{"id":3566,"date":"2025-04-21T02:03:59","date_gmt":"2025-04-20T17:03:59","guid":{"rendered":"https:\/\/www.dogrow.net\/python\/?p=3566"},"modified":"2025-06-23T13:45:34","modified_gmt":"2025-06-23T04:45:34","slug":"blog128-rtx-5070ti%e3%81%a7-pytorch%e3%82%92%e5%8b%95%e3%81%8b%e3%81%99%e3%80%82","status":"publish","type":"post","link":"https:\/\/www.dogrow.net\/python\/blog128-rtx-5070ti%e3%81%a7-pytorch%e3%82%92%e5%8b%95%e3%81%8b%e3%81%99%e3%80%82\/","title":{"rendered":"(128) RTX 5070ti, 5060ti\u3067 PyTorch\u3092\u52d5\u304b\u3059\u3002"},"content":{"rendered":"<h1 class=\"my_h\">\u30101\u3011\u3084\u308a\u305f\u3044\u3053\u3068<\/h1>\n<p>NVIDIA GeForce RTX 5070ti\u3092\u642d\u8f09\u3057\u305f\u30de\u30b7\u30f3\u4e0a\u3067 PyTorch\u3092\u52d5\u304b\u3057\u305f\u3044\u3002<\/p>\n<p>\u8ffd\u8a18\uff082025.06.23\uff09\uff1a RTX 5060ti\u3067\u3082\u307e\u3063\u305f\u304f\u540c\u3058\u624b\u9806\u3067 PyTorch\u304c\u52d5\u3044\u305f\u3002<\/p>\n<h1 class=\"my_h\">\u30102\u3011\u3084\u3063\u3066\u307f\u305f<\/h1>\n<h2 class=\"my_h\">Step 1\/6: NVIDIA\u30c9\u30e9\u30a4\u30d0\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/h2>\n<p>\u904e\u53bb\u8a18\u4e8b <a href=\"https:\/\/www.dogrow.net\/linux\/blog163-ubuntu24-04%e3%81%abnvidia%e3%83%89%e3%83%a9%e3%82%a4%e3%83%90%e3%82%92%e3%82%a4%e3%83%b3%e3%82%b9%e3%83%88%e3%83%bc%e3%83%ab%e3%81%99%e3%82%8b%e3%80%82\/\" target=\"_blank\">(163) Ubuntu 24.04\u306b NVIDIA\u30c9\u30e9\u30a4\u30d0\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/a> \u3092\u53c2\u7167\u306e\u3053\u3068\u3002<\/p>\n<h2 class=\"my_h\">Step 2\/6: CUDA Toolkit\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/h2>\n<p>\u3053\u3061\u3089\u306e\u30b5\u30a4\u30c8\u3067 OS, \u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u6307\u5b9a\u3057\u3066\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\uff06\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<br \/>\n<a href=\"https:\/\/developer.nvidia.com\/cuda-downloads?target_os=Linux&#038;target_arch=x86_64&#038;Distribution=Ubuntu&#038;target_version=24.04&#038;target_type=deb_local\" target=\"_blank\">https:\/\/developer.nvidia.com\/cuda-downloads?target_os=Linux&#038;target_arch=x86_64&#038;Distribution=Ubuntu&#038;target_version=24.04&#038;target_type=deb_local<br \/>\n<img decoding=\"async\" src=\"https:\/\/www.dogrow.net\/python\/wp-content\/uploads\/2025\/04\/Image1-1.jpg\" alt=\"\" class=\"my_add_bs1\" \/><\/a><\/p>\n<p><span class=\"my_fc_crimsonBBig\">\u6ce8\u610f\uff1a<\/span><br \/>\nCUDA Toolkit 12.9\u304c\u516c\u958b\u3055\u308c\u3066\u3044\u308b\u304c\u3001\u672c\u65e5\u6642\u70b9\u3067<br \/>\n<span class=\"my_fc_crimsonBBig\">PyTorch\u306f 12.8\u5bfe\u5fdc\u7248\u307e\u3067\u3057\u304b\u516c\u958b\u3055\u308c\u3066\u3044\u306a\u3044\u306e\u3067 12.8\u3092\u9078\u629e<\/span> \u3059\u308b\u3002<br \/>\n\u904e\u53bb\u30d0\u30fc\u30b8\u30e7\u30f3\u306f <a href=\"https:\/\/developer.nvidia.com\/cuda-toolkit-archive\" target=\"_blank\">Archive of Previous CUDA Releases<\/a> \u304b\u3089\u63a2\u305b\u3070\u3088\u3044\u3002<\/p>\n<p>OS, \u30d0\u30fc\u30b8\u30e7\u30f3\u7b49\u306e\u30d1\u30e9\u30e1\u30fc\u30bf\u30fc\u3092\u9078\u629e\u3059\u308b\u3068\u3001\u753b\u9762\u4e0b\u90e8\u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u624b\u9806\uff08\uff1dBash\u306b\u5165\u529b\u3059\u308b\u30b3\u30de\u30f3\u30c9\uff09\u304c\u8868\u793a\u3055\u308c\u308b\u306e\u3067\u3001\u3053\u306e\u901a\u308a\u306b\u5b9f\u884c\u3059\u308c\u3070\u3088\u3044\u3002<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>wget https:\/\/developer.download.nvidia.com\/compute\/cuda\/repos\/ubuntu2404\/x86_64\/cuda-ubuntu2404.pin<\/span>\r\n$ <span class='my_fc_yellow'>sudo mv cuda-ubuntu2404.pin \/etc\/apt\/preferences.d\/cuda-repository-pin-600<\/span>\r\n$ <span class='my_fc_yellow'>wget https:\/\/developer.download.nvidia.com\/compute\/cuda\/12.8.1\/local_installers\/cuda-repo-ubuntu2404-12-8-local_12.8.1-570.124.06-1_amd64.deb<\/span>\r\n$ <span class='my_fc_yellow'>sudo dpkg -i cuda-repo-ubuntu2404-12-8-local_12.8.1-570.124.06-1_amd64.deb<\/span>\r\n$ <span class='my_fc_yellow'>sudo cp \/var\/cuda-repo-ubuntu2404-12-8-local\/cuda-*-keyring.gpg \/usr\/share\/keyrings\/<\/span>\r\n$ <span class='my_fc_yellow'>sudo apt-get update<\/span>\r\n$ <span class='my_fc_yellow'>sudo apt-get -y install cuda-toolkit-12-8<\/span>\r\n<\/pre>\n<p>\u30d1\u30b9\u3092\u901a\u3059\u3002<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>nano ~\/.bashrc<\/span>\r\n<\/pre>\n<pre class=\"my_pre\">\r\nexport PATH=\"\/usr\/local\/cuda\/bin:$PATH\"\r\nexport LD_LIBRARY_PATH=\"\/usr\/local\/cuda\/lib64:$LD_LIBRARY_PATH\"\r\n<\/pre>\n<p>Shell\u3092\u518d\u8d77\u52d5\u3059\u308b\u3002<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>source .bashrc<\/span>\r\n<\/pre>\n<h2 class=\"my_h\">Step 3\/6: cuDNN\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/h2>\n<p>\u3053\u3061\u3089\u306e\u30b5\u30a4\u30c8\u3067 OS, \u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u6307\u5b9a\u3059\u308b\u3002<br \/>\n \u2192 \u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u30b3\u30de\u30f3\u30c9\u304c\u8868\u793a\u3055\u308c\u308b\u306e\u3067\u3001\u305d\u306e\u901a\u308a\u306b\u5b9f\u884c\u3059\u308b\u3002<br \/>\n<a href=\"https:\/\/developer.nvidia.com\/cudnn\" target=\"_blank\">https:\/\/developer.nvidia.com\/cudnn<br \/>\n<img decoding=\"async\" src=\"https:\/\/www.dogrow.net\/python\/wp-content\/uploads\/2025\/04\/Image1-2.jpg\" alt=\"\" class=\"my_add_bs1\" \/><\/a><\/p>\n<h2 class=\"my_h\">Step 4\/6: TensorRT\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/h2>\n<p>\u3010Q\u3011\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u3066\u3044\u308b CUDA\u30d0\u30fc\u30b8\u30e7\u30f3\u306e\u78ba\u8a8d\u65b9\u6cd5\u306f\uff1f<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>nvcc -V<\/span>\r\nnvcc: NVIDIA (R) Cuda compiler driver\r\nCopyright (c) 2005-2025 NVIDIA Corporation\r\nBuilt on Fri_Feb_21_20:23:50_PST_2025\r\nCuda compilation tools, release 12.8, V12.8.93\r\nBuild cuda_12.8.r12.8\/compiler.35583870_0 \r\n<\/pre>\n<p>\u3010Q\u3011CUDA12.8\u306b\u5bfe\u5fdc\u3059\u308bTensorRT\u306f\uff1f<br \/>\n10.9\u3092\u4f7f\u7528\u3059\u308b\u3002 \u203b2025\u5e744\u670821\u65e5\u6642\u70b9<\/p>\n<p><span class=\"my_fc_crimsonB\">\u203b\u8981\u30e6\u30fc\u30b6\u30fc\u767b\u9332\uff06\u30ed\u30b0\u30a4\u30f3\uff08\u7121\u6599\uff09<\/span><\/p>\n<p>\u6211\u304c\u5bb6\u3067\u306f <span class='my_fc_crimsonB'>Ubuntu24.04 + CUDA 12.8<\/span> \u306e\u7d44\u307f\u5408\u308f\u305b\u306b\u5408\u308f\u305b\u3066\u3001\u4e0b\u8a18\u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\uff06\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<br \/>\n<a href=\"https:\/\/developer.nvidia.com\/tensorrt\/download\/10x\" target=\"_blank\">https:\/\/developer.nvidia.com\/tensorrt\/download\/10x<br \/>\n<img decoding=\"async\" src=\"https:\/\/www.dogrow.net\/python\/wp-content\/uploads\/2025\/04\/i005.jpg\" alt=\"\" class=\"my_add_bs1\" \/><\/a><\/p>\n<p>TensorRT\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u624b\u9806<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>wget https:\/\/developer.nvidia.com\/downloads\/compute\/machine-learning\/tensorrt\/10.9.0\/local_repo\/nv-tensorrt-local-repo-ubuntu2404-10.9.0-cuda-12.8_1.0-1_amd64.deb<\/span>\r\n$ <span class='my_fc_yellow'>sudo dpkg -i nv-tensorrt-local-repo-ubuntu2404-10.9.0-cuda-12.8_1.0-1_amd64.deb<\/span>\r\n$ <span class='my_fc_yellow'>sudo cp \/var\/nv-tensorrt-local-repo-ubuntu2404-10.9.0-cuda-12.8\/*.gpg \/usr\/share\/keyrings\/<\/span>\r\n$ <span class='my_fc_yellow'>sudo apt update<\/span>\r\n$ <span class='my_fc_yellow'>sudo apt install tensorrt python3-libnvinfer<\/span>\r\n<\/pre>\n<p>\u6b63\u3057\u304f\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u305f\u3053\u3068\u3092 Python\u30b7\u30a7\u30eb\u3067\u78ba\u8a8d\u3057\u3066\u307f\u308b\u3002<br \/>\n<pre class=\"my_pre_python\">\n&gt;&gt;&gt; import tensorrt as trt\n&gt;&gt;&gt; trt.__version__\n&#8216;10.9.0.34&#8217;\n<\/pre><\/p>\n<h2 class=\"my_h\">Step 5\/6: PyTorch\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002<\/h2>\n<p>\u25a0PyTorch\u7528\u306e\u4eee\u60f3\u74b0\u5883\u306e\u4f5c\u6210\u3068\u5b9f\u884c <\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>sudo apt install python3-pip<\/span>\r\n$ <span class='my_fc_yellow'>sudo apt install python3-venv <\/span>\r\n<\/pre>\n<p>\u4eee\u60f3\u74b0\u5883\u540d\u306f\u81ea\u7531\u306b\u4ed8\u3051\u3066\u3088\u3044\u3002\u3053\u3053\u3067\u306f myPyTorch \u3068\u3059\u308b\u3002<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n$ <span class='my_fc_yellow'>python3 -m venv myPyTorch<\/span>\r\n$ <span class='my_fc_yellow'>. .\/myPyTorch\/bin\/activate<\/span>\r\n<\/pre>\n<p>\u3053\u3053\u3067\u4e0b\u8a18\u306e\u3088\u3046\u306b PyTorch\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3068<span class='my_fc_redBBig'>\u30c0\u30e1<\/span><\/p>\n<pre class=\"my_pre_bgBlack\">\r\n(myPyTorch) $ <span class='my_fc_yellow'>pip install torch torchvision torchaudio<\/span>\r\n<\/pre>\n<p>Python\u30b7\u30a7\u30eb\u3067 PyTprch\u3092\u4f7f\u3063\u305f\u30d7\u30ed\u30b0\u30e9\u30e0\u3092\u5b9f\u884c\u3059\u308b\u3068\u3001\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u30a8\u30e9\u30fc\u304c\u767a\u751f\u3059\u308b\u3002<\/p>\n<pre class=\"my_pre_python\">\r\n<span class=\"my_fc_red\">\/home\/hoge\/myPyTorch\/lib\/python3.12\/site-packages\/torch\/cuda\/__init__.py:235: UserWarning:\r\nNVIDIA GeForce RTX 5070 Ti with CUDA capability sm_120 is not compatible with the current PyTorch installation.\r\nThe current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_70 sm_75 sm_80 sm_86 sm_90.\r\nIf you want to use the NVIDIA GeForce RTX 5070 Ti GPU with PyTorch, please check the instructions at https:\/\/pytorch.org\/get-started\/locally\/\r\n\r\n  warnings.warn(\r\nTraceback (most recent call last):\r\n  File \"<stdin>\", line 1, in <module>\r\nRuntimeError: CUDA error: no kernel image is available for execution on the device\r\nCUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.\r\nFor debugging consider passing CUDA_LAUNCH_BLOCKING=1\r\nCompile with `TORCH_USE_CUDA_DSA` to enable device-side assertions. <\/span>\r\n<\/pre>\n<p><span class=\"my_fc_crimsonB\">NVIDIA GeForce RTX 5070 Ti GPU\u3092 PyTorch\u3067\u4f7f\u7528\u3057\u305f\u3044\u5834\u5408\u306f <a href=\"https:\/\/pytorch.org\/get-started\/locally\" target=\"_blank\">https:\/\/pytorch.org\/get-started\/locally<\/a> \u3092\u30c1\u30a7\u30c3\u30af\u3057\u3066\u3002<\/span><br \/>\n\u3068\u306e\u3053\u3068\u306a\u306e\u3067\u3001\u6307\u5b9a\u3055\u308c\u305f\u30da\u30fc\u30b8\u3092\u898b\u308c\u3070\u3088\u3044\u3002<\/p>\n<p>\u305d\u306e\u524d\u306b\u3001\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u3057\u307e\u3063\u305f PyTorch\u3092\u30a2\u30f3\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u304a\u304f\u3002<\/p>\n<pre class=\"my_pre_bgBlack\">\r\n(myPyTorch) $ <span class='my_fc_yellow'>pip uninstall torch torchvision torchaudio<\/span>\r\n(myPyTorch) $ <span class='my_fc_yellow'>pip cache purge<\/span>\r\n<\/pre>\n<p>\u518d\u5ea6\u3001PyTorch\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3002\u305f\u3060\u3057\u3001\u4e0a\u8a18\u306e\u901a\u308a PyTorch\u516c\u5f0f\u304c\u516c\u958b\u3057\u3066\u3044\u308b Preview\u7248\uff08RTX 5070ti\u5bfe\u5fdc\uff09\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30e9\u3092\u4f7f\u3046\u3002<br \/>\n<span class=\"my_fc_crimson\">\u5f8c\u65e5\u3001Stable\u7248\u304c\u66f4\u65b0\u3055\u308c\u305f\u3089\u5165\u308c\u66ff\u3048\u3092\u5fd8\u308c\u305a\u306b\u3002<\/span><br \/>\n<span class=\"my_fc_crimsonBBig\">\u8ffd\u8a18: 2025.05.21 CUDA 12.8\u7528 Stable\u7248\u304c\u516c\u958b\u3055\u308c\u305f\u306e\u3067\u3001\u305d\u3061\u3089\u3092\u4f7f\u3046\u3002<\/span><\/p>\n<p>\u6307\u5b9a\u3055\u308c\u305f\u30da\u30fc\u30b8\u3067 CUDA12.8\u3092\u9078\u629e\u3059\u308b\u3068\u3001\u305d\u306e\u4e0b\u306b pip\u30b3\u30de\u30f3\u30c9\u304c\u8868\u793a\u3055\u308c\u308b\u3002\u3053\u306e\u901a\u308a\u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308c\u3070\u3088\u3044\u3002<br \/>\n<a href=\"https:\/\/pytorch.org\/get-started\/locally\/\" target=\"_blank\">https:\/\/pytorch.org\/get-started\/locally\/<br \/>\n<img decoding=\"async\" src=\"https:\/\/www.dogrow.net\/python\/wp-content\/uploads\/2025\/04\/i001.jpg\" alt=\"\" class=\"my_add_bs1\" \/><\/a><\/p>\n<pre class=\"my_pre_bgBlack\">\r\n(myPyTorch) $ <span class='my_fc_yellow'>pip3 install torch torchvision torchaudio --index-url https:\/\/download.pytorch.org\/whl\/cu128<\/span>\r\n<\/pre>\n<p><span class=\"my_fc_crimsonB\">\u8ffd\u8a18: 2025.05.21 \u2191\u2191\u2191 CUDA 12.8\u7528 Stable\u7248\u304c\u516c\u958b\u3055\u308c\u305f\u306e\u3067\u3001\u30b3\u30de\u30f3\u30c9\u3092\u5909\u66f4\u6e08\u307f\u3002<\/span><\/p>\n<h2 class=\"my_h\">Step 6\/6: PyTorch\u3092\u4f7f\u3063\u3066\u307f\u308b\u3002<\/h2>\n<p>Python\u3092\u8d77\u52d5\u3057\u3001Python shell\u4e0a\u3067\u4ee5\u4e0b\u3092\u5b9f\u884c\u3057\u3066\u307f\u308b\u3002<br \/>\n\u5358\u7d14\u306a\u4e8c\u6b21\u5143\u914d\u5217\u306e\u8db3\u3057\u7b97\u3060\u3002<br \/>\n<pre class=\"my_pre_python\">\n&gt;&gt;&gt; import torch\n&gt;&gt;&gt;\n&gt;&gt;&gt; device = torch.device(&#8220;cuda&#8221; if torch.cuda.is_available() else &#8220;cpu&#8221;)\n&gt;&gt;&gt; device\ndevice(type=&#8217;cuda&#8217;)\n&gt;&gt;&gt;\n&gt;&gt;&gt; a = torch.rand(10000, 10000, device=device)\n&gt;&gt;&gt; b = torch.rand(10000, 10000, device=device)\n&gt;&gt;&gt;\n&gt;&gt;&gt; a+b\ntensor([[1.1051, 1.7026, 0.4868,  ..., 1.2196, 0.6159, 1.2123],\n        [1.7712, 0.9288, 1.0838,  ..., 1.1087, 0.8116, 1.3092],\n        [0.1965, 0.6683, 1.2354,  ..., 0.5542, 0.9295, 1.1723],\n        ...,\n        [0.5446, 0.9329, 1.0721,  ..., 0.8439, 0.6026, 1.2240],\n        [0.6722, 1.0614, 0.5205,  ..., 0.3612, 1.4174, 0.6487],\n        [1.1193, 1.4363, 0.9264,  ..., 0.4703, 1.2994, 0.9731]],\n       device=&#8217;cuda:0&#8242;)\n&gt;&gt;&gt;\n<\/pre><\/p>\n<p>OK\u3060\uff01\u6b63\u3057\u304f\u30c6\u30f3\u30bd\u30eb\u8a08\u7b97\u304c\u3067\u304d\u305f\u3002<br \/>\n\u56e0\u307f\u306b\u30c6\u30f3\u30bd\u30eb\uff08tensor\uff09\u3068\u306f\u3001\u30d7\u30ed\u30b0\u30e9\u30e0\u3067\u4f7f\u7528\u3059\u308b\u591a\u6b21\u5143\u914d\u5217\u306e\u3053\u3068\u3002<\/p>\n<p>\u591a\u6b21\u5143\u914d\u5217\u306e\u8a08\u7b97\u304c\u30b7\u30f3\u30d7\u30eb\u306b\u8a18\u8ff0\u3067\u304d\u308b\u306e\u304c\u3001Python, MATLAB, FORTRAN\u306a\u3069\u306e\u6a5f\u68b0\u5b66\u7fd2\u3067\u4f7f\u308f\u308c\u308b\u8a00\u8a9e\u306e\u7279\u5fb4\u3060\u3002<br \/>\n\u521d\u3081\u3066FORTRAN\u306b\u89e6\u308c\u305f\u306e\u304c 1990\u5e74\u3060\u304b\u3089\u3001\u3082\u304635\u5e74\u3082\u524d\u304b\u30fb\u30fb\u30fb<\/p>\n<h1 class=\"my_h\">\u30103\u3011\u8ffd\u52a0\u4f5c\u696d<\/h1>\n<h3 class=\"my_h\">(1) DKMS (Dynamic Kernel Module Support) \u3082\u5165\u308c\u3066\u304a\u304f\u3002<\/h3>\n<p><span class='my_fc_deeppinkBBig'>DKMS<\/span> <span class='my_fc_deeppinkB'>: Dynamic Kernel Module Support<\/span><br \/>\n\u30ab\u30fc\u30cd\u30eb\u304c\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3055\u308c\u308b\u305f\u3073\u306b\u3001\u81ea\u52d5\u3067\u30e2\u30b8\u30e5\u30fc\u30eb\u3092\u518d\u30d3\u30eb\u30c9\uff06\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u304f\u308c\u308b\u3002<br \/>\n<span class='my_fc_crimsonB'>\u203b sudo apt upgrade \u3067\u4f9d\u5b58\u95a2\u4fc2\u304c\u58ca\u308c\u308b\u4e8b\u614b\u3092\u56de\u907f\u3067\u304d\u308b\u3002<\/span><\/p>\n<p><span class='my_fs_big2'>nvidia-dkms-570-open\u3068\u306f\uff1f<\/span><br \/>\nNVIDIA\u306e\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u7248 GPU\u30c9\u30e9\u30a4\u30d0\uff08570\u7cfb\u5217\uff09\u306e\u30ab\u30fc\u30cd\u30eb\u30e2\u30b8\u30e5\u30fc\u30eb \u3092\u81ea\u52d5\u7684\u306b\u30d3\u30eb\u30c9\u30fb\u7ba1\u7406\u3059\u308b\u305f\u3081\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u3002<br \/>\nNVIDIA\u306e\u30c9\u30e9\u30a4\u30d0\u306f\u3001\u30e6\u30fc\u30b6\u7a7a\u9593\u30e9\u30a4\u30d6\u30e9\u30ea\u7fa4\uff08libcuda, libnvidia-ml\u306a\u3069\uff09\u3068\u30ab\u30fc\u30cd\u30eb\u30e2\u30b8\u30e5\u30fc\u30eb\uff08nvidia.ko\uff09\u306e\u4e21\u65b9\u304b\u3089\u6210\u308a\u7acb\u3063\u3066\u3044\u308b\u3002<br \/>\n<span class='my_fc_blue'>nvidia-dkms-570-open<\/span> \u306f\u3001\u3053\u306e\u30ab\u30fc\u30cd\u30eb\u30e2\u30b8\u30e5\u30fc\u30eb\u3092\u30ab\u30fc\u30cd\u30eb\u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u306b\u5408\u308f\u305b\u3066\u81ea\u52d5\u3067\u30d3\u30eb\u30c9\uff06\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u304f\u308c\u308b\u3002<\/p>\n<pre class='my_pre_bgBlack'>\r\n$ <span class='my_fc_yellow'>sudo apt install --reinstall nvidia-dkms-570-open<\/span>\r\n$ <span class='my_fc_yellow'>sudo update-initramfs -u<\/span>\r\n$ <span class='my_fc_yellow'>sudo reboot<\/span>\r\n<\/pre>\n<h1 class=\"my_h\">\u30104\u3011\u5fdc\u7528<\/h1>\n<p>PyTorch\u3092\u4f7f\u3063\u305f\u6a5f\u68b0\u5b66\u7fd2\uff06\u81ea\u52d5\u8a8d\u8b58\u30d7\u30ed\u30b0\u30e9\u30e0\u3092\u52d5\u304b\u3057\u3066\u307f\u305f\u3002<\/p>\n<p><a href=\"https:\/\/www.dogrow.net\/nnet\/blog30-pytorch%e3%81%a7mnist-%e2%98%85rtx5070ti%e3%81%a7%e3%81%aa%e3%81%8b%e3%81%aa%e3%81%8b%e9%ab%98%e9%80%9f%ef%bc%81\/\" target=\"_blank\">(30)\u3010PyTorch\u3067MNIST #1\u3011GeForce RTX 5070ti\u3067\u9ad8\u901f\u5b9f\u884c<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog31-pytorch%e3%81%a7cpu-vs-gpu-%e3%81%a9%e3%81%a1%e3%82%89%e3%81%8c%e9%80%9f%e3%81%84%ef%bc%9f\/\" target=\"_blank\">(31)\u3010PyTorch\u3067MNIST #2\u3011GPU vs CPU \u3069\u3061\u3089\u304c\u901f\u3044\uff1f<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog32-pytorch%e3%81%a7mnist-%e7%90%86%e8%a7%a3%e3%81%ae%e5%8a%a9%e3%81%91%e3%81%ab%e3%81%aa%e3%82%8b%e3%83%9f%e3%83%8b%e3%83%9e%e3%83%a0%e5%ae%9f%e8%a3%85\/\" target=\"_blank\">(32)\u3010PyTorch\u3067MNIST #3\u3011\u7406\u89e3\u306e\u52a9\u3051\u306b\u306a\u308b\u30df\u30cb\u30de\u30e0\u5b9f\u88c5\uff08CPU\u7528\u3001GPU\u7528\uff09<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog33%e3%80%90pytorch%e3%81%a7mnist-4%e3%80%91%e3%83%97%e3%83%ad%e3%82%b0%e3%83%a9%e3%83%a0%e3%81%ae%e4%bf%9d%e5%ae%88%e6%80%a7%e3%82%92%e5%90%91%e4%b8%8a%e3%81%95%e3%81%9b%e3%82%8b%e3%80%82\/\" target=\"_blank\">(33)\u3010PyTorch\u3067MNIST #4\u3011\u30d7\u30ed\u30b0\u30e9\u30e0\u306e\u4fdd\u5b88\u6027\u3092\u5411\u4e0a\u3055\u305b\u308b\u3002<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/34%e3%80%90pytorch%e3%81%a7mnist-5%e3%80%91%e5%ad%a6%e7%bf%92%e6%b8%88%e3%81%bf%e3%83%91%e3%83%a9%e3%83%a1%e3%83%bc%e3%82%bf%e3%83%95%e3%82%a1%e3%82%a4%e3%83%ab%e3%82%92%e6%8c%87%e5%ae%9a%e5%8f%af\/\" target=\"_blank\">(34)\u3010PyTorch\u3067MNIST #5\u3011\u5b66\u7fd2\u6e08\u307f\u30d1\u30e9\u30e1\u30fc\u30bf\u30d5\u30a1\u30a4\u30eb\u3092\u6307\u5b9a\u53ef\u80fd\u306b\u3059\u308b\u3002<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/35%e3%80%90pytorch%e3%81%a7mnist-6%e3%80%91%e4%b8%8d%e6%ad%a3%e8%a7%a3%e7%94%bb%e5%83%8f%e3%81%a8%e5%88%a4%e5%ae%9a%e7%b5%90%e6%9e%9c%e3%82%92jpeg%e7%94%bb%e5%83%8f%e5%87%ba%e5%8a%9b%e3%81%99\/\" target=\"_blank\">(35)\u3010PyTorch\u3067MNIST #6\u3011\u4e0d\u6b63\u89e3\u753b\u50cf\u3068\u5224\u5b9a\u7d50\u679c\u3092JPEG\u753b\u50cf\u51fa\u529b\u3059\u308b\u3002<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog36%e3%80%90pytorch%e3%81%a7mnist-7%e3%80%91%e8%87%aa%e7%ad%86%e7%94%bb%e5%83%8fpng%e3%83%95%e3%82%a1%e3%82%a4%e3%83%ab%e3%82%92%e8%87%aa%e5%8b%95%e8%aa%8d%e8%ad%98%e3%81%95%e3%81%9b%e3%82%8b\/\" target=\"_blank\">(36)\u3010PyTorch\u3067MNIST #7\u3011\u81ea\u7b46\u753b\u50cfPNG\u30d5\u30a1\u30a4\u30eb\u3092\u81ea\u52d5\u8a8d\u8b58\u3055\u305b\u308b\u3002<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog40%e3%80%90pytorch%e3%81%a7mnist-8%e3%80%91web%e3%82%b5%e3%83%bc%e3%83%93%e3%82%b9%e5%8c%96%e3%81%99%e3%82%8b%e3%80%82\/\" target=\"_blank\">(40)\u3010PyTorch\u3067MNIST #8\u3011\u81ea\u52d5\u8a8d\u8b58\u3092Web\u30b5\u30fc\u30d3\u30b9\u5316\u3059\u308b\u3002<\/a><br \/>\n<a href=\"https:\/\/www.dogrow.net\/nnet\/blog38-pytorch%e3%81%a7cifar-10%e3%82%92%e3%82%84%e3%81%a3%e3%81%a6%e3%81%bf%e3%82%8b%e3%80%82\/\" target=\"_blank\">(38)\u3010PyTorch\u3067CIFAR-10 #1\u3011\u81ea\u52d5\u8a8d\u8b58\u3057\u3066\u307f\u308b\u3002<\/a><\/p>\n<h1 class=\"my_h\">\u30105\u3011\u53c2\u8003\u60c5\u5831<\/h1>\n<p><a href=\"https:\/\/pytorch.org\/\" target=\"_blank\">PyTorch\u516c\u5f0f\u30da\u30fc\u30b8<\/a><br \/>\n<a href=\"\" target=\"_blank\"><\/a><\/p>\n<p>\u3068\u3066\u3082\u6709\u76ca\u306a\u60c5\u5831\u3092\u3042\u308a\u304c\u3068\u3046\u3054\u3056\u3044\u307e\u3059\u3002m(_ _)m<br \/>\n<a href=\"https:\/\/qiita.com\/xtrizeShino\/items\/56125ef057c6e88b4302\" target=\"_blank\">https:\/\/qiita.com\/xtrizeShino\/items\/56125ef057c6e88b4302<\/a><\/p>\n<hr class=\"my_hr_bottom\">\n","protected":false},"excerpt":{"rendered":"<p>\u30101\u3011\u3084\u308a\u305f\u3044\u3053\u3068 NVIDIA GeForce RTX 5070ti\u3092\u642d\u8f09\u3057\u305f\u30de\u30b7\u30f3\u4e0a\u3067 PyTorch\u3092\u52d5\u304b\u3057\u305f\u3044\u3002 \u8ffd\u8a18\uff082025.06.23\uff09\uff1a RTX 5060ti\u3067\u3082\u307e\u3063\u305f\u304f\u540c\u3058\u624b\u9806\u3067 PyTorch\u304c\u52d5\u3044\u305f\u2026 <span class=\"read-more\"><a href=\"https:\/\/www.dogrow.net\/python\/blog128-rtx-5070ti%e3%81%a7-pytorch%e3%82%92%e5%8b%95%e3%81%8b%e3%81%99%e3%80%82\/\">\u7d9a\u304d\u3092\u8aad\u3080 &raquo;<\/a><\/span><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[66,65,64],"tags":[],"class_list":["post-3566","post","type-post","status-publish","format-standard","hentry","category-cuda","category-gpu","category-pytorch"],"views":7111,"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/posts\/3566","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/comments?post=3566"}],"version-history":[{"count":82,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/posts\/3566\/revisions"}],"predecessor-version":[{"id":3893,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/posts\/3566\/revisions\/3893"}],"wp:attachment":[{"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/media?parent=3566"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/categories?post=3566"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.dogrow.net\/python\/wp-json\/wp\/v2\/tags?post=3566"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}