From cb01cb39af1528ea6386e3b1ec79138189657834 Mon Sep 17 00:00:00 2001 From: megemini Date: Wed, 19 Nov 2025 13:01:29 +0800 Subject: [PATCH 1/2] =?UTF-8?q?[CppExtension]=20=E6=9B=B4=E6=96=B0?= =?UTF-8?q?=E8=87=AA=E5=AE=9A=E4=B9=89=E7=AE=97=E5=AD=90=E6=96=87=E6=A1=A3?= =?UTF-8?q?=EF=BC=8C=E5=AE=8C=E5=96=84=E5=AE=89=E8=A3=85=E8=B7=AF=E5=BE=84?= =?UTF-8?q?=E5=92=8C=E6=97=A5=E5=BF=97=E4=BF=A1=E6=81=AF?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/guides/custom_op/new_cpp_op_cn.md | 130 +++++++++++++++---------- 1 file changed, 79 insertions(+), 51 deletions(-) diff --git a/docs/guides/custom_op/new_cpp_op_cn.md b/docs/guides/custom_op/new_cpp_op_cn.md index 586e308e377..1ff16a813c0 100644 --- a/docs/guides/custom_op/new_cpp_op_cn.md +++ b/docs/guides/custom_op/new_cpp_op_cn.md @@ -1423,7 +1423,7 @@ PD_BUILD_GRAD_OP(custom_add) ### 使用 `setuptools` 编译 -该方式是对 `python` 内建库中的 `setuptools.setup` 接口的进一步封装，能够自动地生成 Python API 并以 Module 的形式安装到 `site-packages` 目录。编译完成后，支持通过 `import` 语句导入使用。 +该方式是对 `python` 内建库中的 `setuptools.setup` 接口的进一步封装，能够自动地生成 Python API 并以 Module 的形式安装到 `site-packages` 或者 `dist-packages` 目录。编译完成后，支持通过 `import` 语句导入使用。 您需要编写 `setup.py` 文件， 配置自定义算子的编译规则。 @@ -1459,68 +1459,83 @@ setup( 执行 `python setup_cpu.py install` 或者 `python setup_cuda.py install` 即可一键完成自定义算子的编译和安装。 -以 `python setup_cuda.py install` 为例，执行日志如下： +以 `python setup_cpu.py install` 为例，执行日志如下 (省略日志时间，以及 setuptools 版本相关警告等信息)： ``` -running install -running bdist_egg -running egg_info -writing custom_setup_ops.egg-info/PKG-INFO -writing dependency_links to custom_setup_ops.egg-info/dependency_links.txt -writing top-level names to custom_setup_ops.egg-info/top_level.txt -reading manifest file 'custom_setup_ops.egg-info/SOURCES.txt' -writing manifest file 'custom_setup_ops.egg-info/SOURCES.txt' -installing library code to build/custom_setup_ops/bdist.linux-x86_64/egg -running install_lib -running build_ext -/usr/local/lib/python3.7/site-packages/paddle/fluid/layers/utils.py:77: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working - return (isinstance(seq, collections.Sequence) and +[ INFO] dist.py:1018 - running install +[ INFO] dist.py:1018 - running build +[ INFO] dist.py:1018 - running build_ext Compiling user custom op, it will cost a few seconds..... -creating build/custom_setup_ops/bdist.linux-x86_64/egg -copying build/custom_setup_ops/lib.linux-x86_64-3.7/version.txt -> build/custom_setup_ops/bdist.linux-x86_64/egg -copying build/custom_setup_ops/lib.linux-x86_64-3.7/relu_cpu.o -> build/custom_setup_ops/bdist.linux-x86_64/egg -copying build/custom_setup_ops/lib.linux-x86_64-3.7/relu_cuda.o -> build/custom_setup_ops/bdist.linux-x86_64/egg -copying build/custom_setup_ops/lib.linux-x86_64-3.7/relu_cuda.cu.o -> build/custom_setup_ops/bdist.linux-x86_64/egg -copying build/custom_setup_ops/lib.linux-x86_64-3.7/custom_setup_ops.so -> build/custom_setup_ops/bdist.linux-x86_64/egg -creating stub loader for custom_setup_ops.so -byte-compiling build/custom_setup_ops/bdist.linux-x86_64/egg/custom_setup_ops.py to custom_setup_ops.cpython-37.pyc -creating build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -copying custom_setup_ops.egg-info/PKG-INFO -> build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -copying custom_setup_ops.egg-info/SOURCES.txt -> build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -copying custom_setup_ops.egg-info/dependency_links.txt -> build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -copying custom_setup_ops.egg-info/not-zip-safe -> build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -copying custom_setup_ops.egg-info/top_level.txt -> build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO -writing build/custom_setup_ops/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt -creating 'dist/custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg' and adding 'build/custom_setup_ops/bdist.linux-x86_64/egg' to it -removing 'build/custom_setup_ops/bdist.linux-x86_64/egg' (and everything under it) -Processing custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg -creating /usr/local/lib/python3.7/site-packages/custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg -Extracting custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg to /usr/local/lib/python3.7/site-packages -Adding custom-setup-ops 0.0.0 to easy-install.pth file - -Installed /usr/local/lib/python3.7/site-packages/custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg -Processing dependencies for custom-setup-ops==0.0.0 -Finished processing dependencies for custom-setup-ops==0.0.0 +[ INFO] build_ext.py:538 - building 'custom_setup_ops' extension +[ INFO] dir_util.py:58 - creating /paddle/Paddle/build/tmp_setuptools/build/custom_setup_ops/lib.linux-x86_64-cpython-39/build/custom_setup_ops/temp.linux-x86_64-cpython-39 +[ INFO] spawn.py:77 - x86_64-linux-gnu-g++ -pthread -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -fPIC -I/usr/local/lib/python3.9/dist-packages/paddle/include -I/usr/local/lib/python3.9/dist-packages/paddle/include/third_party -I/usr/local/lib/python3.9/dist-packages/paddle/include/paddle/phi/api/include/compat -I/usr/local/lib/python3.9/dist-packages/paddle/include/paddle/phi/api/include/compat/torch/csrc/api/include -I/usr/include/python3.9 -I/usr/include/python3.9 -c /paddle/Paddle/build/tmp_setuptools/relu_cpu.cc -o /paddle/Paddle/build/tmp_setuptools/build/custom_setup_ops/lib.linux-x86_64-cpython-39/build/custom_setup_ops/temp.linux-x86_64-cpython-39/relu_cpu.o -w -DPADDLE_WITH_CUSTOM_KERNEL -DPADDLE_EXTENSION_NAME=custom_setup_ops -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++17 +/paddle/Paddle/build/tmp_setuptools/build/custom_setup_ops/lib.linux-x86_64-cpython-39/build/custom_setup_ops/temp.linux-x86_64-cpython-39/relu_cpu.o is compiled +[ INFO] spawn.py:77 - x86_64-linux-gnu-g++ -pthread -Wno-unused-result -Wsign-compare -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -g -fwrapv -O2 -shared -Wl,-O1 -Wl,-Bsymbolic-functions /paddle/Paddle/build/tmp_setuptools/build/custom_setup_ops/lib.linux-x86_64-cpython-39/build/custom_setup_ops/temp.linux-x86_64-cpython-39/relu_cpu.o -L/usr/local/lib/python3.9/dist-packages/paddle/libs -L/usr/local/lib/python3.9/dist-packages/paddle/base -L/usr/lib/x86_64-linux-gnu -Wl,--enable-new-dtags,-rpath,/usr/local/lib/python3.9/dist-packages/paddle/libs -Wl,--enable-new-dtags,-rpath,/usr/local/lib/python3.9/dist-packages/paddle/base -o build/custom_setup_ops/lib.linux-x86_64-cpython-39/custom_setup_ops.so -l:libpaddle.so +Received len(custom_op) = 1, using custom operator +Removed: build/custom_setup_ops/lib.linux-x86_64-cpython-39/build/custom_setup_ops/temp.linux-x86_64-cpython-39/relu_cpu.o +[ INFO] dist.py:1018 - running install_lib +[ INFO] file_util.py:130 - copying build/custom_setup_ops/lib.linux-x86_64-cpython-39/version.txt -> /usr/local/lib/python3.9/dist-packages +[ INFO] file_util.py:130 - copying build/custom_setup_ops/lib.linux-x86_64-cpython-39/custom_setup_ops.py -> /usr/local/lib/python3.9/dist-packages +[ INFO] file_util.py:130 - copying build/custom_setup_ops/lib.linux-x86_64-cpython-39/custom_setup_ops.so -> /usr/local/lib/python3.9/dist-packages +[ INFO] util.py:485 - byte-compiling /usr/local/lib/python3.9/dist-packages/custom_setup_ops.py to custom_setup_ops.cpython-39.pyc +[ INFO] dist.py:1018 - running install_egg_info +[ INFO] dist.py:1018 - running egg_info +[ INFO] dir_util.py:58 - creating custom_setup_ops.egg-info +[ INFO] egg_info.py:651 - writing custom_setup_ops.egg-info/PKG-INFO +[ INFO] egg_info.py:279 - writing dependency_links to custom_setup_ops.egg-info/dependency_links.txt +[ INFO] egg_info.py:279 - writing top-level names to custom_setup_ops.egg-info/top_level.txt +[ INFO] util.py:332 - writing manifest file 'custom_setup_ops.egg-info/SOURCES.txt' +[ INFO] sdist.py:203 - reading manifest file 'custom_setup_ops.egg-info/SOURCES.txt' +[ INFO] util.py:332 - writing manifest file 'custom_setup_ops.egg-info/SOURCES.txt' +[ INFO] util.py:332 - Copying custom_setup_ops.egg-info to /usr/local/lib/python3.9/dist-packages/custom_setup_ops-0.0.0-py3.9.egg-info +[ INFO] dist.py:1018 - running install_scripts ``` -执行成功后，如日志所示，自定义算子模块 `custom_setup_ops` 被安装至如下目录： +执行成功后，如日志所示，自定义算子模块 `custom_setup_ops` 的安装信息如下： -`/usr/local/lib/python3.7/site-packages/custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg` +``` shell -`custom_setup_ops-0.0.0-py3.7-linux-x86_64.egg` 目录中内容如下： +> pip show custom_setup_ops +Name: custom_setup_ops +Version: 0.0.0 +Summary: +Home-page: +Author: +Author-email: +License: +Location: /usr/local/lib/python3.9/dist-packages +Requires: +Required-by: ``` -custom_setup_ops_pd_.so EGG-INFO/ relu_cpu.o relu_cuda.o -custom_setup_ops.py __pycache__/ relu_cuda.cu.o version.txt + +其目录结构如下： + +``` shell + +> tree /usr/local/lib/python3.9/dist-packages/custom_setup_ops* +/usr/local/lib/python3.9/dist-packages/custom_setup_ops +|-- __init__.py +`-- custom_setup_ops_pd_.so +/usr/local/lib/python3.9/dist-packages/custom_setup_ops-0.0.0-py3.9.egg-info +|-- PKG-INFO +|-- SOURCES.txt +|-- dependency_links.txt +|-- not-zip-safe +`-- top_level.txt + ``` -其中 `custom_setup_ops_pd_.so` 为自定义算子编译生成的动态库，`custom_setup_ops.py` 为根据 `PaddlePaddle` 接口的定义规则，自动生成的自定义算子 python 模块源码，其示例内容为（自动生成的代码后续可能会更新，生成结果可能与示例代码不一致）： +其中 `custom_setup_ops_pd_.so` 为自定义算子编译生成的动态库，`custom_setup_ops/__init__.py` 为根据 `PaddlePaddle` 接口的定义规则，自动生成的自定义算子 python 模块源码，其示例内容为（自动生成的代码后续可能会更新，生成结果可能与示例代码不一致）： ```python + from paddle import _C_ops from paddle.framework import in_dynamic_or_pir_mode from paddle.base.layer_helper import LayerHelper +from paddle.jit.marker import unified +@unified def custom_relu(x): # The output variable's dtype use default value 'float32', # and the actual dtype of output variable will be inferred in runtime. @@ -1586,11 +1601,24 @@ __bootstrap__() 随后，可以直接在构建模型过程中导入使用，简单示例如下： ```python -import paddle -from custom_setup_ops import custom_relu - -x = paddle.randn([4, 10], dtype='float32') -relu_out = custom_relu(x) +In [1]: import paddle + ...: from custom_setup_ops import custom_relu + ...: paddle.set_device('cpu') + ...: + ...: x = paddle.randn([4, 10], dtype='float32') + ...: relu_out = custom_relu(x) + +In [2]: relu_out +Out[2]: +Tensor(shape=[4, 10], dtype=float32, place=Place(cpu), stop_gradient=True, + [[0.02314972, 0.65578228, 0. , 0. , 0.18305063, 0. , + 0.67343038, 0. , 1.16782570, 1.71236455], + [0. , 0.60349381, 0. , 0. , 0. , 0. , + 0.14162211, 0. , 0. , 0.33964530], + [0. , 0. , 0.12062856, 0.18853758, 0.54154527, 0.73217475, + 0. , 0. , 0. , 0. ], + [0.04105225, 0. , 0.67857188, 0.95838499, 1.08346415, 2.47209001, + 0. , 0. , 0.22969440, 1.08237624]]) ``` > 注：`setuptools` 的封装是为了简化自定义算子编译和使用流程，即使不依赖于 `setuptools`，也可以自行编译生成动态库，并封装相应的 python API，然后在基于 `PaddlePaddle` 实现的模型中使用 From fafceedf1e71bbd38fdcbda715c6147821b87b3d Mon Sep 17 00:00:00 2001 From: megemini Date: Wed, 19 Nov 2025 21:10:57 +0800 Subject: [PATCH 2/2] =?UTF-8?q?[CppExtension]=20=E6=9B=B4=E6=96=B0?= =?UTF-8?q?=E8=87=AA=E5=AE=9A=E4=B9=89=E7=AE=97=E5=AD=90=E6=96=87=E6=A1=A3?= =?UTF-8?q?=EF=BC=8C=E4=BC=98=E5=8C=96=E4=BB=A3=E7=A0=81=E7=A4=BA=E4=BE=8B?= =?UTF-8?q?=E6=A0=BC=E5=BC=8F=E5=B9=B6=E6=B7=BB=E5=8A=A0=E7=8E=B0=E4=BB=A3?= =?UTF-8?q?=E6=9E=84=E5=BB=BA=E6=96=B9=E5=BC=8F=E8=AF=B4=E6=98=8E?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/guides/custom_op/new_cpp_op_cn.md | 37 +++++++++++++------------- 1 file changed, 19 insertions(+), 18 deletions(-) diff --git a/docs/guides/custom_op/new_cpp_op_cn.md b/docs/guides/custom_op/new_cpp_op_cn.md index 1ff16a813c0..35866f16324 100644 --- a/docs/guides/custom_op/new_cpp_op_cn.md +++ b/docs/guides/custom_op/new_cpp_op_cn.md @@ -1459,6 +1459,8 @@ setup( 执行 `python setup_cpu.py install` 或者 `python setup_cuda.py install` 即可一键完成自定义算子的编译和安装。 +> 注：在当前目录只有一个 `setup.py` 文件时，可以使用 `pip install . --no-build-isolation` 这种现代化的方式进行构建。 + 以 `python setup_cpu.py install` 为例，执行日志如下 (省略日志时间，以及 setuptools 版本相关警告等信息)： ``` @@ -1601,24 +1603,23 @@ __bootstrap__() 随后，可以直接在构建模型过程中导入使用，简单示例如下： ```python -In [1]: import paddle - ...: from custom_setup_ops import custom_relu - ...: paddle.set_device('cpu') - ...: - ...: x = paddle.randn([4, 10], dtype='float32') - ...: relu_out = custom_relu(x) - -In [2]: relu_out -Out[2]: -Tensor(shape=[4, 10], dtype=float32, place=Place(cpu), stop_gradient=True, - [[0.02314972, 0.65578228, 0. , 0. , 0.18305063, 0. , - 0.67343038, 0. , 1.16782570, 1.71236455], - [0. , 0.60349381, 0. , 0. , 0. , 0. , - 0.14162211, 0. , 0. , 0.33964530], - [0. , 0. , 0.12062856, 0.18853758, 0.54154527, 0.73217475, - 0. , 0. , 0. , 0. ], - [0.04105225, 0. , 0.67857188, 0.95838499, 1.08346415, 2.47209001, - 0. , 0. , 0.22969440, 1.08237624]]) +import paddle +from custom_setup_ops import custom_relu +paddle.set_device('cpu') + +x = paddle.randn([4, 10], dtype='float32') +relu_out = custom_relu(x) + +# `relu_out` should be like: +# Tensor(shape=[4, 10], dtype=float32, place=Place(cpu), stop_gradient=True, +# [[0.02314972, 0.65578228, 0. , 0. , 0.18305063, 0. , +# 0.67343038, 0. , 1.16782570, 1.71236455], +# [0. , 0.60349381, 0. , 0. , 0. , 0. , +# 0.14162211, 0. , 0. , 0.33964530], +# [0. , 0. , 0.12062856, 0.18853758, 0.54154527, 0.73217475, +# 0. , 0. , 0. , 0. ], +# [0.04105225, 0. , 0.67857188, 0.95838499, 1.08346415, 2.47209001, +# 0. , 0. , 0.22969440, 1.08237624]]) ``` > 注：`setuptools` 的封装是为了简化自定义算子编译和使用流程，即使不依赖于 `setuptools`，也可以自行编译生成动态库，并封装相应的 python API，然后在基于 `PaddlePaddle` 实现的模型中使用