><a id="diffusers" 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"><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>Diffusers</span></h3> <table data-svelte-h="svelte-xzva6e"><thead><tr><th>Architecture</th> <th>Task</th></tr></thead> <tbody><tr><td>Stable Diffusion</td> <td>text-to-image, image-to-image, inpaint</td></tr> <tr><td>Stable Diffusion XL Base</td> <td>text-to-image, image-to-image, inpaint</td></tr> <tr><td>Stable Diffusion XL Refiner</td> <td>image-to-image, inpaint</td></tr> <tr><td>SDXL Turbo</td> <td>text-to-image, image-to-image, inpaint</td></tr> <tr><td>LCM</td> <td>text-to-image</td></tr> <tr><td>PixArt-α</td> <td>text-to-image</td></tr> <tr><td>PixArt-Σ</td> <td>text-to-image</td></tr> <tr><td>Flux</td> <td>text-to-image, inpaint</td></tr> <tr><td>Flux Kontext</td> <td>text-to-image, image-to-image</td></tr></tbody></table> <h3 class="relative group"><a id="sentence-transformers" 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="#sentence-transformers"><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>Sentence Transformers</span></h3> <table data-svelte-h="svelte-zeg51l"><thead><tr><th>Architecture</th> <th>Task</th></tr></thead> <tbody><tr><td>Transformer</td> <td>feature-extraction, sentence-similarity</td></tr> <tr><td>CLIP</td> <td>feature-extraction, zero-shot-image-classification</td></tr></tbody></table> <blockquote class="tip"><p data-svelte-h="svelte-kd3kkz">To learn how to export a model for inference, you can check this <a href="https://huggingface.co/docs/optimum-neuron/guides/export_model#selecting-a-task" rel="nofollow">guide</a>.</p></blockquote> <p></p>

	<script>
	{
	__sveltekit_19pvy0q = {
	assets: "/docs/optimum.neuron/pr_1097/en",
	base: "/docs/optimum.neuron/pr_1097/en",
	env: {}
	};

	const element = document.currentScript.parentElement;

	const data = [null,null];

	Promise.all([
	import("/docs/optimum.neuron/pr_1097/en/_app/immutable/entry/start.82e2ff17.js"),
	import("/docs/optimum.neuron/pr_1097/en/_app/immutable/entry/app.9b0c0103.js")
	]).then(([kit, app]) => {
	kit.start(app, element, {
	node_ids: [0, 45],
	data,
	form: null,
	error: null
	});
	});
	}
	</script>

Xet Storage Details

Size:: 20.8 kB
Xet hash:: 02c754a0f2a339b0334eddd3b9bca445035095b7398dec6f23f8fb80ec267f1c

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