Windows 下用 build_win.cmd 直接编译 GPU 版caffe

本文详细介绍了在Windows环境下安装和配置Caffe的全过程,包括CUDA和cuDNN的安装,选择合适的Python版本和IDE,下载并配置Caffe源码,以及编译和调试Caffe的过程。
部署运行你感兴趣的模型镜像

1.安装CUDA

首先我们得从CUDA官网,下载对应版本得CUDA,这里打开默认是CUDA 10,若我们想下载低版本的 CUDA,如下图,选择 Legacy Releases:

这里以 CUDA 8.0 为例:

点击下载,下载完是一个 .exe 文件,只要硬件符合直接双击安装即可。

2.安装cudnn

这里自行下载,需要在 NVIDIA 的cudnn官网 注册一个账号,免费的,以后可能都会用到,所以这里就不给出百度云分享了,可以去官网注册一下。 完成注册登录以后进行下载页面:

根据自己CUDA 的版本选择对应的版本,以 cudnn 5.1 为例:

下载完是一个 .zip 的压缩包,直接解压后是3个子文件夹:

这时候我们找到第1步下载安装的 CUDA 路径,以我的电脑为例:

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0

打开上述文件夹,找到对应于 cudnn 解压出来的 3 个子文件夹

将 cudnn 子文件夹中的文件分别放到上面三个对应名字的文件夹里,这样 cudnn 就配置完成了。

3.前期准备

可以看到caffe官方贴出的windows版安装对安装环境版本的说明:

###注:这里Python的版本只支持python2.7 或者 python3.5

所以我们需要先准备好:

a) VS2015

Tip: 如果电脑安装的是 VS2013, Python只能用2.7版本的,否则编译时会 version 报错

b) anaconda3: 因为现在官网最新下载的anaconda3都是python3.6版本的了,而目前caffe官方只支持到python3.5,这里给出anaconda3-python3.5版本的百度云下载链接。[密码:y05j]

c) anaconda2:直接官网下载安装即可。

###注:b 和 c 任选其一即可,后面根据自己选择的python版本进行对应配置,个人推荐下载anaconda3。

4.下载windows版caffe

Github 官方下载地址:下载

打开界面以后:

如上图所示:我们选择左边的Branch,选择windows然后选择下载就可以了。

5.修改Caffe配置文件

下载压缩包以后解压缩,可以看到:

进入 scripts 文件夹选择build_win.cmd文件,用文本编辑器进行编辑:

在build_win.cmd中根据我们的需求修改我们的配置(主要是69行-100行之间),因为我需要使用pycaffe,所以事先安装了 anaconda3[python3.5版本],这里我将87行改成了PYTHON VERSION=3 ),官方说明caffe的配置python只支持3.5或者2.7版本,所以安装时候需要注意一下,这里将76行, CPU_ONLY 设置成 0,99行 RUN_INSTALL 设置成1

6.编译caffe

修改好build_win.cmd文件后保存退出,然后在caffe-windows文件下打开命令行工具,输入:

.\script\build_win.cmd

开始编译caffe(这里我使用的是Cmder,界面会比windows自带的cmd界面友好一些,而且可以加入右键菜单直接打开,类似于Ubuntu的Terminal,很方便,推荐大家使用,具体安装方法参见我的另一篇博客-Cmder 加入右键菜单),会有下面的输出:

可以看到第一次编译的时候,caffe会下载依赖库到图片中画红线的路径,所下载的依赖包143M左右,但下载过程中也可能会多次出现下载报错的问题,不是很稳定,这个时候我们可以:

1)先ctrl+c结束这个进程,然后以python版本为区分,可以在我分享的百度云文件下载自己对应版本的依赖包。
Python2.7的:libraries_v140_x64_py27_1.1.0.tar.bz2[ 密码:4yf0 ]
Python3.5的:libraries_v140_x64_py35_1.1.0.tar.bz2[ 密码:35jj ]

2)耐心等待它自己下载完。

上面两步任选其一,然后下载完成后直接将压缩包放在上面输出的对应路径下即可,以我的电脑为例:

然后再次输入 .\script\build_win.cmd ,可以看到:

这里下载就变成了开始提取文件,稍加等待,caffe就开始编译了。

最后编译成功会得到:

警告数可以忽略,每次编译可能都会不一样,只要是 0个错误,就说明caffe编译成功了,就可以正常使用了。

###Debug版和Release版通过修改build_win.cmd 第 81 行就可以了。步骤相同,直接编译即可。

编译以后刚才强调的将第99行 RUN INSTALL 设置成1,在build/install/bin 目录下就会生成:

可以看到像caffe.execompute_image_mean.exe 等常用的执行文件就在这里了(设置为0就没有这些文件),C++使用时,添加相应路径即可。

相关博客:

Windows 下用 build_win.cmd 直接编译CPU版caffe

Windows下 Caffe C++接口的调用

Windows下 Pycaffe 的配置与使用

您可能感兴趣的与本文相关的镜像

Python3.10

Python3.10

Conda
Python

Python 是一种高级、解释型、通用的编程语言,以其简洁易读的语法而闻名,适用于广泛的应用,包括Web开发、数据分析、人工智能和自动化脚本

(.venv) PS E:\PyTorch_Build\pytorch\build> # 搜索整个构建目录中的所有构建文件 (.venv) PS E:\PyTorch_Build\pytorch\build> $buildRoot = "E:\PyTorch_Build\pytorch\build" (.venv) PS E:\PyTorch_Build\pytorch\build> $allBuildFiles = Get-ChildItem -Path $buildRoot -Recurse -Include *.pyd, *.dll, *.lib (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> if ($allBuildFiles.Count -eq 0) { >> # 如果没有找到任何构建文件,重新运行构建过程 >> Write-Host "未找到任何构建文件,重新运行构建过程..." >> cd "E:\PyTorch_Build\pytorch" >> python setup.py install >> $allBuildFiles = Get-ChildItem -Path $buildRoot -Recurse -Include *.pyd, *.dll, *.lib >> } (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> # 打印找到的文件 (.venv) PS E:\PyTorch_Build\pytorch\build> Write-Host "找到的构建文件:" 找到的构建文件: (.venv) PS E:\PyTorch_Build\pytorch\build> $allBuildFiles | ForEach-Object { Write-Host " - $($_.FullName)" } - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\charset_normalizer\md__mypyc.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\charset_normalizer\md.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\markupsafe\_speedups.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\lib\npymath.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_multiarray_tests.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_multiarray_tests.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_multiarray_umath.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_multiarray_umath.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_operand_flag_tests.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_operand_flag_tests.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_rational_tests.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_rational_tests.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_simd.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_simd.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_struct_ufunc_tests.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_struct_ufunc_tests.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_umath_tests.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\_core\_umath_tests.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\fft\_pocketfft_umath.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\fft\_pocketfft_umath.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\linalg\_umath_linalg.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\linalg\_umath_linalg.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\linalg\lapack_lite.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\linalg\lapack_lite.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\lib\npyrandom.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_bounded_integers.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_bounded_integers.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_common.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_common.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_generator.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_generator.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_mt19937.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_mt19937.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_pcg64.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_pcg64.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_philox.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_philox.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_sfc64.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\_sfc64.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\bit_generator.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\bit_generator.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\mtrand.cp310-win_amd64.lib - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy\random\mtrand.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy.libs\libscipy_openblas64_-13e2df515630b4a41f92893938845698.dll - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\numpy.libs\msvcp140-263139962577ecda4cd9469ca360a746.dll - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\optree\_C.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\psutil\_psutil_windows.pyd - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\yaml\_yaml.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\bin\c10_cuda.dll - E:\PyTorch_Build\pytorch\build\bin\c10.dll - E:\PyTorch_Build\pytorch\build\bin\caffe2_nvrtc.dll - E:\PyTorch_Build\pytorch\build\bin\nnapi_backend.dll - E:\PyTorch_Build\pytorch\build\bin\shm.dll - E:\PyTorch_Build\pytorch\build\bin\torch_cpu.dll - E:\PyTorch_Build\pytorch\build\bin\torch_cuda.dll - E:\PyTorch_Build\pytorch\build\bin\torch_global_deps.dll - E:\PyTorch_Build\pytorch\build\bin\torch_python.dll - E:\PyTorch_Build\pytorch\build\bin\torch.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-console-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-datetime-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-debug-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-errorhandling-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-file-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-file-l1-2-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-file-l2-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-handle-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-heap-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-interlocked-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-libraryloader-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-localization-l1-2-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-memory-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-namedpipe-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-processenvironment-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-processthreads-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-processthreads-l1-1-1.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-profile-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-rtlsupport-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-string-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-synch-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-synch-l1-2-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-sysinfo-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-timezone-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-core-util-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-conio-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-convert-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-environment-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-filesystem-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-heap-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-locale-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-math-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-multibyte-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-private-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-process-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-runtime-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-stdio-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-string-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-time-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\api-ms-win-crt-utility-l1-1-0.dll - E:\PyTorch_Build\pytorch\build\Dependencies\ClrPhlib.dll - E:\PyTorch_Build\pytorch\build\Dependencies\dbghelp.dll - E:\PyTorch_Build\pytorch\build\Dependencies\DependenciesLib.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Dragablz.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Mono.Cecil.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Mono.Cecil.Mdb.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Mono.Cecil.Pdb.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Mono.Cecil.Rocks.dll - E:\PyTorch_Build\pytorch\build\Dependencies\MSVCP140.dll - E:\PyTorch_Build\pytorch\build\Dependencies\NDesk.Options.dll - E:\PyTorch_Build\pytorch\build\Dependencies\Newtonsoft.Json.dll - E:\PyTorch_Build\pytorch\build\Dependencies\ucrtbase.dll - E:\PyTorch_Build\pytorch\build\Dependencies\VCRUNTIME140_1.dll - E:\PyTorch_Build\pytorch\build\Dependencies\VCRUNTIME140.dll - E:\PyTorch_Build\pytorch\build\functorch\functorch.pyd - E:\PyTorch_Build\pytorch\build\lib\c10_cuda.lib - E:\PyTorch_Build\pytorch\build\lib\c10.lib - E:\PyTorch_Build\pytorch\build\lib\caffe2_nvrtc.lib - E:\PyTorch_Build\pytorch\build\lib\Caffe2_perfkernels_avx2.lib - E:\PyTorch_Build\pytorch\build\lib\cpuinfo_internals.lib - E:\PyTorch_Build\pytorch\build\lib\cpuinfo.lib - E:\PyTorch_Build\pytorch\build\lib\fmt.lib - E:\PyTorch_Build\pytorch\build\lib\functorch.lib - E:\PyTorch_Build\pytorch\build\lib\kineto.lib - E:\PyTorch_Build\pytorch\build\lib\libittnotify.lib - E:\PyTorch_Build\pytorch\build\lib\libjitprofiling.lib - E:\PyTorch_Build\pytorch\build\lib\libprotobuf-lite.lib - E:\PyTorch_Build\pytorch\build\lib\libprotobuf.lib - E:\PyTorch_Build\pytorch\build\lib\libprotoc.lib - E:\PyTorch_Build\pytorch\build\lib\microkernels-all.lib - E:\PyTorch_Build\pytorch\build\lib\microkernels-prod.lib - E:\PyTorch_Build\pytorch\build\lib\mimalloc.lib - E:\PyTorch_Build\pytorch\build\lib\nnapi_backend.lib - E:\PyTorch_Build\pytorch\build\lib\onnx_proto.lib - E:\PyTorch_Build\pytorch\build\lib\onnx.lib - E:\PyTorch_Build\pytorch\build\lib\pthreadpool.lib - E:\PyTorch_Build\pytorch\build\lib\shm.lib - E:\PyTorch_Build\pytorch\build\lib\torch_cpu.lib - E:\PyTorch_Build\pytorch\build\lib\torch_cuda.lib - E:\PyTorch_Build\pytorch\build\lib\torch_python.lib - E:\PyTorch_Build\pytorch\build\lib\torch.lib - E:\PyTorch_Build\pytorch\build\lib\XNNPACK.lib - E:\PyTorch_Build\pytorch\build\lib.win-amd64-cpython-310\functorch\_C.cp310-win_amd64.pyd - E:\PyTorch_Build\pytorch\build\sleef\lib\sleef.lib (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> $torchDir = "E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\torch" (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> foreach ($file in $allBuildFiles) { >> # 创建目标路径的目录结构 >> $relativePath = $file.FullName.Substring($buildRoot.Length) >> $destPath = Join-Path $torchDir $relativePath >> $destDir = [System.IO.Path]::GetDirectoryName($destPath) >> >> if (-not (Test-Path $destDir)) { >> New-Item -ItemType Directory -Path $destDir -Force | Out-Null >> } >> >> # 复制文件 >> Copy-Item -Path $file.FullName -Destination $destPath -Force >> Write-Host "已复制: $($file.Name) -> $relativePath" >> } 已复制: md__mypyc.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\charset_normalizer\md__mypyc.cp310-win_amd64.pyd 已复制: md.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\charset_normalizer\md.cp310-win_amd64.pyd 已复制: _speedups.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\markupsafe\_speedups.cp310-win_amd64.pyd 已复制: npymath.lib -> \.venv\Lib\site-packages\numpy\_core\lib\npymath.lib 已复制: _multiarray_tests.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_multiarray_tests.cp310-win_amd64.lib 已复制: _multiarray_tests.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_multiarray_tests.cp310-win_amd64.pyd 已复制: _multiarray_umath.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_multiarray_umath.cp310-win_amd64.lib 已复制: _multiarray_umath.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_multiarray_umath.cp310-win_amd64.pyd 已复制: _operand_flag_tests.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_operand_flag_tests.cp310-win_amd64.lib 已复制: _operand_flag_tests.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_operand_flag_tests.cp310-win_amd64.pyd 已复制: _rational_tests.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_rational_tests.cp310-win_amd64.lib 已复制: _rational_tests.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_rational_tests.cp310-win_amd64.pyd 已复制: _simd.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_simd.cp310-win_amd64.lib 已复制: _simd.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_simd.cp310-win_amd64.pyd 已复制: _struct_ufunc_tests.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_struct_ufunc_tests.cp310-win_amd64.lib 已复制: _struct_ufunc_tests.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_struct_ufunc_tests.cp310-win_amd64.pyd 已复制: _umath_tests.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\_core\_umath_tests.cp310-win_amd64.lib 已复制: _umath_tests.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\_core\_umath_tests.cp310-win_amd64.pyd 已复制: _pocketfft_umath.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\fft\_pocketfft_umath.cp310-win_amd64.lib 已复制: _pocketfft_umath.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\fft\_pocketfft_umath.cp310-win_amd64.pyd 已复制: _umath_linalg.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\linalg\_umath_linalg.cp310-win_amd64.lib 已复制: _umath_linalg.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\linalg\_umath_linalg.cp310-win_amd64.pyd 已复制: lapack_lite.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\linalg\lapack_lite.cp310-win_amd64.lib 已复制: lapack_lite.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\linalg\lapack_lite.cp310-win_amd64.pyd 已复制: npyrandom.lib -> \.venv\Lib\site-packages\numpy\random\lib\npyrandom.lib 已复制: _bounded_integers.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_bounded_integers.cp310-win_amd64.lib 已复制: _bounded_integers.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_bounded_integers.cp310-win_amd64.pyd 已复制: _common.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_common.cp310-win_amd64.lib 已复制: _common.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_common.cp310-win_amd64.pyd 已复制: _generator.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_generator.cp310-win_amd64.lib 已复制: _generator.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_generator.cp310-win_amd64.pyd 已复制: _mt19937.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_mt19937.cp310-win_amd64.lib 已复制: _mt19937.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_mt19937.cp310-win_amd64.pyd 已复制: _pcg64.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_pcg64.cp310-win_amd64.lib 已复制: _pcg64.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_pcg64.cp310-win_amd64.pyd 已复制: _philox.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_philox.cp310-win_amd64.lib 已复制: _philox.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_philox.cp310-win_amd64.pyd 已复制: _sfc64.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\_sfc64.cp310-win_amd64.lib 已复制: _sfc64.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\_sfc64.cp310-win_amd64.pyd 已复制: bit_generator.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\bit_generator.cp310-win_amd64.lib 已复制: bit_generator.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\bit_generator.cp310-win_amd64.pyd 已复制: mtrand.cp310-win_amd64.lib -> \.venv\Lib\site-packages\numpy\random\mtrand.cp310-win_amd64.lib 已复制: mtrand.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\numpy\random\mtrand.cp310-win_amd64.pyd 已复制: libscipy_openblas64_-13e2df515630b4a41f92893938845698.dll -> \.venv\Lib\site-packages\numpy.libs\libscipy_openblas64_-13e2df515630b4a41f92893938845698.dll 已复制: msvcp140-263139962577ecda4cd9469ca360a746.dll -> \.venv\Lib\site-packages\numpy.libs\msvcp140-263139962577ecda4cd9469ca360a746.dll 已复制: _C.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\optree\_C.cp310-win_amd64.pyd 已复制: _psutil_windows.pyd -> \.venv\Lib\site-packages\psutil\_psutil_windows.pyd 已复制: _yaml.cp310-win_amd64.pyd -> \.venv\Lib\site-packages\yaml\_yaml.cp310-win_amd64.pyd 已复制: c10_cuda.dll -> \bin\c10_cuda.dll 已复制: c10.dll -> \bin\c10.dll 已复制: caffe2_nvrtc.dll -> \bin\caffe2_nvrtc.dll 已复制: nnapi_backend.dll -> \bin\nnapi_backend.dll 已复制: shm.dll -> \bin\shm.dll 已复制: torch_cpu.dll -> \bin\torch_cpu.dll 已复制: torch_cuda.dll -> \bin\torch_cuda.dll 已复制: torch_global_deps.dll -> \bin\torch_global_deps.dll 已复制: torch_python.dll -> \bin\torch_python.dll 已复制: torch.dll -> \bin\torch.dll 已复制: api-ms-win-core-console-l1-1-0.dll -> \Dependencies\api-ms-win-core-console-l1-1-0.dll 已复制: api-ms-win-core-datetime-l1-1-0.dll -> \Dependencies\api-ms-win-core-datetime-l1-1-0.dll 已复制: api-ms-win-core-debug-l1-1-0.dll -> \Dependencies\api-ms-win-core-debug-l1-1-0.dll 已复制: api-ms-win-core-errorhandling-l1-1-0.dll -> \Dependencies\api-ms-win-core-errorhandling-l1-1-0.dll 已复制: api-ms-win-core-file-l1-1-0.dll -> \Dependencies\api-ms-win-core-file-l1-1-0.dll 已复制: api-ms-win-core-file-l1-2-0.dll -> \Dependencies\api-ms-win-core-file-l1-2-0.dll 已复制: api-ms-win-core-file-l2-1-0.dll -> \Dependencies\api-ms-win-core-file-l2-1-0.dll 已复制: api-ms-win-core-handle-l1-1-0.dll -> \Dependencies\api-ms-win-core-handle-l1-1-0.dll 已复制: api-ms-win-core-heap-l1-1-0.dll -> \Dependencies\api-ms-win-core-heap-l1-1-0.dll 已复制: api-ms-win-core-interlocked-l1-1-0.dll -> \Dependencies\api-ms-win-core-interlocked-l1-1-0.dll 已复制: api-ms-win-core-libraryloader-l1-1-0.dll -> \Dependencies\api-ms-win-core-libraryloader-l1-1-0.dll 已复制: api-ms-win-core-localization-l1-2-0.dll -> \Dependencies\api-ms-win-core-localization-l1-2-0.dll 已复制: api-ms-win-core-memory-l1-1-0.dll -> \Dependencies\api-ms-win-core-memory-l1-1-0.dll 已复制: api-ms-win-core-namedpipe-l1-1-0.dll -> \Dependencies\api-ms-win-core-namedpipe-l1-1-0.dll 已复制: api-ms-win-core-processenvironment-l1-1-0.dll -> \Dependencies\api-ms-win-core-processenvironment-l1-1-0.dll 已复制: api-ms-win-core-processthreads-l1-1-0.dll -> \Dependencies\api-ms-win-core-processthreads-l1-1-0.dll 已复制: api-ms-win-core-processthreads-l1-1-1.dll -> \Dependencies\api-ms-win-core-processthreads-l1-1-1.dll 已复制: api-ms-win-core-profile-l1-1-0.dll -> \Dependencies\api-ms-win-core-profile-l1-1-0.dll 已复制: api-ms-win-core-rtlsupport-l1-1-0.dll -> \Dependencies\api-ms-win-core-rtlsupport-l1-1-0.dll 已复制: api-ms-win-core-string-l1-1-0.dll -> \Dependencies\api-ms-win-core-string-l1-1-0.dll 已复制: api-ms-win-core-synch-l1-1-0.dll -> \Dependencies\api-ms-win-core-synch-l1-1-0.dll 已复制: api-ms-win-core-synch-l1-2-0.dll -> \Dependencies\api-ms-win-core-synch-l1-2-0.dll 已复制: api-ms-win-core-sysinfo-l1-1-0.dll -> \Dependencies\api-ms-win-core-sysinfo-l1-1-0.dll 已复制: api-ms-win-core-timezone-l1-1-0.dll -> \Dependencies\api-ms-win-core-timezone-l1-1-0.dll 已复制: api-ms-win-core-util-l1-1-0.dll -> \Dependencies\api-ms-win-core-util-l1-1-0.dll 已复制: api-ms-win-crt-conio-l1-1-0.dll -> \Dependencies\api-ms-win-crt-conio-l1-1-0.dll 已复制: api-ms-win-crt-convert-l1-1-0.dll -> \Dependencies\api-ms-win-crt-convert-l1-1-0.dll 已复制: api-ms-win-crt-environment-l1-1-0.dll -> \Dependencies\api-ms-win-crt-environment-l1-1-0.dll 已复制: api-ms-win-crt-filesystem-l1-1-0.dll -> \Dependencies\api-ms-win-crt-filesystem-l1-1-0.dll 已复制: api-ms-win-crt-heap-l1-1-0.dll -> \Dependencies\api-ms-win-crt-heap-l1-1-0.dll 已复制: api-ms-win-crt-locale-l1-1-0.dll -> \Dependencies\api-ms-win-crt-locale-l1-1-0.dll 已复制: api-ms-win-crt-math-l1-1-0.dll -> \Dependencies\api-ms-win-crt-math-l1-1-0.dll 已复制: api-ms-win-crt-multibyte-l1-1-0.dll -> \Dependencies\api-ms-win-crt-multibyte-l1-1-0.dll 已复制: api-ms-win-crt-private-l1-1-0.dll -> \Dependencies\api-ms-win-crt-private-l1-1-0.dll 已复制: api-ms-win-crt-process-l1-1-0.dll -> \Dependencies\api-ms-win-crt-process-l1-1-0.dll 已复制: api-ms-win-crt-runtime-l1-1-0.dll -> \Dependencies\api-ms-win-crt-runtime-l1-1-0.dll 已复制: api-ms-win-crt-stdio-l1-1-0.dll -> \Dependencies\api-ms-win-crt-stdio-l1-1-0.dll 已复制: api-ms-win-crt-string-l1-1-0.dll -> \Dependencies\api-ms-win-crt-string-l1-1-0.dll 已复制: api-ms-win-crt-time-l1-1-0.dll -> \Dependencies\api-ms-win-crt-time-l1-1-0.dll 已复制: api-ms-win-crt-utility-l1-1-0.dll -> \Dependencies\api-ms-win-crt-utility-l1-1-0.dll 已复制: ClrPhlib.dll -> \Dependencies\ClrPhlib.dll 已复制: dbghelp.dll -> \Dependencies\dbghelp.dll 已复制: DependenciesLib.dll -> \Dependencies\DependenciesLib.dll 已复制: Dragablz.dll -> \Dependencies\Dragablz.dll 已复制: Mono.Cecil.dll -> \Dependencies\Mono.Cecil.dll 已复制: Mono.Cecil.Mdb.dll -> \Dependencies\Mono.Cecil.Mdb.dll 已复制: Mono.Cecil.Pdb.dll -> \Dependencies\Mono.Cecil.Pdb.dll 已复制: Mono.Cecil.Rocks.dll -> \Dependencies\Mono.Cecil.Rocks.dll 已复制: MSVCP140.dll -> \Dependencies\MSVCP140.dll 已复制: NDesk.Options.dll -> \Dependencies\NDesk.Options.dll 已复制: Newtonsoft.Json.dll -> \Dependencies\Newtonsoft.Json.dll 已复制: ucrtbase.dll -> \Dependencies\ucrtbase.dll 已复制: VCRUNTIME140_1.dll -> \Dependencies\VCRUNTIME140_1.dll 已复制: VCRUNTIME140.dll -> \Dependencies\VCRUNTIME140.dll 已复制: functorch.pyd -> \functorch\functorch.pyd 已复制: c10_cuda.lib -> \lib\c10_cuda.lib 已复制: c10.lib -> \lib\c10.lib 已复制: caffe2_nvrtc.lib -> \lib\caffe2_nvrtc.lib 已复制: Caffe2_perfkernels_avx2.lib -> \lib\Caffe2_perfkernels_avx2.lib 已复制: cpuinfo_internals.lib -> \lib\cpuinfo_internals.lib 已复制: cpuinfo.lib -> \lib\cpuinfo.lib 已复制: fmt.lib -> \lib\fmt.lib 已复制: functorch.lib -> \lib\functorch.lib 已复制: kineto.lib -> \lib\kineto.lib 已复制: libittnotify.lib -> \lib\libittnotify.lib 已复制: libjitprofiling.lib -> \lib\libjitprofiling.lib 已复制: libprotobuf-lite.lib -> \lib\libprotobuf-lite.lib 已复制: libprotobuf.lib -> \lib\libprotobuf.lib 已复制: libprotoc.lib -> \lib\libprotoc.lib 已复制: microkernels-all.lib -> \lib\microkernels-all.lib 已复制: microkernels-prod.lib -> \lib\microkernels-prod.lib 已复制: mimalloc.lib -> \lib\mimalloc.lib 已复制: nnapi_backend.lib -> \lib\nnapi_backend.lib 已复制: onnx_proto.lib -> \lib\onnx_proto.lib 已复制: onnx.lib -> \lib\onnx.lib 已复制: pthreadpool.lib -> \lib\pthreadpool.lib 已复制: shm.lib -> \lib\shm.lib 已复制: torch_cpu.lib -> \lib\torch_cpu.lib 已复制: torch_cuda.lib -> \lib\torch_cuda.lib 已复制: torch_python.lib -> \lib\torch_python.lib 已复制: torch.lib -> \lib\torch.lib 已复制: XNNPACK.lib -> \lib\XNNPACK.lib 已复制: _C.cp310-win_amd64.pyd -> \lib.win-amd64-cpython-310\functorch\_C.cp310-win_amd64.pyd 已复制: sleef.lib -> \sleef\lib\sleef.lib (.venv) PS E:\PyTorch_Build\pytorch\build> # 验证核心文件是否已复制 (.venv) PS E:\PyTorch_Build\pytorch\build> $criticalFiles = @( >> "_C.pyd", >> "_C_flatbuffer.pyd", >> "torch_python.dll", >> "c10.dll", >> "c10_cuda.dll" >> ) (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> foreach ($file in $criticalFiles) { >> $filePath = Join-Path $torchDir $file >> if (Test-Path $filePath) { >> Write-Host "[成功] 核心文件存在: $file" >> } else { >> Write-Host "[错误] 核心文件缺失: $file" >> >> # 尝试搜索整个系统 >> $found = Get-ChildItem -Path $buildRoot -Recurse -Filter $file -ErrorAction SilentlyContinue >> if ($found) { >> Write-Host " 文件位于: $($found.FullName)" >> Copy-Item -Path $found.FullName -Destination $torchDir -Force >> Write-Host " 已手动复制文件" >> } else { >> Write-Host " 系统中未找到此文件" >> } >> } >> } [错误] 核心文件缺失: _C.pyd 系统中未找到此文件 [错误] 核心文件缺失: _C_flatbuffer.pyd 系统中未找到此文件 [错误] 核心文件缺失: torch_python.dll 文件位于: E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\torch\bin\torch_python.dll E:\PyTorch_Build\pytorch\build\bin\torch_python.dll 已手动复制文件 [错误] 核心文件缺失: c10.dll 文件位于: E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\torch\bin\c10.dll E:\PyTorch_Build\pytorch\build\bin\c10.dll 已手动复制文件 [错误] 核心文件缺失: c10_cuda.dll 文件位于: E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages\torch\bin\c10_cuda.dll E:\PyTorch_Build\pytorch\build\bin\c10_cuda.dll 已手动复制文件 (.venv) PS E:\PyTorch_Build\pytorch\build> # 保存为 check_build.ps1 (.venv) PS E:\PyTorch_Build\pytorch\build> $buildRoot = "E:\PyTorch_Build\pytorch\build" (.venv) PS E:\PyTorch_Build\pytorch\build> $requiredFiles = @( >> "lib.win-amd64-cpython-310\torch\_C.pyd", >> "lib.win-amd64-cpython-310\torch\lib\c10.dll", >> "lib.win-amd64-cpython-310\torch\lib\torch_cuda.dll" >> ) (.venv) PS E:\PyTorch_Build\pytorch\build> (.venv) PS E:\PyTorch_Build\pytorch\build> Write-Host "=== 构建完整性检查 ===" === 构建完整性检查 === (.venv) PS E:\PyTorch_Build\pytorch\build> foreach ($file in $requiredFiles) { >> $fullPath = Join-Path $buildRoot $file >> $exists = Test-Path $fullPath >> >> if ($exists) { >> Write-Host "[✓] $file" >> } else { >> Write-Host "[✗] $file" >> Write-Host " 文件缺失,请检查构建日志" >> >> # 检查构建日志 >> $logPath = Join-Path $buildRoot "build.log" >> if (Test-Path $logPath) { >> $errorLines = Get-Content $logPath | Select-String "error" -CaseSensitive >> if ($errorLines) { >> Write-Host " 构建日志中的错误:" >> $errorLines | Select-Object -First 5 | ForEach-Object { Write-Host " $_" } >> } >> } >> } >> } [✗] lib.win-amd64-cpython-310\torch\_C.pyd 文件缺失,请检查构建日志 [✗] lib.win-amd64-cpython-310\torch\lib\c10.dll 文件缺失,请检查构建日志 [✗] lib.win-amd64-cpython-310\torch\lib\torch_cuda.dll 文件缺失,请检查构建日志 (.venv) PS E:\PyTorch_Build\pytorch\build> # 查看构建日志 (.venv) PS E:\PyTorch_Build\pytorch\build> Get-Content "E:\PyTorch_Build\pytorch\build\build.log" -Tail 50 Get-Content: Cannot find path 'E:\PyTorch_Build\pytorch\build\build.log' because it does not exist. (.venv) PS E:\PyTorch_Build\pytorch\build> # 检查CUDA编译器 (.venv) PS E:\PyTorch_Build\pytorch\build> nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2025 NVIDIA Corporation Built on Wed_Jul_16_20:06:48_Pacific_Daylight_Time_2025 Cuda compilation tools, release 13.0, V13.0.48 Build cuda_13.0.r13.0/compiler.36260728_0 (.venv) PS E:\PyTorch_Build\pytorch\build> # 检查CUDA示例编译 (.venv) PS E:\PyTorch_Build\pytorch\build> cd "E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\extras\demo_suite" Set-Location: Cannot find path 'E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\extras\demo_suite' because it does not exist. (.venv) PS E:\PyTorch_Build\pytorch\build> .\bandwidthTest.exe .\bandwidthTest.exe: The term '.\bandwidthTest.exe' is not recognized as a name of a cmdlet, function, script file, or executable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again. (.venv) PS E:\PyTorch_Build\pytorch\build> # 检查Python开发头文件 (.venv) PS E:\PyTorch_Build\pytorch\build> python -c "import sysconfig; print(sysconfig.get_config_var('INCLUDEPY'))" E:\Python310\Include (.venv) PS E:\PyTorch_Build\pytorch\build> # 检查MSVC编译器 (.venv) PS E:\PyTorch_Build\pytorch\build> cl.exe /? cl.exe: The term 'cl.exe' is not recognized as a name of a cmdlet, function, script file, or executable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again. (.venv) PS E:\PyTorch_Build\pytorch\build> # 运行所有检查 (.venv) PS E:\PyTorch_Build\pytorch\build> .\check_build.ps1 .\check_build.ps1: The term '.\check_build.ps1' is not recognized as a name of a cmdlet, function, script file, or executable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again. (.venv) PS E:\PyTorch_Build\pytorch\build> python torch_diagnose.py ================================================== PyTorch 环境诊断工具 ================================================== [系统信息] 操作系统: Windows-10-10.0.26100-SP0 Python本: 3.10.10 (tags/v3.10.10:aad5f6a, Feb 7 2023, 17:20:36) [MSC v.1929 64 bit (AMD64)] Python路径: E:\PyTorch_Build\pytorch\build\.venv\Scripts\python.exe 工作目录: E:\PyTorch_Build\pytorch\build [环境变量] PATH: E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\PyTorch_Build\pytorch\build\lib;E:\OpenBLAS\bin;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\PyTorch_Build\pytorch\build\.venv\Scripts;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;C:\Program Files\PowerShell\7;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin\x64;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin\x64;;E:\PyTorch_Build\pytorch\build\lib.win-amd64-cpython-310\torch\lib;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin;C:\Users\Administrator\AppData\Local\Microsoft\dotnet;C:\Users\Administrator\AppData\Local\Microsoft\dotnet;C:\Users\Administrator\AppData\Local\Microsoft\dotnet\;C:\Program Files\dotnet;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0\;C:\WINDOWS\System32\OpenSSH\;E:\Python310;C:\Program Files\dotnet\;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\Users\Administrator\.dotnet\tools;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\Users\Administrator\.dotnet\tools;E:\Python310\Scripts;E:\Python310\Scripts;C:\Program Files\PowerShell\7\;E:\Program Files\Microsoft VS Code\bin;E:\Program Files\Git\cmd;C:\Program Files\NVIDIA Corporation\Nsight Compute 2025.3.0\;E:\Program Files\CMake\bin;C:\Program Files\Microsoft SQL Server\150\Tools\Binn\;C:\Program Files\Microsoft SQL Server\Client SDK\ODBC\170\Tools\Binn\;C:\Program Files (x86)\Incredibuild;E:\PyTorch_Build\pytorch\build\lib.win-amd64-cpython-310\torch\lib;E:\Program Files (x86)\Windows Kits\10\Windows Performance Toolkit\;C:\ProgramData\chocolatey\bin;E:\Program Files\Rust\.cargo\bin;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\Users\Administrator\.dotnet\tools PYTHONPATH: E:\PyTorch_Build\pytorch Torch目录: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch [核心文件检查] [✗] 文件存在: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch\_C.pyd [✗] 文件存在: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch\_C_flatbuffer.pyd [✓] 文件存在: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch\torch_python.dll [✓] 文件存在: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch\c10.dll [✓] 文件存在: E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages\torch\c10_cuda.dll [DLL依赖检查] [✗] DLL加载失败: _C.pyd - Could not find module '_C.pyd' (or one of its dependencies). Try using the full path with constructor syntax. [✗] DLL加载失败: _C_flatbuffer.pyd - Could not find module '_C_flatbuffer.pyd' (or one of its dependencies). Try using the full path with constructor syntax. [✗] DLL加载失败: torch_python.dll - Could not find module 'torch_python.dll' (or one of its dependencies). Try using the full path with constructor syntax. [✗] DLL加载失败: c10.dll - Could not find module 'c10.dll' (or one of its dependencies). Try using the full path with constructor syntax. [✗] DLL加载失败: c10_cuda.dll - Could not find module 'c10_cuda.dll' (or one of its dependencies). Try using the full path with constructor syntax. [模块导入测试] [✗] 模块导入失败: torch - DLL load failed while importing _C: 找不到指定的模块。 [错误详情] Traceback (most recent call last): File "E:\PyTorch_Build\pytorch\build\torch_diagnose.py", line 83, in main import torch File "E:\PyTorch_Build\pytorch\torch\__init__.py", line 423, in <module> from torch._C import * # noqa: F403 ImportError: DLL load failed while importing _C: 找不到指定的模块。 [调试信息] Python库路径: - E:\Python310\lib - E:\PyTorch_Build\pytorch\build\.venv\lib\site-packages PATH中的关键目录: - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\PyTorch_Build\pytorch\build\.venv\Lib\site-packages - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\PyTorch_Build\pytorch\build\lib - E:\OpenBLAS\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\PyTorch_Build\pytorch\build\.venv\Scripts - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin\x64 - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin\x64 - E:\PyTorch_Build\pytorch\build\lib.win-amd64-cpython-310\torch\lib - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.0\bin - E:\Program Files\Microsoft VS Code\bin - E:\Program Files\CMake\bin - C:\Program Files\Microsoft SQL Server\150\Tools\Binn\ - C:\Program Files\Microsoft SQL Server\Client SDK\ODBC\170\Tools\Binn\ - E:\PyTorch_Build\pytorch\build\lib.win-amd64-cpython-310\torch\lib - C:\ProgramData\chocolatey\bin - E:\Program Files\Rust\.cargo\bin (.venv) PS E:\PyTorch_Build\pytorch\build>
09-02
在使用python3 setup.py install --user时出现-- Could NOT find CUDA (missing: CUDA_INCLUDE_DIRS CUDA_CUDART_LIBRARY) (found version "12.6") CMake Warning at /home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake/Caffe2/public/cuda.cmake:31 (message): Caffe2: CUDA cannot be found. Depending on whether you are building Caffe2 or a Caffe2 dependent library, the next warning / error will give you more info. Call Stack (most recent call first): /home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake/Caffe2/Caffe2Config.cmake:86 (include) /home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:68 (find_package) cmake/TorchAudioHelper.cmake:1 (find_package) CMakeLists.txt:91 (include) CMake Error at /home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake/Caffe2/Caffe2Config.cmake:90 (message): Your installed Caffe2 version uses CUDA but I cannot find the CUDA libraries. Please set the proper CUDA prefixes and / or install CUDA. Call Stack (most recent call first): /home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:68 (find_package) cmake/TorchAudioHelper.cmake:1 (find_package) CMakeLists.txt:91 (include) -- Configuring incomplete, errors occurred! You have changed variables that require your cache to be deleted. Configure will be re-run and you may have to reset some variables. The following variables have changed: CMAKE_CUDA_COMPILER= :/usr/local/cuda-12.6/bin/nvcc -- Generating done (0.0s) CMake Generate step failed. Build files cannot be regenerated correctly. Traceback (most recent call last): File "/home/nvidia/audio/setup.py", line 144, in <module> _main() File "/home/nvidia/audio/setup.py", line 99, in _main setup( File "/usr/lib/python3/dist-packages/setuptools/__init__.py", line 153, in setup return distutils.core.setup(**attrs) File "/usr/lib/python3.10/distutils/core.py", line 148, in setup dist.run_commands() File "/usr/lib/python3.10/distutils/dist.py", line 966, in run_commands self.run_command(cmd) File "/usr/lib/python3.10/distutils/dist.py", line 985, in run_command cmd_obj.run() File "/usr/lib/python3/dist-packages/setuptools/command/install.py", line 74, in run self.do_egg_install() File "/usr/lib/python3/dist-packages/setuptools/command/install.py", line 116, in do_egg_install self.run_command('bdist_egg') File "/usr/lib/python3.10/distutils/cmd.py", line 313, in run_command self.distribution.run_command(command) File "/usr/lib/python3.10/distutils/dist.py", line 985, in run_command cmd_obj.run() File "/usr/lib/python3/dist-packages/setuptools/command/bdist_egg.py", line 164, in run cmd = self.call_command('install_lib', warn_dir=0) File "/usr/lib/python3/dist-packages/setuptools/command/bdist_egg.py", line 150, in call_command self.run_command(cmdname) File "/usr/lib/python3.10/distutils/cmd.py", line 313, in run_command self.distribution.run_command(command) File "/usr/lib/python3.10/distutils/dist.py", line 985, in run_command cmd_obj.run() File "/usr/lib/python3/dist-packages/setuptools/command/install_lib.py", line 23, in run self.build() File "/usr/lib/python3.10/distutils/command/install_lib.py", line 109, in build self.run_command('build_ext') File "/usr/lib/python3.10/distutils/cmd.py", line 313, in run_command self.distribution.run_command(command) File "/usr/lib/python3.10/distutils/dist.py", line 985, in run_command cmd_obj.run() File "/home/nvidia/audio/tools/setup_helpers/extension.py", line 70, in run super().run() File "/usr/lib/python3/dist-packages/setuptools/command/build_ext.py", line 79, in run _build_ext.run(self) File "/usr/lib/python3.10/distutils/command/build_ext.py", line 340, in run self.build_extensions() File "/usr/lib/python3.10/distutils/command/build_ext.py", line 449, in build_extensions self._build_extensions_serial() File "/usr/lib/python3.10/distutils/command/build_ext.py", line 474, in _build_extensions_serial self.build_extension(ext) File "/home/nvidia/audio/tools/setup_helpers/extension.py", line 155, in build_extension subprocess.check_call(["cmake", str(_ROOT_DIR)] + cmake_args, cwd=self.build_temp) File "/usr/lib/python3.10/subprocess.py", line 369, in check_call raise CalledProcessError(retcode, cmd) subprocess.CalledProcessError: Command '['cmake', '/home/nvidia/audio', '-DCMAKE_BUILD_TYPE=Release', '-DCMAKE_PREFIX_PATH=/home/nvidia/.local/lib/python3.10/site-packages/torch/share/cmake', '-DCMAKE_INSTALL_PREFIX=/home/nvidia/audio/build/lib.linux-aarch64-3.10/', '-DCMAKE_VERBOSE_MAKEFILE=ON', '-DPython_INCLUDE_DIR=/usr/include/python3.10', '-DBUILD_CPP_TEST=OFF', '-DBUILD_RIR:BOOL=ON', '-DBUILD_RNNT:BOOL=ON', '-DBUILD_ALIGN:BOOL=ON', '-DBUILD_CUDA_CTC_DECODER:BOOL=ON', '-DBUILD_TORCHAUDIO_PYTHON_EXTENSION:BOOL=ON', '-DBUILD_TORIO_PYTHON_EXTENSION:BOOL=ON', '-DUSE_ROCM:BOOL=OFF', '-DUSE_CUDA:BOOL=ON', '-DUSE_OPENMP:BOOL=ON', "-DCMAKE_CUDA_COMPILER=':/usr/local/cuda-12.6/bin/nvcc'", "-DCUDA_TOOLKIT_ROOT_DIR=':/usr/local/cuda-12.6'", '-GNinja']' returned non-zero exit status 1.
最新发布
09-24
评论 4
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值