{"id":14608,"date":"2026-05-29T17:00:15","date_gmt":"2026-05-29T08:00:15","guid":{"rendered":"https:\/\/blogs.mathworks.com\/japan-community\/?p=14608"},"modified":"2026-05-29T17:00:15","modified_gmt":"2026-05-29T08:00:15","slug":"from-pytorch-litert-to-c-c-and-cuda-source-code-jp","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/japan-community\/2026\/05\/29\/from-pytorch-litert-to-c-c-and-cuda-source-code-jp\/","title":{"rendered":"PyTorch &#038; LiteRT \u3092 C\/C++\/CUDA \u30b3\u30fc\u30c9\u306b\u81ea\u52d5\u5909\u63db"},"content":{"rendered":"<div class=\"rtcContent\">\n<p><span style=\"color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-size: 14px; white-space-collapse: preserve;\">\u203b\u3053\u306e\u6295\u7a3f\u306f 2026 \u5e74 5 \u6708 22 \u65e5\u306b <\/span><a style=\"font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-size: 14px; white-space-collapse: preserve;\" href=\"https:\/\/blogs.mathworks.com\/deep-learning\/2026\/05\/22\/from-pytorch-litert-to-c-c-and-cuda-source-code\/?from=jp\"> The Artificial Intelligence \u3078 \u6295\u7a3f<\/a><span style=\"color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-size: 14px; white-space-collapse: preserve;\">\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<\/span><\/p>\n<h6>&#8212;<\/h6>\n<div class=\"rtcContent\">\n<div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><\/div>\n<div>\n<h6><\/h6>\n<table style=\"background-color: #e2f0ff;\">\n<tbody>\n<tr>\n<td style=\"vertical-align: middle; padding: 10px;\"><strong>\u30b2\u30b9\u30c8\u30e9\u30a4\u30bf\u30fc: <a href=\"https:\/\/www.linkedin.com\/in\/christoph-stockhammer\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Christoph Stockhammer<\/a><\/strong><\/p>\n<h6><\/h6>\n<p><span style=\"font-weight: bold;\">Christoph Stockhammer <\/span>\u306f MathWorks \u3067 AI \u6d3b\u7528\u3092\u62c5\u5f53\u3059\u308b\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u30a8\u30f3\u30b8\u30cb\u30a2\u3067\u3059\u3002Christoph \u306f\u30df\u30e5\u30f3\u30d8\u30f3\u5de5\u79d1\u5927\u5b66\u3067\u6570\u5b66\u306e\u4fee\u58eb\u53f7\u3092\u53d6\u5f97\u3057\u3066\u3044\u307e\u3059\u3002<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h6><\/h6>\n<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><\/div>\n<\/div>\n<div><\/div>\n<div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">PyTorch \u30e2\u30c7\u30eb\u3092\u7d44\u307f\u8fbc\u307f\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306b\u30c7\u30d7\u30ed\u30a4\u3057\u3088\u3046\u3068\u3057\u305f\u3053\u3068\u304c\u3042\u308b\u306a\u3089\u7d0d\u5f97\u3057\u3066\u3044\u305f\u3060\u3051\u308b\u3068\u601d\u3044\u307e\u3059\u3002\u30e2\u30c7\u30eb\u305d\u306e\u3082\u306e\u306f\u6226\u3044\u306e\u534a\u5206\u306b\u3059\u304e\u307e\u305b\u3093\u3002\u672c\u5f53\u306e\u8ab2\u984c\u306f\u3001\u30e9\u30f3\u30bf\u30a4\u30e0\u3001\u5171\u6709\u30e9\u30a4\u30d6\u30e9\u30ea\u3001\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u56fa\u6709\u306e\u30d3\u30eb\u30c9\u3001\u305d\u3057\u3066\u30bf\u30fc\u30b2\u30c3\u30c8\u30c7\u30d0\u30a4\u30b9\u306b\u5408\u3046\u30c4\u30fc\u30eb\u30c1\u30a7\u30fc\u30f3\u304c\u5fc5\u8981\u306b\u306a\u3063\u305f\u3068\u304d\u306b\u59cb\u307e\u308a\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">MATLAB R2026a \u304b\u3089\u4ee3\u308f\u308a\u3068\u306a\u308b\u65b9\u6cd5\u304c\u5229\u7528\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002\u30c7\u30d0\u30a4\u30b9\u306b\u30e9\u30f3\u30bf\u30a4\u30e0\u3092\u8f09\u305b\u308b\u4ee3\u308f\u308a\u306b\u3001<strong>PyTorch ExportedProgram<\/strong> \u3068 <strong>LiteRT<\/strong> \u30e2\u30c7\u30eb\u304b\u3089 <strong>\u30b9\u30bf\u30f3\u30c9\u30a2\u30ed\u30f3\u306e C\/C++\uff08\u304a\u3088\u3073 CUDA\u00ae\uff09\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9<\/strong> \u3092\u76f4\u63a5\u751f\u6210\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308a\u307e\u3057\u305f\u3002\u30a4\u30f3\u30bf\u30fc\u30d7\u30ea\u30bf\u306f\u4e0d\u8981\u3067\u3059\u3002\u63a8\u8ad6\u30a8\u30f3\u30b8\u30f3\u3082\u4e0d\u8981\u3067\u3059\u3002MATLAB \u3078\u306e\u30a4\u30f3\u30dd\u30fc\u30c8\u51e6\u7406\u3082\u4e0d\u8981\u3067\u3059\u3002\u65e2\u5b58\u306e\u30c4\u30fc\u30eb\u30c1\u30a7\u30fc\u30f3\u3067\u30b3\u30f3\u30d1\u30a4\u30eb\u3067\u304d\u308b\u3001\u53ef\u8aad\u6027\u3068\u79fb\u690d\u6027\u306e\u3042\u308b\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u3060\u3051\u3067\u6e08\u307f\u307e\u3059\u3002<\/div>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">\u306a\u305c\u30b9\u30bf\u30f3\u30c9\u30a2\u30ed\u30f3\u30b3\u30fc\u30c9\u751f\u6210\u306a\u306e\u304b?<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u7d44\u307f\u8fbc\u307f\u30bf\u30fc\u30b2\u30c3\u30c8\u306f\u3001\u30de\u30a4\u30b3\u30f3\u304b\u3089 Raspberry Pi \u3084 NVIDIA\u00ae Jetson \u306e\u3088\u3046\u306a\u30c7\u30d0\u30a4\u30b9\u307e\u3067\u3001\u4e88\u6e2c\u53ef\u80fd\u6027\u3092\u975e\u5e38\u306b\u91cd\u8996\u3057\u307e\u3059\u3002\u30e1\u30e2\u30ea\u4f7f\u7528\u91cf\u3001\u8d77\u52d5\u6642\u9593\u3001\u30d0\u30a4\u30ca\u30ea\u30b5\u30a4\u30ba\u306f\u3001\u67d4\u8edf\u6027\u3088\u308a\u91cd\u8981\u306b\u306a\u308b\u3053\u3068\u304c\u3088\u304f\u3042\u308a\u307e\u3059\u3002LiteRT\uff08for Microcontrollers\uff09\u3084 ONNX Runtime \u306e\u3088\u3046\u306a\u30e9\u30f3\u30bf\u30a4\u30e0\u30d9\u30fc\u30b9\u306e\u30bd\u30ea\u30e5\u30fc\u30b7\u30e7\u30f3\u306f\u975e\u5e38\u306b\u512a\u308c\u3066\u3044\u307e\u3059\u304c\u3001\u305d\u308c\u3067\u3082\u7d14\u7c8b\u306a\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u4ee5\u4e0a\u306e\u51e6\u7406\u80fd\u529b\u3068\u30e1\u30e2\u30ea\u4f7f\u7528\u91cf\u3092\u5fc5\u8981\u3068\u3059\u308b\u4f9d\u5b58\u95a2\u4fc2\u3084\u62bd\u8c61\u5316\u30ec\u30a4\u30e4\u30fc\u3092\u6301\u3061\u8fbc\u307f\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u30b9\u30bf\u30f3\u30c9\u30a2\u30ed\u30f3\u30b3\u30fc\u30c9\u751f\u6210\u306f\u305d\u306e\u30ec\u30a4\u30e4\u30fc\u3092\u5b8c\u5168\u306b\u53d6\u308a\u9664\u304d\u307e\u3059\u3002\u751f\u6210\u3055\u308c\u308b\u30b3\u30fc\u30c9\u306b\u306f\u3001\u30e2\u30c7\u30eb\u304c\u5fc5\u8981\u3068\u3059\u308b\u3082\u306e\u3060\u3051\u3001\u3064\u307e\u308a\u30eb\u30fc\u30d7\u3001\u6f14\u7b97\u3001\u30c7\u30fc\u30bf\u3060\u3051\u304c\u542b\u307e\u308c\u307e\u3059\u3002\u305d\u306e\u305f\u3081\u3001\u89e3\u6790\u3001\u30c7\u30d0\u30c3\u30b0\u3001\u8a8d\u8a3c\u3001\u65e2\u5b58\u306e\u7d44\u307f\u8fbc\u307f\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u3068\u306e\u7d71\u5408\u304c\u5bb9\u6613\u306b\u306a\u308a\u307e\u3059\u3002<\/div>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">\u821e\u53f0\u88cf: \u79d8\u8a23\u3068\u306a\u308b MLIR<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u306e\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u652f\u3048\u308b\u4e2d\u6838\u6280\u8853\u306f <strong>MLIR\uff08Multi-Level Intermediate Representation\uff09<\/strong> \u3067\u3059\u3002MATLAB Coder \u306f PyTorch ExportedProgram \u3068 LiteRT \u30e2\u30c7\u30eb\u3092 <a href=\"https:\/\/mlir.llvm.org\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">MLIR<\/a> \u306b\u843d\u3068\u3057\u8fbc\u307f\u3001\u30b0\u30e9\u30d5\u69cb\u9020\u3068\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u3092\u610f\u8b58\u3057\u305f\u4e00\u9023\u306e\u6700\u9069\u5316\u3092\u9069\u7528\u3057\u305f\u3046\u3048\u3067\u3001C\/C++ \u307e\u305f\u306f CUDA \u306e\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3057\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u6700\u9069\u5316\u304c IR \u30ec\u30d9\u30eb\u3067\u884c\u308f\u308c\u308b\u305f\u3081\u3001\u751f\u6210\u30b3\u30fc\u30c9\u306f\u6b21\u306e\u5229\u70b9\u3092\u6d3b\u7528\u3067\u304d\u307e\u3059\u3002<\/div>\n<ul style=\"margin: 10px 0px 20px; padding-left: 0px; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-size: 14px;\">\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\">\u30e1\u30e2\u30ea\u30c8\u30e9\u30d5\u30a3\u30c3\u30af\u3092\u6e1b\u3089\u3059\u305f\u3081\u306e\u6f14\u7b97\u5b50\u878d\u5408<\/li>\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\">\u30de\u30eb\u30c1\u30b3\u30a2 CPU \u4e0a\u3067 OpenMP \u3092\u4f7f\u3063\u305f\u4e26\u5217\u5b9f\u884c<\/li>\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\">\u30d9\u30af\u30c8\u30eb\u5316\uff08\u305f\u3068\u3048\u3070 ARM\u00ae Neon \u3084 Intel\u00ae AVX\uff09<\/li>\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\">\u5229\u7528\u53ef\u80fd\u306a\u5834\u5408\u306e\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u56fa\u6709\u30ab\u30fc\u30cd\u30eb<\/li>\n<\/ul>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u305d\u306e\u7d50\u679c\u3001<strong>\u79fb\u690d\u6027\u3068\u52b9\u7387\u6027<\/strong>\u3092\u4e21\u7acb\u3067\u304d\u308b\u30b3\u30fc\u30c9\u306b\u306a\u308a\u307e\u3059\u3002<\/div>\n<\/div>\n<div><\/div>\n<div>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">PyTorch \u3068 LiteRT \u306e\u76f4\u63a5\u30b5\u30dd\u30fc\u30c8<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u306e\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306f 2 \u3064\u306e\u5165\u529b\u5f62\u5f0f\u3092\u30b5\u30dd\u30fc\u30c8\u3057\u307e\u3059\u3002<\/div>\n<ul style=\"margin: 10px 0px 20px; padding-left: 0px; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-size: 14px;\">\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\"><strong>PyTorch ExportedProgram<\/strong>: PyTorch \u30e2\u30c7\u30eb\u3092\u30af\u30ea\u30fc\u30f3\u304b\u3064\u5b89\u5b9a\u3057\u3066\u8868\u73fe\u3059\u308b\u305f\u3081\u306e\u5f62\u5f0f<\/li>\n<li style=\"margin-left: 56px; line-height: 21px; min-height: 0px; text-align: left; white-space: pre-wrap;\"><strong>LiteRT<\/strong>: TensorFlow \u7531\u6765\u3067\u3082 PyTorch \u30e2\u30c7\u30eb\u304b\u3089\u5909\u63db\u3057\u305f\u3082\u306e\u3067\u3082\u5bfe\u5fdc<\/li>\n<\/ul>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">\u4f8b: \u591a\u5c64\u30d1\u30fc\u30bb\u30d7\u30c8\u30ed\u30f3 (MLP)<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u5168\u7d50\u5408\u5c64\u306b\u57fa\u3065\u304f\u30d5\u30a3\u30fc\u30c9\u30d5\u30a9\u30ef\u30fc\u30c9\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306f\u3001\u6700\u3082\u4e00\u822c\u7684\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306e 1 \u3064\u3067\u3042\u308a\u3001\u5165\u9580\u7528\u306e\u4f8b\u3068\u3057\u3066\u6700\u9069\u3067\u3059\u3002\u305d\u3053\u3067 1 \u3064\u898b\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">torch \u3067\u306f\u3001\u3053\u306e\u3088\u3046\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306f\u6570\u884c\u306e\u30b3\u30fc\u30c9\u3067\u5b9a\u7fa9\u3067\u304d\u307e\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto;\"><code>self.net = nn.Sequential(\r\n    nn.Linear(in_features, hidden1),\r\n    nn.ReLU(),\r\n    nn.Linear(hidden1, hidden2),\r\n    nn.ReLU(),\r\n    nn.Linear(hidden2, out_features)\r\n)<\/code><\/pre>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u901a\u5e38\u306a\u3089\u5b9f\u30c7\u30fc\u30bf\u3067\u30e2\u30c7\u30eb\u3092\u5b66\u7fd2\u3055\u305b\u307e\u3059\u304c\u3001\u3053\u3053\u3067\u306e\u76ee\u7684\u306b\u306f\u3001\u5143\u306e\uff08\u30e9\u30f3\u30c0\u30e0\u521d\u671f\u5316\u3055\u308c\u305f\uff09\u91cd\u307f\u3068\u30d0\u30a4\u30a2\u30b9\u3092\u4fdd\u6301\u3057\u305f\u307e\u307e\u30e2\u30c7\u30eb\u3092\u30c7\u30a3\u30b9\u30af\u306b\u66f8\u304d\u51fa\u3059\u3060\u3051\u3067\u5341\u5206\u3067\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto;\"><code>example_inputs = (torch.randn(batch_size, in_features),)\r\nexported_program = torch.export.export(model, example_inputs)\r\nout_file = \"three_layer_mlp.pt2\"\r\ntorch.export.save(exported_program, out_file)<\/code><\/pre>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u308c\u306b\u3088\u308a &#8220;three_layer_mlp.pt2&#8221; \u3068\u3044\u3046\u30d5\u30a1\u30a4\u30eb\u304c\u751f\u6210\u3055\u308c\u3001<a href=\"https:\/\/netron.app\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Netron<\/a> \u306e\u3088\u3046\u306a\u30c4\u30fc\u30eb\u3067\u53ef\u8996\u5316\u3067\u304d\u307e\u3059\u3002\u4e88\u60f3\u3069\u304a\u308a\u3001\u305d\u306e\u4e2d\u306b\u306f 3 \u3064\u306e\u5168\u7d50\u5408\u5c64\uff08&#8221;linear&#8221; \u3068\u8868\u793a\uff09\u304c\u3042\u308a\u3001\u305d\u306e\u9593\u306b relu \u6d3b\u6027\u5316\u304c\u5165\u3063\u3066\u3044\u308b\u3053\u3068\u304c\u78ba\u8a8d\u3067\u304d\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><img decoding=\"async\" class=\"alignnone wp-image-19299\" style=\"vertical-align: baseline; max-width: 200px; height: auto;\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2026\/05\/RLU_upload.png\" alt=\"Netron visualization of three-layer MLP showing linear layers with relu activations\" width=\"200\" \/><\/div>\n<\/div>\n<div><\/div>\n<div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u305d\u308c\u3067\u306f\u3001\u3053\u306e\u30e2\u30c7\u30eb\u3092 MATLAB \u3067\u3082\u8aad\u307f\u8fbc\u3093\u3067\u307f\u307e\u3057\u3087\u3046\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto; color: #333333;\">&gt;&gt; mlp = loadPyTorchExportedProgram('three_layer_mlp.pt2')<\/pre>\n<div style=\"background-color: #ffffff; padding: 10px; border: none; font-family: monospace; font-size: 13px; overflow-x: auto; color: #555555 !important; white-space: normal;\">\n<p>Loading the model. This may take a few minutes.<\/p>\n<p>mlp =<\/p>\n<p>PyTorchExportedProgram contained in three_layer_mlp.pt2:<\/p>\n<p>Input Specifications<br \/>\n______________________________________<\/p>\n<p>Input Name Size Type<br \/>\n_____ _____ ________ ________<\/p>\n<p>1 &#8220;in1&#8221; &#8220;1 x 16&#8221; &#8220;single&#8221;<\/p>\n<p>Output Specifications<br \/>\n_______________________________________<\/p>\n<p>Output Name Size Type<br \/>\n______ ______ _______ ________<\/p>\n<p>1 &#8220;out1&#8221; &#8220;1 x 8&#8221; &#8220;single&#8221;<\/p>\n<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u308c\u306f\u554f\u984c\u306a\u3055\u305d\u3046\u3067\u3059\u3002\u5165\u529b\u7279\u5fb4\u91cf\u306e\u6570\uff0816\uff09\u3068\u51fa\u529b\u7279\u5fb4\u91cf\u306e\u6570\uff088\uff09\u304c\u6b63\u3057\u3044\u3053\u3068\u304c\u308f\u304b\u308a\u307e\u3059\u3002MATLAB \u5185\u304b\u3089\u3044\u304f\u3064\u304b\u63a8\u8ad6\u30c6\u30b9\u30c8\u3082\u5b9f\u884c\u3057\u3066\u304a\u304f\u306e\u304c\u3088\u3044\u3067\u3057\u3087\u3046\u3002\u305f\u3068\u3048\u3070\u3001\u30e9\u30f3\u30c0\u30e0\u30c7\u30fc\u30bf\u3092\u4f7f\u3063\u3066\u30e2\u30c7\u30eb\u306b\u4e88\u6e2c\u3055\u305b\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto; color: #333333;\">&gt;&gt; invoke(mlp, randn(1,16,'single'))<\/pre>\n<div style=\"background-color: #ffffff; padding: 10px; border: none; font-family: monospace; font-size: 13px; overflow-x: auto; color: #555555 !important; white-space: normal;\">\n<p>ans =<\/p>\n<p>1 x 8 single row vector<\/p>\n<p>-0.0442 -0.2474 0.3094 0.1759 -0.0768 0.0293 0.1398 -0.0767<\/p>\n<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u6b21\u306b\u3001\u3053\u306e\u30e2\u30c7\u30eb\u306e\u4e88\u6e2c\u95a2\u6570\u304b\u3089 C \u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3057\u305f\u3044\u3068\u601d\u3044\u307e\u3059\u3002\u305d\u306e\u305f\u3081\u306b\u306f\u3001\u4e0a\u306e\u30b3\u30de\u30f3\u30c9\u3092 MATLAB \u95a2\u6570\u306b\u307e\u3068\u3081\u308b\u3060\u3051\u3067\u6e08\u307f\u307e\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto;\"><code>function predictions = predictModel(inputFeatures)\r\n    mlp = loadPyTorchExportedProgram('three_layer_mlp.pt2');\r\n    predictions = invoke(mlp, inputFeatures);\r\nend<\/code><\/pre>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u6700\u5f8c\u306e\u30b9\u30c6\u30c3\u30d7\u3068\u3057\u3066\u30011 \u3064\u306e MATLAB \u30b3\u30de\u30f3\u30c9\u3067 C \u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u30d5\u30a1\u30a4\u30eb\u3092\u751f\u6210\u3057\u307e\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto; color: #333333;\">&gt;&gt; codegen -c predictModel.m -args {zeros(1,16,'single')}<\/pre>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u79c1\u306e\u30ce\u30fc\u30c8 PC \u306b\u306f Intel i7 \u30d7\u30ed\u30bb\u30c3\u30b5\u304c\u642d\u8f09\u3055\u308c\u3066\u3044\u307e\u3059\u3002\u6027\u80fd\u3092\u6539\u5584\u3059\u308b\u305f\u3081\u306b\u3001\u30b3\u30fc\u30c9\u751f\u6210\u5668\u306f\u79c1\u306e\u30d7\u30ed\u30bb\u30c3\u30b5\u69cb\u6210\u3092\u81ea\u52d5\u7684\u306b\u8a8d\u8b58\u3057\u307e\u3059\uff08\u5fc5\u8981\u306a\u3089\u5225\u306e\u69cb\u6210\u3092\u624b\u52d5\u3067\u6307\u5b9a\u3057\u3066\u4e0a\u66f8\u304d\u3059\u308b\u3053\u3068\u3082\u3067\u304d\u307e\u3059\uff09\u3002\u3053\u306e\u60c5\u5831\u3092\u4f7f\u3063\u3066\u3001\u30b3\u30fc\u30c9\u751f\u6210\u5668\u306f\u5168\u7d50\u5408\u5c64\u306e\u4e2d\u6838\u3068\u306a\u308b\u884c\u5217\u30d9\u30af\u30c8\u30eb\u7a4d\u306e\u8a08\u7b97\u306b\u3001\u79c1\u306e\u30d7\u30ed\u30bb\u30c3\u30b5\u304c\u30b5\u30dd\u30fc\u30c8\u3059\u308b AVX \u547d\u4ee4\u3092\u6d3b\u7528\u3057\u307e\u3059\u3002\u305d\u306e\u305f\u3081\u3001\u5358\u7d14\u306a for \u30eb\u30fc\u30d7\u3067\u306f\u306a\u304f\u3001\u751f\u6210\u3055\u308c\u305f C \u30b3\u30fc\u30c9\u3067\u306f\u6b21\u306e\u3088\u3046\u306a\u8a18\u8ff0\u306b\u306a\u308a\u307e\u3059\u3002<\/div>\n<pre style=\"background-color: #f5f5f5; padding: 10px; border-radius: 4px; font-family: monospace; font-size: 13px; overflow-x: auto;\"><code>c = _mm256_add_ps(c, _mm256_mul_ps(_mm256_loadu_ps(&amp;A[idxA]), b));<\/code><\/pre>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u3053\u3067 _mm256_mul_ps \u306f Intel intrinsic \u3092\u7528\u3044\u305f\u5358\u7cbe\u5ea6\uff08FP32\uff09\u306e packed \u4e57\u7b97\u6f14\u7b97\u3067\u3042\u308a\u3001\u6027\u80fd\u3092\u5927\u304d\u304f\u5411\u4e0a\u3055\u305b\u308b\u53ef\u80fd\u6027\u304c\u3042\u308a\u307e\u3059\u3002\u3053\u308c\u306f\u3001\u751f\u6210\u30b3\u30fc\u30c9\u304c\u5b9f\u884c\u5bfe\u8c61\u306e\u7279\u5b9a\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u5411\u3051\u306b\u6700\u9069\u5316\u3055\u308c\u308b\u65b9\u6cd5\u306e\u4e00\u4f8b\u306b\u3059\u304e\u307e\u305b\u3093\u3002<\/div>\n<\/div>\n<div><\/div>\n<div>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">\u63a8\u8ad6\u3060\u3051\u3067\u306f\u3042\u308a\u307e\u305b\u3093<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u5b9f\u969b\u306e\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u3067\u306f\u3001\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u304c\u5358\u72ec\u3067\u52d5\u4f5c\u3059\u308b\u3053\u3068\u306f\u307b\u3068\u3093\u3069\u3042\u308a\u307e\u305b\u3093\u3002\u524d\u51e6\u7406\u3001\u5f8c\u51e6\u7406\u3001\u305d\u306e\u4ed6\u306e\u5236\u5fa1\u30ed\u30b8\u30c3\u30af\u306a\u3069\u3082\u3001\u307b\u307c\u5fc5\u305a\u542b\u307e\u308c\u307e\u3059\u3002\u81ea\u52d5\u30b3\u30fc\u30c9\u751f\u6210\u306e\u3088\u3044\u70b9\u306e 1 \u3064\u306f\u3001<em>\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u5168\u4f53<\/em>\u3001\u3064\u307e\u308a\u4fe1\u53f7\u51e6\u7406\u3001\u7279\u5fb4\u62bd\u51fa\u3001\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u63a8\u8ad6\u3092\u30011 \u56de\u306e\u6c7a\u5b9a\u7684\u306a\u624b\u9806\u3067\u30b3\u30fc\u30c9\u751f\u6210\u3067\u304d\u308b\u3053\u3068\u3067\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">MATLAB \u30b3\u30fc\u30c9\u3084 Simulink \u30e2\u30c7\u30eb\u3092\u542b\u3080\u30b5\u30f3\u30d7\u30eb\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u304c\u3001\u3053\u306e\u6a5f\u80fd\u3068\u3068\u3082\u306b\u63d0\u4f9b\u3055\u308c\u3066\u3044\u307e\u3059\u3002\u305f\u3068\u3048\u3070\u3001<a href=\"https:\/\/www.mathworks.com\/help\/coder\/ug\/monocular-depth-estimation-using-depth-anything-v2-pytorch-model.html\" target=\"_blank\" rel=\"noopener\">Depth Anything V2 PyTorch Model \u3092\u7528\u3044\u305f\u5358\u773c\u6df1\u5ea6\u63a8\u5b9a<\/a> \u306f\u3001\u81ea\u52d5\u904b\u8ee2\u3084\u30ca\u30d3\u30b2\u30fc\u30b7\u30e7\u30f3\u306e\u3088\u3046\u306a\u7528\u9014\u306b\u5229\u7528\u3067\u304d\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><img decoding=\"async\" class=\"alignnone wp-image-19295\" style=\"vertical-align: baseline; max-width: 800px; height: auto;\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2026\/05\/Depth-Anything-V2-output.png\" alt=\"Monocular depth estimation example showing original image and depth map side by side\" width=\"800\" \/><\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><a href=\"https:\/\/www.mathworks.com\/help\/coder\/ug\/segmentation-and-object-detection-using-yolo-v11-litert-model.html\" target=\"_blank\" rel=\"noopener\">YOLO v11 LiteRT \u30e2\u30c7\u30eb\u3092\u7528\u3044\u305f\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\u3068\u7269\u4f53\u691c\u51fa<\/a> \u306e\u4f8b\u3067\u306f\u3001NVIDIA cuDNN \u3084 TensorRT \u30e9\u30a4\u30d6\u30e9\u30ea\u306b\u4f9d\u5b58\u305b\u305a\u306b\u3001\u753b\u50cf\u306e\u30bb\u30b0\u30e1\u30f3\u30c6\u30fc\u30b7\u30e7\u30f3\u3084\u691c\u51fa\u306e\u305f\u3081\u306b\u7269\u4f53\u3092\u8b58\u5225\u3057\u3066\u8f2a\u90ed\u3092\u63cf\u304f\u65b9\u6cd5\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><img decoding=\"async\" class=\"alignnone size-full wp-image-19297\" style=\"vertical-align: baseline;\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2026\/05\/YOLOv11-output.png\" alt=\"YOLO v11 segmentation and object detection example\" \/><\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u6700\u5f8c\u306b\u3001<a href=\"https:\/\/www.mathworks.com\/help\/coder\/ug\/predict-battery-state-of-charge-using-litert-model.html\" target=\"_blank\" rel=\"noopener\">LiteRT \u30e2\u30c7\u30eb\u3092\u7528\u3044\u305f\u30d0\u30c3\u30c6\u30ea\u30fc\u5145\u96fb\u72b6\u614b\u306e\u4e88\u6e2c<\/a> \u306e\u4f8b\u3067\u306f\u3001\u96fb\u6c17\u81ea\u52d5\u8eca\u3084\u305d\u306e\u4ed6\u306e\u30d0\u30c3\u30c6\u30ea\u30fc\u99c6\u52d5\u30c7\u30d0\u30a4\u30b9\u306e\u30a8\u30cd\u30eb\u30ae\u30fc\u7ba1\u7406\u30b7\u30b9\u30c6\u30e0\u306b\u304a\u3051\u308b\u91cd\u8981\u6307\u6a19\u3067\u3042\u308b\u30d0\u30c3\u30c6\u30ea\u30fc\u306e State of Charge\uff08SOC\uff09\u3092\u4e88\u6e2c\u3059\u308b AI \u30e2\u30c7\u30eb\u3092\u30c7\u30d7\u30ed\u30a4\u3059\u308b\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><img decoding=\"async\" class=\"alignnone size-full wp-image-19294\" style=\"vertical-align: baseline;\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2026\/05\/BSoC-output.png\" alt=\"Predicted BSOC vs Ground truth chart\" \/><\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u4eca\u3067\u306f C \u3084 C++ \u306e\u30bd\u30fc\u30b9\u3092\u5f97\u308b\u624b\u6bb5\u3068\u3057\u3066\u751f\u6210 AI \u30c4\u30fc\u30eb\u3092\u691c\u8a0e\u3059\u308b\u3053\u3068\u3082\u3067\u304d\u308b\u3067\u3057\u3087\u3046\u3002\u4e0a\u3067\u8aac\u660e\u3057\u305f\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306e\u5927\u304d\u306a\u9055\u3044\u306e 1 \u3064\u306f\u3001\u305d\u308c\u304c <em>\u6c7a\u5b9a\u7684<\/em> \u3067\u3042\u308a <em>\u8ffd\u8de1\u53ef\u80fd<\/em> \u3067\u3042\u308b\u3053\u3068\u3067\u3059\u3002\u540c\u3058\u69cb\u6210\u3067 C \u30bd\u30fc\u30b9\u3092 10 \u56de\u751f\u6210\u3059\u308c\u3070\u300110 \u56de\u3068\u3082\u307e\u3063\u305f\u304f\u540c\u3058 C \u30bd\u30fc\u30b9\u304c\u751f\u6210\u3055\u308c\u307e\u3059\u3002\u30d0\u30c3\u30af\u30a8\u30f3\u30c9\u306e\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\u3001\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u3001\u30d7\u30ed\u30f3\u30d7\u30c8\u306b\u306f\u4f9d\u5b58\u3057\u307e\u305b\u3093\u3002<\/div>\n<\/div>\n<div><\/div>\n<div>\n<h2 style=\"margin: 3px 10px 5px 4px; padding: 0px; line-height: 25px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 20px; font-weight: bold; text-align: left;\">\u30e9\u30f3\u30bf\u30a4\u30e0\u306b\u4ee3\u308f\u308b\u5b9f\u7528\u7684\u306a\u9078\u629e\u80a2<\/h2>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u79c1\u305f\u3061\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3067\u306f\u3001\u751f\u6210\u30b3\u30fc\u30c9\u306f\u30e9\u30f3\u30bf\u30a4\u30e0\u30d9\u30fc\u30b9\u306e\u30a2\u30d7\u30ed\u30fc\u30c1\u3068\u304b\u306a\u308a\u8fd1\u3044\u6027\u80fd\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002\u540c\u3058\u304f\u3089\u3044\u91cd\u8981\u306a\u306e\u306f\u3001\u3053\u306e\u30b3\u30fc\u30c9\u304c\u4eba\u9593\u306b\u8aad\u307f\u3084\u3059\u304f\u3001\u8a2d\u5b9a\u53ef\u80fd\u3067\u3001\u5fc5\u8981\u306b\u5fdc\u3058\u3066\u518d\u5165\u53ef\u80fd\u3067\u3042\u308b\u3053\u3068\u3067\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u3053\u308c\u306f\u3001Raspberry Pi 4 \u4e0a\u3067\u3044\u304f\u3064\u304b\u306e\u4ee3\u8868\u7684\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306b\u3064\u3044\u3066\u3001\u81ea\u52d5\u751f\u6210\u30b3\u30fc\u30c9\u3068 LiteRT \u306e\u6027\u80fd\u3092\u6bd4\u8f03\u3057\u305f\u30b0\u30e9\u30d5\u3067\u3059\u3002\u521d\u56de\u30ea\u30ea\u30fc\u30b9\u3067\u306f\u3001\u30e6\u30fc\u30b6\u30fc\u304c\u591a\u7a2e\u591a\u69d8\u306a\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u3092\u4f7f\u3046\u50be\u5411\u3092\u8e0f\u307e\u3048\u3001\u3067\u304d\u308b\u3060\u3051\u591a\u304f\u306e\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u3068\u30ec\u30a4\u30e4\u30fc\u3092\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u3053\u3068\u306b\u91cd\u70b9\u3092\u7f6e\u3044\u3066\u3044\u307e\u3059\u3002\u591a\u6570\u306e\u6700\u9069\u5316\u3092\u8a08\u753b\u3057\u3066\u304a\u308a\u3001\u4eca\u5f8c\u306e\u30ea\u30ea\u30fc\u30b9\u3067\u306f\u30a4\u30f3\u30bf\u30fc\u30d7\u30ea\u30bf\u306e\u6027\u80fd\u306b\u4e26\u3073\u3001\u3055\u3089\u306b\u4e0a\u56de\u308b\u3053\u3068\u3092\u898b\u8fbc\u3093\u3067\u3044\u307e\u3059\u3002<\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\"><img decoding=\"async\" class=\"alignnone wp-image-19296\" style=\"vertical-align: baseline; max-width: 800px; height: auto;\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2026\/05\/Raspberry-Pi-4-benchmark-R2026a.png\" alt=\"Performance comparison chart: Generated Code vs LiteRT C++ interpreter on Raspberry Pi 4 ARM Cortex-A for various network architectures\" width=\"800\" \/><\/div>\n<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif, Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u307e\u3068\u3081\u308b\u3068\u3001\u900f\u660e\u6027\u3001\u79fb\u690d\u6027\u3001\u305d\u3057\u3066\u7d44\u307f\u8fbc\u307f\u30b7\u30b9\u30c6\u30e0\u3068\u306e\u5bc6\u63a5\u306a\u7d71\u5408\u3092\u91cd\u8996\u3059\u308b AI \u30c7\u30d7\u30ed\u30a4\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u63a2\u3057\u3066\u3044\u308b\u306a\u3089\u3001MATLAB Coder \u3092\u7528\u3044\u305f PyTorch \u304a\u3088\u3073 LiteRT \u304b\u3089\u306e\u30b9\u30bf\u30f3\u30c9\u30a2\u30ed\u30f3\u30b3\u30fc\u30c9\u751f\u6210\u306f\u5341\u5206\u306b\u691c\u8a0e\u3059\u308b\u4fa1\u5024\u304c\u3042\u308a\u307e\u3059\u3002<\/div>\n<\/div>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<div class=\"overview-image\"><img src=\"https:\/\/blogs.mathworks.com\/japan-community\/files\/2026\/05\/YOLOv11-output.png\" class=\"img-responsive attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" decoding=\"async\" loading=\"lazy\" \/><\/div>\n<p>\n\u203b\u3053\u306e\u6295\u7a3f\u306f 2026 \u5e74 5 \u6708 22 \u65e5\u306b  The Artificial Intelligence \u3078 \u6295\u7a3f\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<br \/>\n&#8212;<\/p>\n<p>\u30b2\u30b9\u30c8\u30e9\u30a4\u30bf\u30fc: Christoph Stockhammer<\/p>\n<p>Christoph Stockhammer \u306f MathWorks \u3067 AI \u6d3b\u7528\u3092\u62c5\u5f53\u3059\u308b\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u30a8\u30f3\u30b8\u30cb\u30a2\u3067\u3059\u3002Christoph&#8230; <a class=\"read-more\" href=\"https:\/\/blogs.mathworks.com\/japan-community\/2026\/05\/29\/from-pytorch-litert-to-c-c-and-cuda-source-code-jp\/\">read more >><\/a><\/p>\n","protected":false},"author":159,"featured_media":14612,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[182,61],"tags":[],"_links":{"self":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/14608"}],"collection":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/users\/159"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/comments?post=14608"}],"version-history":[{"count":4,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/14608\/revisions"}],"predecessor-version":[{"id":14613,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/14608\/revisions\/14613"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/media\/14612"}],"wp:attachment":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/media?parent=14608"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/categories?post=14608"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/tags?post=14608"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}