Hugging Face's logo

Buckets:
hf-doc-build
/
doc-dev

Files
xet
hf-doc-build/doc-dev / accelerate /pr_4021 /en /usage_guides /quantization.html
HuggingFaceDocBuilder's picture
HuggingFaceDocBuilder
download
raw
37.2 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Model quantization&quot;,&quot;local&quot;:&quot;model-quantization&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;bitsandbytes Integration&quot;,&quot;local&quot;:&quot;bitsandbytes-integration&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Pre-Requisites&quot;,&quot;local&quot;:&quot;pre-requisites&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;How it works&quot;,&quot;local&quot;:&quot;how-it-works&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Saving and loading 8-bit model&quot;,&quot;local&quot;:&quot;saving-and-loading-8-bit-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Offload modules to cpu and disk&quot;,&quot;local&quot;:&quot;offload-modules-to-cpu-and-disk&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Fine-tune a quantized model&quot;,&quot;local&quot;:&quot;fine-tune-a-quantized-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Example demo - running GPT2 1.5b on a Google Colab&quot;,&quot;local&quot;:&quot;example-demo---running-gpt2-15b-on-a-google-colab&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2}],&quot;depth&quot;:1}">
 <link href="/docs/accelerate/pr_4021/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/entry/start.8a49e72b.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/scheduler.b9285784.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/singletons.7547c222.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/index.6d423e5c.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/paths.d42c9205.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/entry/app.1df4d18e.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/preload-helper.b0bd19d1.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/index.26bc89a1.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/nodes/0.0e7c56e8.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/each.e59479a4.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/nodes/56.db981804.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.7a0ae628.js">
 <link rel="modulepreload" href="/docs/accelerate/pr_4021/en/_app/immutable/chunks/CodeBlock.844ff9c3.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Model quantization&quot;,&quot;local&quot;:&quot;model-quantization&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;bitsandbytes Integration&quot;,&quot;local&quot;:&quot;bitsandbytes-integration&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Pre-Requisites&quot;,&quot;local&quot;:&quot;pre-requisites&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;How it works&quot;,&quot;local&quot;:&quot;how-it-works&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Saving and loading 8-bit model&quot;,&quot;local&quot;:&quot;saving-and-loading-8-bit-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Offload modules to cpu and disk&quot;,&quot;local&quot;:&quot;offload-modules-to-cpu-and-disk&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Fine-tune a quantized model&quot;,&quot;local&quot;:&quot;fine-tune-a-quantized-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;Example demo - running GPT2 1.5b on a Google Colab&quot;,&quot;local&quot;:&quot;example-demo---running-gpt2-15b-on-a-google-colab&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2}],&quot;depth&quot;: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="model-quantization" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#model-quantization"><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 quantization</span></h1> <h2 class="relative group"><a id="bitsandbytes-integration" 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="#bitsandbytes-integration"><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>bitsandbytes Integration</span></h2> <p data-svelte-h="svelte-1uk3wou">Accelerate brings <code>bitsandbytes</code> quantization to your model. You can now load any pytorch model in 8-bit or 4-bit with a few lines of code.</p> <p data-svelte-h="svelte-1p54xze">If you want to use Transformers models with <code>bitsandbytes</code>, you should follow this <a href="https://huggingface.co/docs/transformers/main_classes/quantization" rel="nofollow">documentation</a>.</p> <p data-svelte-h="svelte-hrid2v">To learn more about how the <code>bitsandbytes</code> quantization works, check out the blog posts on <a href="https://huggingface.co/blog/hf-bitsandbytes-integration" rel="nofollow">8-bit quantization</a> and <a href="https://huggingface.co/blog/4bit-transformers-bitsandbytes" rel="nofollow">4-bit quantization</a>.</p> <h3 class="relative group"><a id="pre-requisites" 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="#pre-requisites"><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>Pre-Requisites</span></h3> <p data-svelte-h="svelte-50fa4n">You will need to install the following requirements:</p> <ul data-svelte-h="svelte-t2geh8"><li>Install <code>bitsandbytes</code> library</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 -->pip install bitsandbytes<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-s1psu3">For non-cuda devices, you can refer to the bitsandbytes installation guide <a href="https://huggingface.co/docs/bitsandbytes/main/en/installation#multi-backend" rel="nofollow">here</a>.</p> <ul data-svelte-h="svelte-l0cm28"><li>Install latest <code>accelerate</code> from source</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 -->pip install git+https://github.com/huggingface/accelerate.git<!-- HTML_TAG_END --></pre></div> <ul data-svelte-h="svelte-1c2t6dx"><li>Install <code>minGPT</code> and <code>huggingface_hub</code> to run examples</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 -->git <span class="hljs-built_in">clone</span> https://github.com/karpathy/minGPT.git
pip install minGPT/
pip install huggingface_hub<!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="how-it-works" 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="#how-it-works"><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>How it works</span></h3> <p data-svelte-h="svelte-sp5qz0">First, we need to initialize our model. To save memory, we can initialize an empty model using the context manager <a href="/docs/accelerate/pr_4021/en/package_reference/big_modeling#accelerate.init_empty_weights">init_empty_weights()</a>.</p> <p data-svelte-h="svelte-1fwkss5">Let’s take the GPT2 model from minGPT library.</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> accelerate <span class="hljs-keyword">import</span> init_empty_weights
<span class="hljs-keyword">from</span> mingpt.model <span class="hljs-keyword">import</span> GPT
model_config = GPT.get_default_config()
model_config.model_type = <span class="hljs-string">&#x27;gpt2-xl&#x27;</span>
model_config.vocab_size = <span class="hljs-number">50257</span>
model_config.block_size = <span class="hljs-number">1024</span>
<span class="hljs-keyword">with</span> init_empty_weights():
 empty_model = GPT(model_config)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-bbyjp5">Then, we need to get the path to the weights of your model. The path can be the state_dict file (e.g. “pytorch_model.bin”) or a folder containing the sharded checkpoints.</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> huggingface_hub <span class="hljs-keyword">import</span> snapshot_download
weights_location = snapshot_download(repo_id=<span class="hljs-string">&quot;marcsun13/gpt2-xl-linear-sharded&quot;</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-175gqdq">Finally, you need to set your quantization configuration with <a href="/docs/accelerate/pr_4021/en/package_reference/utilities#accelerate.utils.BnbQuantizationConfig">BnbQuantizationConfig</a>.</p> <p data-svelte-h="svelte-4hyyr7">Here’s an example for 8-bit quantization:</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> accelerate.utils <span class="hljs-keyword">import</span> BnbQuantizationConfig
bnb_quantization_config = BnbQuantizationConfig(load_in_8bit=<span class="hljs-literal">True</span>, llm_int8_threshold = <span class="hljs-number">6</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1qtcdkv">Here’s an example for 4-bit quantization:</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> accelerate.utils <span class="hljs-keyword">import</span> BnbQuantizationConfig
bnb_quantization_config = BnbQuantizationConfig(load_in_4bit=<span class="hljs-literal">True</span>, bnb_4bit_compute_dtype=torch.bfloat16, bnb_4bit_use_double_quant=<span class="hljs-literal">True</span>, bnb_4bit_quant_type=<span class="hljs-string">&quot;nf4&quot;</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1ny2zj4">To quantize your empty model with the selected configuration, you need to use <a href="/docs/accelerate/pr_4021/en/package_reference/utilities#accelerate.utils.load_and_quantize_model">load_and_quantize_model()</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> accelerate.utils <span class="hljs-keyword">import</span> load_and_quantize_model
quantized_model = load_and_quantize_model(empty_model, weights_location=weights_location, bnb_quantization_config=bnb_quantization_config)<!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="saving-and-loading-8-bit-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="#saving-and-loading-8-bit-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>Saving and loading 8-bit model</span></h3> <p data-svelte-h="svelte-in50zg">You can save your 8-bit model with accelerate using <a href="/docs/accelerate/pr_4021/en/package_reference/accelerator#accelerate.Accelerator.save_model">save_model()</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> accelerate <span class="hljs-keyword">import</span> Accelerator
accelerate = Accelerator()
new_weights_location = <span class="hljs-string">&quot;path/to/save_directory&quot;</span>
accelerate.save_model(quantized_model, new_weights_location)
quantized_model_from_saved = load_and_quantize_model(empty_model, weights_location=new_weights_location, bnb_quantization_config=bnb_quantization_config, device_map = <span class="hljs-string">&quot;auto&quot;</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1x0s82b">Note that 4-bit model serialization is currently not supported.</p> <h3 class="relative group"><a id="offload-modules-to-cpu-and-disk" 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="#offload-modules-to-cpu-and-disk"><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>Offload modules to cpu and disk</span></h3> <p data-svelte-h="svelte-1vy35js">You can offload some modules to cpu/disk if you don’t have enough space on the GPU to store the entire model on your GPUs.
This uses big model inference under the hood. Check this <a href="https://huggingface.co/docs/accelerate/usage_guides/big_modeling" rel="nofollow">documentation</a> for more details.</p> <p data-svelte-h="svelte-11qxtld">For 8-bit quantization, the selected modules will be converted to 8-bit precision.</p> <p data-svelte-h="svelte-qaw7ck">For 4-bit quantization, the selected modules will be kept in <code>torch_dtype</code> that the user passed in <code>BnbQuantizationConfig</code>. We will add support to convert these offloaded modules in 4-bit when 4-bit serialization will be possible.</p> <p data-svelte-h="svelte-gyjts5">You just need to pass a custom <code>device_map</code> in order to offload modules on cpu/disk. The offload modules will be dispatched on the GPU when needed. Here’s an example :</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 -->device_map = {
 <span class="hljs-string">&quot;transformer.wte&quot;</span>: <span class="hljs-number">0</span>,
 <span class="hljs-string">&quot;transformer.wpe&quot;</span>: <span class="hljs-number">0</span>,
 <span class="hljs-string">&quot;transformer.drop&quot;</span>: <span class="hljs-number">0</span>,
 <span class="hljs-string">&quot;transformer.h&quot;</span>: <span class="hljs-string">&quot;cpu&quot;</span>,
 <span class="hljs-string">&quot;transformer.ln_f&quot;</span>: <span class="hljs-string">&quot;disk&quot;</span>,
 <span class="hljs-string">&quot;lm_head&quot;</span>: <span class="hljs-string">&quot;disk&quot;</span>,
}<!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="fine-tune-a-quantized-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="#fine-tune-a-quantized-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>Fine-tune a quantized model</span></h3> <p data-svelte-h="svelte-rtwtrf">It is not possible to perform pure 8bit or 4bit training on these models. However, you can train these models by leveraging parameter efficient fine tuning methods (PEFT) and train for example adapters on top of them. Please have a look at <a href="https://github.com/huggingface/peft" rel="nofollow">peft</a> library for more details.</p> <p data-svelte-h="svelte-1c6wgo7">Currently, you can’t add adapters on top of any quantized model. However, with the official support of adapters with Transformers models, you can fine-tune quantized models. If you want to fine-tune a Transformers model , follow this <a href="https://huggingface.co/docs/transformers/main_classes/quantization" rel="nofollow">documentation</a> instead. Check out this <a href="https://colab.research.google.com/drive/1VoYNfYDKcKRQRor98Zbf2-9VQTtGJ24k?usp=sharing" rel="nofollow">demo</a> on how to fine-tune a 4-bit Transformers model.</p> <p data-svelte-h="svelte-1n8dj9v">Note that you don’t need to pass <code>device_map</code> when loading the model for training. It will automatically load your model on your GPU. Please note that <code>device_map=auto</code> should be used for inference only.</p> <h3 class="relative group"><a id="example-demo---running-gpt2-15b-on-a-google-colab" 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="#example-demo---running-gpt2-15b-on-a-google-colab"><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>Example demo - running GPT2 1.5b on a Google Colab</span></h3> <p data-svelte-h="svelte-19uggsw">Check out the Google Colab <a href="https://colab.research.google.com/drive/1T1pOgewAWVpR9gKpaEWw4orOrzPFb3yM?usp=sharing" rel="nofollow">demo</a> for running quantized models on a GPT2 model. The GPT2-1.5B model checkpoint is in FP32 which uses 6GB of memory. After quantization, it uses 1.6GB with 8-bit modules and 1.2GB with 4-bit modules.</p> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/accelerate/blob/main/docs/source/usage_guides/quantization.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_1q7nz6m = {
 assets: "/docs/accelerate/pr_4021/en",
 base: "/docs/accelerate/pr_4021/en",
 env: {}
 };
 const element = document.currentScript.parentElement;
 const data = [null,null];
 Promise.all([
 import("/docs/accelerate/pr_4021/en/_app/immutable/entry/start.8a49e72b.js"),
 import("/docs/accelerate/pr_4021/en/_app/immutable/entry/app.1df4d18e.js")
 ]).then(([kit, app]) => {
 kit.start(app, element, {
 node_ids: [0, 56],
 data,
 form: null,
 error: null
 });
 });
 }
 </script>

Xet Storage Details

Size:
37.2 kB
·
Xet hash:
773025a33c33a4352cb719f562b1ed85b001f14083354b1977128bfd47caf6c8

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