datasets/v2.6.2/en/image_process.html

<meta charset="utf-8" /><meta http-equiv="content-security-policy" content=""><meta name="hf:doc:metadata" content="{&quot;local&quot;:&quot;process-image-data&quot;,&quot;sections&quot;:[{&quot;local&quot;:&quot;map&quot;,&quot;title&quot;:&quot;Map&quot;},{&quot;local&quot;:&quot;apply-transforms&quot;,&quot;title&quot;:&quot;Apply transforms&quot;}],&quot;title&quot;:&quot;Process image data&quot;}" data-svelte="svelte-1phssyn">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/assets/pages/__layout.svelte-hf-doc-builder.css">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/start-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/chunks/vendor-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/chunks/paths-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/pages/__layout.svelte-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/pages/image_process.mdx-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/chunks/Tip-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/chunks/IconCopyLink-hf-doc-builder.js">
	<link rel="modulepreload" href="/docs/datasets/v2.6.2/en/_app/chunks/CodeBlock-hf-doc-builder.js"> 





<h1 class="relative group"><a id="process-image-data" 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="#process-image-data"><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>Process image data
	</span></h1>

<p>This guide shows specific methods for processing image datasets. Learn how to:</p>
<ul><li>Use <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a> with image dataset.</li>
<li>Apply data augmentations to a dataset with <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.set_transform">set_transform()</a>.</li></ul>
<p>For a guide on how to process any type of dataset, take a look at the <a class="underline decoration-sky-400 decoration-2 font-semibold" href="./process">general process guide</a>.</p>
<h2 class="relative group"><a id="map" 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="#map"><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>Map
	</span></h2>

<p>The <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a> function can apply transforms over an entire dataset.</p>
<p>For example, create a basic <a href="https://pytorch.org/vision/stable/generated/torchvision.transforms.Resize.html" rel="nofollow"><code>Resize</code></a> function:</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><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">def</span> <span class="hljs-title function_">transforms</span>(<span class="hljs-params">examples</span>):
<span class="hljs-meta">... </span>    examples[<span class="hljs-string">&quot;pixel_values&quot;</span>] = [image.convert(<span class="hljs-string">&quot;RGB&quot;</span>).resize((<span class="hljs-number">100</span>,<span class="hljs-number">100</span>)) <span class="hljs-keyword">for</span> image <span class="hljs-keyword">in</span> examples[<span class="hljs-string">&quot;image&quot;</span>]]
<span class="hljs-meta">... </span>    <span class="hljs-keyword">return</span> examples<!-- HTML_TAG_END --></pre></div>
<p>Now use the <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a> function to resize the entire dataset, and set <code>batched=True</code> to speed up the process by accepting batches of examples. The transform returns <code>pixel_values</code> as a cacheable <code>PIL.Image</code> object:</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><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span>dataset = dataset.<span class="hljs-built_in">map</span>(transforms, remove_columns=[<span class="hljs-string">&quot;image&quot;</span>], batched=<span class="hljs-literal">True</span>)
<span class="hljs-meta">&gt;&gt;&gt; </span>dataset[<span class="hljs-number">0</span>]
{<span class="hljs-string">&#x27;label&#x27;</span>: <span class="hljs-number">6</span>,
 <span class="hljs-string">&#x27;pixel_values&#x27;</span>: &lt;PIL.PngImagePlugin.PngImageFile image mode=RGB size=100x100 at <span class="hljs-number">0x7F058237BB10</span>&gt;}<!-- HTML_TAG_END --></pre></div>
<p>The cache file saves time because you don’t have to execute the same transform twice. The <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a> function is best for operations you only run once per training - like resizing an image - instead of using it for operations executed for each epoch, like data augmentations.</p>
<p><a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a> takes up some memory, but you can reduce its memory requirements with the following parameters:</p>
<ul><li><a href="./package_reference/main_classes#datasets.DatasetDict.map.batch_size"><code>batch_size</code></a> determines the number of examples that are processed in one call to the transform function.</li>
<li><a href="./package_reference/main_classes#datasets.DatasetDict.map.writer_batch_size"><code>writer_batch_size</code></a> determines the number of processed examples that are kept in memory before they are stored away.</li></ul>
<p>Both parameter values default to 1000, which can be expensive if you are storing images. Lower these values to use less memory when you use <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.map">map()</a>.</p>
<h2 class="relative group"><a id="apply-transforms" 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="#apply-transforms"><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>Apply transforms
	</span></h2>

<p>🤗 Datasets applies data augmentations from any library or package to your dataset. Transforms can be applied on-the-fly on batches of data with <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.set_transform">set_transform()</a>, which consumes less disk space.</p>


<div class="course-tip  bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p>The following example uses <a href="https://pytorch.org/vision/stable/index.html" rel="nofollow">torchvision</a>, but feel free to use other data augmentation libraries like <a href="https://albumentations.ai/docs/" rel="nofollow">Albumentations</a>, <a href="https://kornia.readthedocs.io/en/latest/" rel="nofollow">Kornia</a>, and <a href="https://imgaug.readthedocs.io/en/latest/" rel="nofollow">imgaug</a>.</p></div>
<p>For example, if you’d like to change the color properties of an image randomly:</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><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> torchvision.transforms <span class="hljs-keyword">import</span> Compose, ColorJitter, ToTensor

<span class="hljs-meta">&gt;&gt;&gt; </span>jitter = Compose(
<span class="hljs-meta">... </span>    [
<span class="hljs-meta">... </span>         ColorJitter(brightness=<span class="hljs-number">0.25</span>, contrast=<span class="hljs-number">0.25</span>, saturation=<span class="hljs-number">0.25</span>, hue=<span class="hljs-number">0.7</span>),
<span class="hljs-meta">... </span>         ToTensor(),
<span class="hljs-meta">... </span>    ]
<span class="hljs-meta">... </span>)<!-- HTML_TAG_END --></pre></div>
<p>Create a function to apply the <code>ColorJitter</code> transform:</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><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">def</span> <span class="hljs-title function_">transforms</span>(<span class="hljs-params">examples</span>):
<span class="hljs-meta">... </span>    examples[<span class="hljs-string">&quot;pixel_values&quot;</span>] = [jitter(image.convert(<span class="hljs-string">&quot;RGB&quot;</span>)) <span class="hljs-keyword">for</span> image <span class="hljs-keyword">in</span> examples[<span class="hljs-string">&quot;image&quot;</span>]]
<span class="hljs-meta">... </span>    <span class="hljs-keyword">return</span> examples<!-- HTML_TAG_END --></pre></div>
<p>Apply the transform with the <a href="/docs/datasets/v2.6.2/en/package_reference/main_classes#datasets.Dataset.set_transform">set_transform()</a> function:</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><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span>dataset.set_transform(transforms)<!-- HTML_TAG_END --></pre></div>


		<script type="module" data-hydrate="z66z53">
		import { start } from "/docs/datasets/v2.6.2/en/_app/start-hf-doc-builder.js";
		start({
			target: document.querySelector('[data-hydrate="z66z53"]').parentNode,
			paths: {"base":"/docs/datasets/v2.6.2/en","assets":"/docs/datasets/v2.6.2/en"},
			session: {},
			route: false,
			spa: false,
			trailing_slash: "never",
			hydrate: {
				status: 200,
				error: null,
				nodes: [
					import("/docs/datasets/v2.6.2/en/_app/pages/__layout.svelte-hf-doc-builder.js"),
						import("/docs/datasets/v2.6.2/en/_app/pages/image_process.mdx-hf-doc-builder.js")
				],
				params: {}
			}
		});
	</script>