Buckets:

hf-doc-build
/

doc

Files

xet

hf-doc-build/doc / transformers /main /ja /tasks /semantic_segmentation.html

HuggingFaceDocBuilder

about 2 hours ago

download

raw

69.3 kB

	<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"Semantic segmentation","local":"semantic-segmentation","sections":[{"title":"Load SceneParse150 dataset","local":"load-sceneparse150-dataset","sections":[],"depth":2},{"title":"Preprocess","local":"preprocess","sections":[],"depth":2},{"title":"Evaluate","local":"evaluate","sections":[],"depth":2},{"title":"Train","local":"train","sections":[],"depth":2},{"title":"Inference","local":"inference","sections":[],"depth":2}],"depth":1}">
	<link href="/docs/transformers/main/ja/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/entry/start.b596590e.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/scheduler.0c7da001.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/singletons.11a17a36.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/paths.351f5b4a.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/entry/app.ede2e8a7.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/preload-helper.68965f5e.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/index.e0be4b2b.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/nodes/0.0325648c.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/each.e59479a4.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/nodes/141.46420c87.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/Tip.f6e0bcd3.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/CopyLLMTxtMenu.6767f26f.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/globals.7f7f1b26.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.27cfd80f.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/Youtube.cf53141f.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/CodeBlock.418ebe13.js">
	<link rel="modulepreload" href="/docs/transformers/main/ja/_app/immutable/chunks/DocNotebookDropdown.367a2ff8.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"Semantic segmentation","local":"semantic-segmentation","sections":[{"title":"Load SceneParse150 dataset","local":"load-sceneparse150-dataset","sections":[],"depth":2},{"title":"Preprocess","local":"preprocess","sections":[],"depth":2},{"title":"Evaluate","local":"evaluate","sections":[],"depth":2},{"title":"Train","local":"train","sections":[],"depth":2},{"title":"Inference","local":"inference","sections":[],"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> <div class="flex space-x-1 " style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"> <div class="relative colab-dropdown "> <button class=" " type="button"> <img alt="Open In Colab" class="!m-0" src="https://colab.research.google.com/assets/colab-badge.svg"> </button> </div> <div class="relative colab-dropdown "> <button class=" " type="button"> <img alt="Open In Studio Lab" class="!m-0" src="https://studiolab.sagemaker.aws/studiolab.svg"> </button> </div></div> <h1 class="relative group"><a id="semantic-segmentation" 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="#semantic-segmentation"><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>Semantic segmentation</span></h1> <iframe class="w-full xl:w-4/6 h-80" src="https://www.youtube-nocookie.com/embed/dKE8SIt9C-w" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe> <p data-svelte-h="svelte-aptwcv">セマンティックセグメンテーションでは、画像の個々のピクセルにラベルまたはクラスを割り当てます。セグメンテーションにはいくつかのタイプがありますが、セマンティックセグメンテーションの場合、同じオブジェクトの一意のインスタンス間の区別は行われません。両方のオブジェクトに同じラベルが付けられます (たとえば、<code>car-1</code>と<code>car-2</code>の代わりに<code>car</code>)。セマンティックセグメンテーションの一般的な現実世界のアプリケーションには、歩行者や重要な交通情報を識別するための自動運転車のトレーニング、医療画像内の細胞と異常の識別、衛星画像からの環境変化の監視などが含まれます。</p> <p data-svelte-h="svelte-w5jzhi">このガイドでは、次の方法を説明します。</p> <ol data-svelte-h="svelte-1jiwjp7"><li><a href="https://huggingface.co/datasets/scene_parse_150" rel="nofollow">SceneParse150</a> データセットの <a href="https://huggingface.co/docs/transformers/main/en/model_doc/segformer#segformer" rel="nofollow">SegFormer</a> を微調整します。</li> <li>微調整したモデルを推論に使用します。</li></ol> <blockquote class="tip"><p data-svelte-h="svelte-1n9t13i">このタスクと互換性のあるすべてのアーキテクチャとチェックポイントを確認するには、<a href="https://huggingface.co/tasks/image-segmentation" rel="nofollow">タスクページ</a> を確認することをお勧めします。</p></blockquote> <p data-svelte-h="svelte-1lya3k8">始める前に、必要なライブラリがすべてインストールされていることを確認してください。</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="language-bash "><!-- HTML_TAG_START -->pip install -q datasets transformers evaluate<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-193zy02">モデルをアップロードしてコミュニティと共有できるように、Hugging Face アカウントにログインすることをお勧めします。プロンプトが表示されたら、トークンを入力してログインします。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> notebook_login

	<span class="hljs-meta">>>> </span>notebook_login()<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="load-sceneparse150-dataset" 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="#load-sceneparse150-dataset"><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>Load SceneParse150 dataset</span></h2> <p data-svelte-h="svelte-1q1mg88">まず、SceneParse150 データセットの小さいサブセットを 🤗 データセットライブラリから読み込みます。これにより、完全なデータセットのトレーニングにさらに時間を費やす前に、実験してすべてが機能することを確認する機会が得られます。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> datasets <span class="hljs-keyword">import</span> load_dataset

	<span class="hljs-meta">>>> </span>ds = load_dataset(<span class="hljs-string">"scene_parse_150"</span>, split=<span class="hljs-string">"train[:50]"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-m4w3yq"><code>train_test_split</code> メソッドを使用して、データセットの <code>train</code> 分割をトレインセットとテストセットに分割します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>ds = ds.train_test_split(test_size=<span class="hljs-number">0.2</span>)
	<span class="hljs-meta">>>> </span>train_ds = ds[<span class="hljs-string">"train"</span>]
	<span class="hljs-meta">>>> </span>test_ds = ds[<span class="hljs-string">"test"</span>]<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1r6oj5w">次に、例を見てみましょう。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>train_ds[<span class="hljs-number">0</span>]
	{<span class="hljs-string">'image'</span>: <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=512x683 at <span class="hljs-number">0x7F9B0C201F90</span>>,
	<span class="hljs-string">'annotation'</span>: <PIL.PngImagePlugin.PngImageFile image mode=L size=512x683 at <span class="hljs-number">0x7F9B0C201DD0</span>>,
	<span class="hljs-string">'scene_category'</span>: <span class="hljs-number">368</span>}<!-- HTML_TAG_END --></pre></div> <ul data-svelte-h="svelte-1qjmioa"><li><code>image</code>: シーンの PIL イメージ。</li> <li><code>annotation</code>: セグメンテーションマップの PIL イメージ。モデルのターゲットでもあります。</li> <li><code>scene_category</code>: “kitchen”や”office”などの画像シーンを説明するカテゴリ ID。このガイドでは、<code>image</code>と<code>annotation</code>のみが必要になります。どちらも PIL イメージです。</li></ul> <p data-svelte-h="svelte-aqhl7k">また、ラベル ID をラベルクラスにマップする辞書を作成することもできます。これは、後でモデルを設定するときに役立ちます。ハブからマッピングをダウンロードし、<code>id2label</code> および <code>label2id</code> ディクショナリを作成します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> json
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> pathlib <span class="hljs-keyword">import</span> Path
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> hf_hub_download

	<span class="hljs-meta">>>> </span>repo_id = <span class="hljs-string">"huggingface/label-files"</span>
	<span class="hljs-meta">>>> </span>filename = <span class="hljs-string">"ade20k-id2label.json"</span>
	<span class="hljs-meta">>>> </span>id2label = json.loads(Path(hf_hub_download(repo_id, filename, repo_type=<span class="hljs-string">"dataset"</span>)).read_text())
	<span class="hljs-meta">>>> </span>id2label = {<span class="hljs-built_in">int</span>(k): v <span class="hljs-keyword">for</span> k, v <span class="hljs-keyword">in</span> id2label.items()}
	<span class="hljs-meta">>>> </span>label2id = {v: k <span class="hljs-keyword">for</span> k, v <span class="hljs-keyword">in</span> id2label.items()}
	<span class="hljs-meta">>>> </span>num_labels = <span class="hljs-built_in">len</span>(id2label)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="preprocess" 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="#preprocess"><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>Preprocess</span></h2> <p data-svelte-h="svelte-5gp94r">次のステップでは、SegFormer 画像プロセッサをロードして、モデルの画像と注釈を準備します。このデータセットのような一部のデータセットは、バックグラウンドクラスとしてゼロインデックスを使用します。ただし、実際には背景クラスは 150 個のクラスに含まれていないため、<code>do_reduce_labels=True</code>を設定してすべてのラベルから 1 つを引く必要があります。ゼロインデックスは <code>255</code> に置き換えられるため、SegFormer の損失関数によって無視されます。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoImageProcessor

	<span class="hljs-meta">>>> </span>checkpoint = <span class="hljs-string">"nvidia/mit-b0"</span>
	<span class="hljs-meta">>>> </span>image_processor = AutoImageProcessor.from_pretrained(checkpoint, do_reduce_labels=<span class="hljs-literal">True</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-14eknvb">モデルを過学習に対してより堅牢にするために、画像データセットにいくつかのデータ拡張を適用するのが一般的です。このガイドでは、<a href="https://pytorch.org/vision/stable/index.html" rel="nofollow">torchvision</a> の <a href="https://pytorch.org/vision/stable/generated/torchvision.transforms.ColorJitter.html" rel="nofollow"><code>ColorJitter</code></a> 関数を使用します。 ) を使用して画像の色のプロパティをランダムに変更しますが、任意の画像ライブラリを使用することもできます。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> torchvision.transforms <span class="hljs-keyword">import</span> ColorJitter

	<span class="hljs-meta">>>> </span>jitter = 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">0.1</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-166rb6n">次に、モデルの画像と注釈を準備するための 2 つの前処理関数を作成します。これらの関数は、画像を<code>pixel_values</code>に変換し、注釈を<code>labels</code>に変換します。トレーニングセットの場合、画像を画像プロセッサに提供する前に <code>jitter</code> が適用されます。テストセットの場合、テスト中にデータ拡張が適用されないため、画像プロセッサは<code>images</code>を切り取って正規化し、<code>ラベル</code>のみを切り取ります。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">def</span> <span class="hljs-title function_">train_transforms</span>(<span class="hljs-params">example_batch</span>):
	<span class="hljs-meta">... </span> images = [jitter(x) <span class="hljs-keyword">for</span> x <span class="hljs-keyword">in</span> example_batch[<span class="hljs-string">"image"</span>]]
	<span class="hljs-meta">... </span> labels = [x <span class="hljs-keyword">for</span> x <span class="hljs-keyword">in</span> example_batch[<span class="hljs-string">"annotation"</span>]]
	<span class="hljs-meta">... </span> inputs = image_processor(images, labels)
	<span class="hljs-meta">... </span> <span class="hljs-keyword">return</span> inputs


	<span class="hljs-meta">>>> </span><span class="hljs-keyword">def</span> <span class="hljs-title function_">val_transforms</span>(<span class="hljs-params">example_batch</span>):
	<span class="hljs-meta">... </span> images = [x <span class="hljs-keyword">for</span> x <span class="hljs-keyword">in</span> example_batch[<span class="hljs-string">"image"</span>]]
	<span class="hljs-meta">... </span> labels = [x <span class="hljs-keyword">for</span> x <span class="hljs-keyword">in</span> example_batch[<span class="hljs-string">"annotation"</span>]]
	<span class="hljs-meta">... </span> inputs = image_processor(images, labels)
	<span class="hljs-meta">... </span> <span class="hljs-keyword">return</span> inputs<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-wbpxor">データセット全体に<code>jitter</code>を適用するには、🤗 Datasets <code>set_transform</code> 関数を使用します。変換はオンザフライで適用されるため、高速で消費するディスク容量が少なくなります。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>train_ds.set_transform(train_transforms)
	<span class="hljs-meta">>>> </span>test_ds.set_transform(val_transforms)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="evaluate" 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="#evaluate"><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>Evaluate</span></h2> <p data-svelte-h="svelte-cdztvg">トレーニング中にメトリクスを含めると、多くの場合、モデルのパフォーマンスを評価するのに役立ちます。 🤗 <a href="https://huggingface.co/docs/evaluate/index" rel="nofollow">Evaluate</a> ライブラリを使用して、評価メソッドをすばやくロードできます。このタスクでは、<a href="https://huggingface.co/spaces/evaluate-metric/accuracy" rel="nofollow">Mean Intersection over Union</a> (IoU) メトリックをロードします (🤗 Evaluate <a href="https://huggingface.co/docs/evaluate/a_quick_tour" rel="nofollow">クイックツアー</a> を参照して、メトリクスをロードして計算する方法の詳細を確認してください)。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> evaluate

	<span class="hljs-meta">>>> </span>metric = evaluate.load(<span class="hljs-string">"mean_iou"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1d4c85c">次に、メトリクスを <code>compute</code> する関数を作成します。予測を次のように変換する必要があります
	最初にロジットを作成し、次に <code>compute</code> を呼び出す前にラベルのサイズに一致するように再形成します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> numpy <span class="hljs-keyword">as</span> np
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> torch
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> torch <span class="hljs-keyword">import</span> nn

	<span class="hljs-meta">>>> </span><span class="hljs-keyword">def</span> <span class="hljs-title function_">compute_metrics</span>(<span class="hljs-params">eval_pred</span>):
	<span class="hljs-meta">... </span> <span class="hljs-keyword">with</span> torch.no_grad():
	<span class="hljs-meta">... </span> logits, labels = eval_pred
	<span class="hljs-meta">... </span> logits_tensor = torch.from_numpy(logits)
	<span class="hljs-meta">... </span> logits_tensor = nn.functional.interpolate(
	<span class="hljs-meta">... </span> logits_tensor,
	<span class="hljs-meta">... </span> size=labels.shape[-<span class="hljs-number">2</span>:],
	<span class="hljs-meta">... </span> mode=<span class="hljs-string">"bilinear"</span>,
	<span class="hljs-meta">... </span> align_corners=<span class="hljs-literal">False</span>,
	<span class="hljs-meta">... </span> ).argmax(dim=<span class="hljs-number">1</span>)

	<span class="hljs-meta">... </span> pred_labels = logits_tensor.detach().cpu().numpy()
	<span class="hljs-meta">... </span> metrics = metric.compute(
	<span class="hljs-meta">... </span> predictions=pred_labels,
	<span class="hljs-meta">... </span> references=labels,
	<span class="hljs-meta">... </span> num_labels=num_labels,
	<span class="hljs-meta">... </span> ignore_index=<span class="hljs-number">255</span>,
	<span class="hljs-meta">... </span> reduce_labels=<span class="hljs-literal">False</span>,
	<span class="hljs-meta">... </span> )
	<span class="hljs-meta">... </span> <span class="hljs-keyword">for</span> key, value <span class="hljs-keyword">in</span> metrics.items():
	<span class="hljs-meta">... </span> <span class="hljs-keyword">if</span> <span class="hljs-built_in">type</span>(value) <span class="hljs-keyword">is</span> np.ndarray:
	<span class="hljs-meta">... </span> metrics[key] = value.tolist()
	<span class="hljs-meta">... </span> <span class="hljs-keyword">return</span> metrics<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-18cw5xr">これで<code>compute_metrics</code>関数の準備が整いました。トレーニングをセットアップするときにこの関数に戻ります。</p> <h2 class="relative group"><a id="train" 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="#train"><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>Train</span></h2> <blockquote class="tip"><p data-svelte-h="svelte-1m3nznm"><a href="/docs/transformers/main/ja/main_classes/trainer#transformers.Trainer">Trainer</a> を使用したモデルの微調整に慣れていない場合は、<a href="../training#finetune-with-trainer">ここ</a> の基本的なチュートリアルをご覧ください。</p></blockquote> <p data-svelte-h="svelte-2ojowm">これでモデルのトレーニングを開始する準備が整いました。 <a href="/docs/transformers/main/ja/model_doc/auto#transformers.AutoModelForSemanticSegmentation">AutoModelForSemanticSegmentation</a> を使用して SegFormer をロードし、ラベル ID とラベルクラス間のマッピングをモデルに渡します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoModelForSemanticSegmentation, TrainingArguments, Trainer

	<span class="hljs-meta">>>> </span>model = AutoModelForSemanticSegmentation.from_pretrained(checkpoint, id2label=id2label, label2id=label2id)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1j8bgyv">この時点で残っている手順は次の 3 つだけです。</p> <ol data-svelte-h="svelte-10zi60h"><li><a href="/docs/transformers/main/ja/main_classes/trainer#transformers.TrainingArguments">TrainingArguments</a> でトレーニングハイパーパラメータを定義します。 <code>image</code> 列が削除されるため、未使用の列を削除しないことが重要です。 <code>image</code> 列がないと、<code>pixel_values</code> を作成できません。この動作を防ぐには、<code>remove_unused_columns=False</code>を設定してください。他に必要なパラメータは、モデルの保存場所を指定する <code>output_dir</code> だけです。 <code>push_to_hub=True</code>を設定して、このモデルをハブにプッシュします (モデルをアップロードするには、Hugging Face にサインインする必要があります)。各エポックの終了時に、<a href="/docs/transformers/main/ja/main_classes/trainer#transformers.Trainer">Trainer</a> は IoU メトリックを評価し、トレーニングチェックポイントを保存します。</li> <li>トレーニング引数を、モデル、データセット、トークナイザー、データ照合器、および <code>compute_metrics</code> 関数とともに <a href="/docs/transformers/main/ja/main_classes/trainer#transformers.Trainer">Trainer</a> に渡します。</li> <li><a href="/docs/transformers/main/ja/main_classes/trainer#transformers.Trainer.train">train()</a> を呼び出してモデルを微調整します。</li></ol> <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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>training_args = TrainingArguments(
	<span class="hljs-meta">... </span> output_dir=<span class="hljs-string">"segformer-b0-scene-parse-150"</span>,
	<span class="hljs-meta">... </span> learning_rate=<span class="hljs-number">6e-5</span>,
	<span class="hljs-meta">... </span> num_train_epochs=<span class="hljs-number">50</span>,
	<span class="hljs-meta">... </span> per_device_train_batch_size=<span class="hljs-number">2</span>,
	<span class="hljs-meta">... </span> per_device_eval_batch_size=<span class="hljs-number">2</span>,
	<span class="hljs-meta">... </span> save_total_limit=<span class="hljs-number">3</span>,
	<span class="hljs-meta">... </span> eval_strategy=<span class="hljs-string">"steps"</span>,
	<span class="hljs-meta">... </span> save_strategy=<span class="hljs-string">"steps"</span>,
	<span class="hljs-meta">... </span> save_steps=<span class="hljs-number">20</span>,
	<span class="hljs-meta">... </span> eval_steps=<span class="hljs-number">20</span>,
	<span class="hljs-meta">... </span> logging_steps=<span class="hljs-number">1</span>,
	<span class="hljs-meta">... </span> eval_accumulation_steps=<span class="hljs-number">5</span>,
	<span class="hljs-meta">... </span> remove_unused_columns=<span class="hljs-literal">False</span>,
	<span class="hljs-meta">... </span> push_to_hub=<span class="hljs-literal">True</span>,
	<span class="hljs-meta">... </span>)

	<span class="hljs-meta">>>> </span>trainer = Trainer(
	<span class="hljs-meta">... </span> model=model,
	<span class="hljs-meta">... </span> args=training_args,
	<span class="hljs-meta">... </span> train_dataset=train_ds,
	<span class="hljs-meta">... </span> eval_dataset=test_ds,
	<span class="hljs-meta">... </span> compute_metrics=compute_metrics,
	<span class="hljs-meta">... </span>)

	<span class="hljs-meta">>>> </span>trainer.train()<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-ngexm3">トレーニングが完了したら、 <a href="/docs/transformers/main/ja/main_classes/trainer#transformers.Trainer.push_to_hub">push_to_hub()</a> メソッドを使用してモデルをハブに共有し、誰もがモデルを使用できるようにします。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>trainer.push_to_hub()<!-- HTML_TAG_END --></pre></div> <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-cyrfc8">モデルを微調整したので、それを推論に使用できるようになりました。</p> <p data-svelte-h="svelte-21b5vl">推論のために画像をロードします。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>image = ds[<span class="hljs-number">0</span>][<span class="hljs-string">"image"</span>]
	<span class="hljs-meta">>>> </span>image<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-11jfm1f"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/semantic-seg-image.png" alt="Image of bedroom"></div> <p data-svelte-h="svelte-1bvi7mu">推論用に微調整されたモデルを試す最も簡単な方法は、それを <a href="/docs/transformers/main/ja/main_classes/pipelines#transformers.pipeline">pipeline()</a> で使用することです。モデルを使用して画像セグメンテーション用の <code>pipeline</code>をインスタンス化し、それに画像を渡します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> pipeline

	<span class="hljs-meta">>>> </span>segmenter = pipeline(<span class="hljs-string">"image-segmentation"</span>, model=<span class="hljs-string">"my_awesome_seg_model"</span>)
	<span class="hljs-meta">>>> </span>segmenter(image)
	[{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'wall'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062690</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'sky'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062A50</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'floor'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062B50</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'ceiling'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062A10</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'bed '</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062E90</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'windowpane'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062390</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'cabinet'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062550</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'chair'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062D90</span>>},
	{<span class="hljs-string">'score'</span>: <span class="hljs-literal">None</span>,
	<span class="hljs-string">'label'</span>: <span class="hljs-string">'armchair'</span>,
	<span class="hljs-string">'mask'</span>: <PIL.Image.Image image mode=L size=640x427 at <span class="hljs-number">0x7FD5B2062E10</span>>}]<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-fs04zi">必要に応じて、<code>pipeline</code>の結果を手動で複製することもできます。画像を画像プロセッサで処理し、<code>pixel_values</code> を GPU に配置します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>device = torch.device(<span class="hljs-string">"cuda"</span> <span class="hljs-keyword">if</span> torch.cuda.is_available() <span class="hljs-keyword">else</span> <span class="hljs-string">"cpu"</span>) <span class="hljs-comment"># use GPU if available, otherwise use a CPU</span>
	<span class="hljs-meta">>>> </span>encoding = image_processor(image, return_tensors=<span class="hljs-string">"pt"</span>)
	<span class="hljs-meta">>>> </span>pixel_values = encoding.pixel_values.to(device)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1rvunpz">入力をモデルに渡し、<code>logits</code>を返します。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>outputs = model(pixel_values=pixel_values)
	<span class="hljs-meta">>>> </span>logits = outputs.logits.cpu()<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-oolpf2">次に、ロジットを元の画像サイズに再スケールします。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span>upsampled_logits = nn.functional.interpolate(
	<span class="hljs-meta">... </span> logits,
	<span class="hljs-meta">... </span> size=image.size[::-<span class="hljs-number">1</span>],
	<span class="hljs-meta">... </span> mode=<span class="hljs-string">"bilinear"</span>,
	<span class="hljs-meta">... </span> align_corners=<span class="hljs-literal">False</span>,
	<span class="hljs-meta">... </span>)

	<span class="hljs-meta">>>> </span>pred_seg = upsampled_logits.argmax(dim=<span class="hljs-number">1</span>)[<span class="hljs-number">0</span>]<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-e1i4a5">結果を視覚化するには、<a href="https://github.com/tensorflow/models/blob/3f1ca33afe3c1631b733ea7e40c294273b9e406d/research/deeplab/utils/get_dataset_colormap.py#L51" rel="nofollow">データセットカラーパレット</a> を、それぞれをマップする <code>ade_palette()</code> としてロードします。クラスを RGB 値に変換します。次に、画像と予測されたセグメンテーションマップを組み合わせてプロットできます。</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="language-py "><!-- HTML_TAG_START --><span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> matplotlib.pyplot <span class="hljs-keyword">as</span> plt
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">import</span> numpy <span class="hljs-keyword">as</span> np

	<span class="hljs-meta">>>> </span>color_seg = np.zeros((pred_seg.shape[<span class="hljs-number">0</span>], pred_seg.shape[<span class="hljs-number">1</span>], <span class="hljs-number">3</span>), dtype=np.uint8)
	<span class="hljs-meta">>>> </span>palette = np.array(ade_palette())
	<span class="hljs-meta">>>> </span><span class="hljs-keyword">for</span> label, color <span class="hljs-keyword">in</span> <span class="hljs-built_in">enumerate</span>(palette):
	<span class="hljs-meta">... </span> color_seg[pred_seg == label, :] = color
	<span class="hljs-meta">>>> </span>color_seg = color_seg[..., ::-<span class="hljs-number">1</span>] <span class="hljs-comment"># convert to BGR</span>

	<span class="hljs-meta">>>> </span>img = np.array(image) * <span class="hljs-number">0.5</span> + color_seg * <span class="hljs-number">0.5</span> <span class="hljs-comment"># plot the image with the segmentation map</span>
	<span class="hljs-meta">>>> </span>img = img.astype(np.uint8)

	<span class="hljs-meta">>>> </span>plt.figure(figsize=(<span class="hljs-number">15</span>, <span class="hljs-number">10</span>))
	<span class="hljs-meta">>>> </span>plt.imshow(img)
	<span class="hljs-meta">>>> </span>plt.show()<!-- HTML_TAG_END --></pre></div> <div class="flex justify-center" data-svelte-h="svelte-nsecok"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/semantic-seg-preds.png" alt="Image of bedroom overlaid with segmentation map"></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/ja/tasks/semantic_segmentation.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_1e3a2vv = {
	assets: "/docs/transformers/main/ja",
	base: "/docs/transformers/main/ja",
	env: {}
	};

	const element = document.currentScript.parentElement;

	const data = [null,null];

	Promise.all([
	import("/docs/transformers/main/ja/_app/immutable/entry/start.b596590e.js"),
	import("/docs/transformers/main/ja/_app/immutable/entry/app.ede2e8a7.js")
	]).then(([kit, app]) => {
	kit.start(app, element, {
	node_ids: [0, 141],
	data,
	form: null,
	error: null
	});
	});
	}
	</script>

Xet Storage Details

Size:: 69.3 kB
Xet hash:: 66a811651923cf1d4e064e791c218c469422d21e4d05afb242f500ca0eeeb25f

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