transformers/pr_33913/it/perf_train_cpu_many.html

<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Addestramento effciente su multiple CPU&quot;,&quot;local&quot;:&quot;addestramento-effciente-su-multiple-cpu&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Intel® oneCCL Bindings per PyTorch&quot;,&quot;local&quot;:&quot;intel-oneccl-bindings-per-pytorch&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Intel® oneCCL Bindings per l’installazione PyTorch:&quot;,&quot;local&quot;:&quot;intel-oneccl-bindings-per-linstallazione-pytorch&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Intel® MPI library&quot;,&quot;local&quot;:&quot;intel-mpi-library&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Installazione IPEX:&quot;,&quot;local&quot;:&quot;installazione-ipex&quot;,&quot;sections&quot;:[],&quot;depth&quot;:4}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Utilizzo in Trainer&quot;,&quot;local&quot;:&quot;utilizzo-in-trainer&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}">
		<link href="/docs/transformers/pr_33913/it/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/entry/start.cef8f86b.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/scheduler.36a0863c.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/singletons.39cd3b38.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/index.733708bb.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/paths.64312358.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/entry/app.d399d34f.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/index.9c13489a.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/nodes/0.7859fb6f.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/each.e59479a4.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/nodes/23.6d374ad6.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/Tip.3b06990e.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/CodeBlock.05d8ec32.js">
		<link rel="modulepreload" href="/docs/transformers/pr_33913/it/_app/immutable/chunks/EditOnGithub.e88f2b7b.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Addestramento effciente su multiple CPU&quot;,&quot;local&quot;:&quot;addestramento-effciente-su-multiple-cpu&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Intel® oneCCL Bindings per PyTorch&quot;,&quot;local&quot;:&quot;intel-oneccl-bindings-per-pytorch&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Intel® oneCCL Bindings per l’installazione PyTorch:&quot;,&quot;local&quot;:&quot;intel-oneccl-bindings-per-linstallazione-pytorch&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Intel® MPI library&quot;,&quot;local&quot;:&quot;intel-mpi-library&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Installazione IPEX:&quot;,&quot;local&quot;:&quot;installazione-ipex&quot;,&quot;sections&quot;:[],&quot;depth&quot;:4}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Utilizzo in Trainer&quot;,&quot;local&quot;:&quot;utilizzo-in-trainer&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"><!-- HEAD_svelte-u9bgzb_END -->      <p></p>   <h1 class="relative group"><a id="addestramento-effciente-su-multiple-cpu" 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="#addestramento-effciente-su-multiple-cpu"><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>Addestramento effciente su multiple CPU</span></h1> <p data-svelte-h="svelte-1kfu2az">Quando l’addestramento su una singola CPU è troppo lento, possiamo usare CPU multiple. Quasta guida si concentra su DDP basato su PyTorch abilitando l’addetramento distribuito su CPU in maniera efficiente.</p>  <h2 class="relative group"><a id="intel-oneccl-bindings-per-pytorch" 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="#intel-oneccl-bindings-per-pytorch"><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>Intel® oneCCL Bindings per PyTorch</span></h2> <p data-svelte-h="svelte-1lu0lvx"><a href="https://github.com/oneapi-src/oneCCL" rel="nofollow">Intel® oneCCL</a> (collective communications library) è una libreria per l’addestramento efficiente del deep learning in distribuito e implementa collettivi come allreduce, allgather, alltoall. Per maggiori informazioni su oneCCL, fai riferimento a <a href="https://spec.oneapi.com/versions/latest/elements/oneCCL/source/index.html" rel="nofollow">oneCCL documentation</a> e <a href="https://spec.oneapi.com/versions/latest/elements/oneCCL/source/index.html" rel="nofollow">oneCCL specification</a>.</p> <p data-svelte-h="svelte-ctw5mc">Il modulo <code>oneccl_bindings_for_pytorch</code> (<code>torch_ccl</code> precedentemente alla versione 1.12)  implementa PyTorch C10D ProcessGroup API e può essere caricato dinamicamente com external ProcessGroup e funziona solo su piattaforma Linux al momento.</p> <p data-svelte-h="svelte-12jdnef">Qui trovi informazioni più dettagliate per <a href="https://github.com/intel/torch-ccl" rel="nofollow">oneccl_bind_pt</a>.</p>  <h3 class="relative group"><a id="intel-oneccl-bindings-per-linstallazione-pytorch" 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="#intel-oneccl-bindings-per-linstallazione-pytorch"><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>Intel® oneCCL Bindings per l’installazione PyTorch:</span></h3> <p data-svelte-h="svelte-awoh7n">I file wheel sono disponibili per le seguenti versioni di Python:</p> <table data-svelte-h="svelte-1yk106u"><thead><tr><th align="center">Extension Version</th> <th align="center">Python 3.6</th> <th align="center">Python 3.7</th> <th align="center">Python 3.8</th> <th align="center">Python 3.9</th> <th align="center">Python 3.10</th></tr></thead> <tbody><tr><td align="center">1.13.0</td> <td align="center"></td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td></tr> <tr><td align="center">1.12.100</td> <td align="center"></td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td></tr> <tr><td align="center">1.12.0</td> <td align="center"></td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td></tr> <tr><td align="center">1.11.0</td> <td align="center"></td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td></tr> <tr><td align="center">1.10.0</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center">√</td> <td align="center"></td></tr></tbody></table> <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 oneccl_bind_pt=={pytorch_version} -f https://developer.intel.com/ipex-whl-stable-cpu<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-xq4hi3">dove <code>{pytorch_version}</code> deve essere la tua versione di PyTorch, per l’stanza 1.13.0.
Verifica altri approcci per <a href="https://github.com/intel/torch-ccl" rel="nofollow">oneccl_bind_pt installation</a>.
Le versioni di oneCCL e PyTorch devono combaciare.</p>  <div class="course-tip course-tip-orange bg-gradient-to-br dark:bg-gradient-to-r before:border-orange-500 dark:before:border-orange-800 from-orange-50 dark:from-gray-900 to-white dark:to-gray-950 border border-orange-50 text-orange-700 dark:text-gray-400"><p data-svelte-h="svelte-is6c7w">oneccl_bindings_for_pytorch 1.12.0 prebuilt wheel does not work with PyTorch 1.12.1 (it is for PyTorch 1.12.0)
PyTorch 1.12.1 should work with oneccl_bindings_for_pytorch 1.12.100</p></div>  <h2 class="relative group"><a id="intel-mpi-library" 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="#intel-mpi-library"><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>Intel® MPI library</span></h2> <p data-svelte-h="svelte-1v3lzwy">Usa questa implementazione basata su standard MPI per fornire una architettura flessibile, efficiente, scalabile su cluster per Intel®. Questo componente è parte di Intel® oneAPI HPC Toolkit.</p> <p data-svelte-h="svelte-7y5acl">oneccl_bindings_for_pytorch è installato insieme al set di strumenti MPI. Necessità di reperire l’ambiente prima di utilizzarlo.</p> <p data-svelte-h="svelte-37i02y">per Intel® oneCCL &gt;= 1.12.0</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 -->oneccl_bindings_for_pytorch_path=$(python -c <span class="hljs-string">&quot;from oneccl_bindings_for_pytorch import cwd; print(cwd)&quot;</span>)
<span class="hljs-built_in">source</span> <span class="hljs-variable">$oneccl_bindings_for_pytorch_path</span>/env/setvars.sh<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-2y4a8e">per Intel® oneCCL con versione &lt; 1.12.0</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 -->torch_ccl_path=$(python -c <span class="hljs-string">&quot;import torch; import torch_ccl; import os;  print(os.path.abspath(os.path.dirname(torch_ccl.__file__)))&quot;</span>)
<span class="hljs-built_in">source</span> <span class="hljs-variable">$torch_ccl_path</span>/env/setvars.sh<!-- HTML_TAG_END --></pre></div>  <h4 class="relative group"><a id="installazione-ipex" 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="#installazione-ipex"><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>Installazione IPEX:</span></h4> <p data-svelte-h="svelte-ucejiq">IPEX fornisce ottimizzazioni delle prestazioni per l’addestramento della CPU sia con Float32 che con BFloat16; puoi fare riferimento a <a href="./perf_train_cpu">single CPU section</a>.</p> <p data-svelte-h="svelte-1j4jwwj">Il seguente “Utilizzo in Trainer” prende come esempio mpirun nella libreria Intel® MPI.</p>  <h2 class="relative group"><a id="utilizzo-in-trainer" 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="#utilizzo-in-trainer"><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>Utilizzo in Trainer</span></h2> <p data-svelte-h="svelte-c1ynuc">Per abilitare l’addestramento distribuito multi CPU nel Trainer con il ccl backend, gli utenti devono aggiungere <strong><code>--ddp_backend ccl</code></strong> negli argomenti del comando.</p> <p data-svelte-h="svelte-15bz7m">Vediamo un esempio per il <a href="https://github.com/huggingface/transformers/tree/main/examples/pytorch/question-answering" rel="nofollow">question-answering example</a></p> <p data-svelte-h="svelte-1aptuua">Il seguente comando abilita due processi sul nodo Xeon, con un processo in esecuzione per ogni socket. Le variabili OMP_NUM_THREADS/CCL_WORKER_COUNT possono essere impostate per una prestazione ottimale.</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 --> export CCL_WORKER_COUNT=1
 export MASTER_ADDR=127.0.0.1
 mpirun -n 2 -genv OMP_NUM_THREADS=23 \
 python3 run_qa.py \
 --model_name_or_path google-bert/bert-large-uncased \
 --dataset_name squad \
 --do_train \
 --do_eval \
 --per_device_train_batch_size 12  \
 --learning_rate 3e-5  \
 --num_train_epochs 2  \
 --max_seq_length 384 \
 --doc_stride 128  \
 --output_dir /tmp/debug_squad/ \
 --no_cuda \
 --ddp_backend ccl \
 --use_ipex<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-8wv2la">Il seguente comando abilita l’addestramento per un totale di quattro processi su due Xeon (node0 e node1, prendendo node0 come processo principale), ppn (processes per node) è impostato a 2, on un processo in esecuzione per ogni socket. Le variabili OMP_NUM_THREADS/CCL_WORKER_COUNT possono essere impostate per una prestazione ottimale.</p> <p data-svelte-h="svelte-1m5a0mz">In node0, è necessario creare un file di configurazione che contenga gli indirizzi IP di ciascun nodo (per esempio hostfile) e passare il percorso del file di configurazione come parametro.</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 --> cat hostfile
 xxx.xxx.xxx.xxx #node0 ip
 xxx.xxx.xxx.xxx #node1 ip<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-lx9bti">A questo punto, esegui il seguente comando nel nodo0 e <strong>4DDP</strong> sarà abilitato in node0 e node1 con BF16 auto mixed precision:</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 --> export CCL_WORKER_COUNT=1
 export MASTER_ADDR=xxx.xxx.xxx.xxx #node0 ip
 mpirun -f hostfile -n 4 -ppn 2 \
 -genv OMP_NUM_THREADS=23 \
 python3 run_qa.py \
 --model_name_or_path google-bert/bert-large-uncased \
 --dataset_name squad \
 --do_train \
 --do_eval \
 --per_device_train_batch_size 12  \
 --learning_rate 3e-5  \
 --num_train_epochs 2  \
 --max_seq_length 384 \
 --doc_stride 128  \
 --output_dir /tmp/debug_squad/ \
 --no_cuda \
 --ddp_backend ccl \
 --use_ipex \
 --bf16<!-- 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/transformers/blob/main/docs/source/it/perf_train_cpu_many.md" target="_blank"><span data-svelte-h="svelte-1kd6by1">&lt;</span> <span data-svelte-h="svelte-x0xyl0">&gt;</span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a>  <p></p> 
			
			<script>
				{
					__sveltekit_728xau = {
						assets: "/docs/transformers/pr_33913/it",
						base: "/docs/transformers/pr_33913/it",
						env: {}
					};

					const element = document.currentScript.parentElement;

					const data = [null,null];

					Promise.all([
						import("/docs/transformers/pr_33913/it/_app/immutable/entry/start.cef8f86b.js"),
						import("/docs/transformers/pr_33913/it/_app/immutable/entry/app.d399d34f.js")
					]).then(([kit, app]) => {
						kit.start(app, element, {
							node_ids: [0, 23],
							data,
							form: null,
							error: null
						});
					});
				}
			</script>