We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
(此 ISSUE 为 PaddlePaddle Hackathon 第三期活动的任务 ISSUE,更多详见 【PaddlePaddle Hackathon 第三期】任务总览)
每个框架都有自己的模型和算子表达。OpenVINO 对 PaddlePaddle 的支持需要从 Paddle 的算子映射转换到 OpenVINO 的算子。在这个过程中,我们将熟悉深度学习神经网络的算子表达和计算。开发请参考 贡献指南,任务列表如下:
提交地址:https://github.com/openvinotoolkit/openvino
OpenVINO算子库文档: https://github.com/openvinotoolkit/openvino/blob/master/docs/ops/opset9.md
OpenVINO算子参考实现: https://github.com/openvinotoolkit/openvino/tree/master/src/core/reference/include/ngraph/runtime/reference
PaddlePaddle算子库文档: https://www.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html
PaddlePaddle算子参考实现: https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/fluid/tests/unittests
Paddle2ONNX算子映射参考代码 https://github.com/PaddlePaddle/Paddle2ONNX/tree/develop/paddle2onnx/legacy/op_mapper https://github.com/PaddlePaddle/Paddle2ONNX/tree/develop/paddle2onnx/mapper
可以先生成测试模型用Paddle VisualDL查看paddle算子的输入输出以及属性: https://www.paddlepaddle.org.cn/paddle/visualdl/demo/graph
OpenVINO源码编译方法: 参考:https://github.com/openvinotoolkit/openvino/wiki
$ git clone https://github.com/openvinotoolkit/openvino.git $ cd openvino $ git submodule update --init --recursive $ chmod +x install_build_dependencies.sh $./install_build_dependencies.sh $ export OPENVINO_BASEDIR=`pwd` $ mkdir build $ cd build $ cmake \ -DCMAKE_BUILD_TYPE= Release -DCMAKE_INSTALL_PREFIX="${OPENVINO_BASEDIR}/openvino_dist" \ -DPYTHON_EXECUTABLE=$(which python3) \ -DENABLE_MYRIAD=OFF \ -DENABLE_VPU=OFF \ -DENABLE_PYTHON=ON \ -DNGRAPH_PYTHON_BUILD_ENABLE=ON \ -DENABLE_DEBUG_CAPS=ON \ -DENABLE_CPU_DEBUG_CAPS=ON \ -DENABLE_TESTS=ON \ .. $ make -j$(nproc); make install
单测测试方法:
$ cd bin/intel64/Release $ ./paddle_tests --gtest_filter=PaddleFuzzyOpTest/FrontEndFuzzyOpTest.testOpFuzzy/*
The text was updated successfully, but these errors were encountered:
No branches or pull requests
(此 ISSUE 为 PaddlePaddle Hackathon 第三期活动的任务 ISSUE,更多详见 【PaddlePaddle Hackathon 第三期】任务总览)
每个框架都有自己的模型和算子表达。OpenVINO 对 PaddlePaddle 的支持需要从 Paddle 的算子映射转换到 OpenVINO 的算子。在这个过程中,我们将熟悉深度学习神经网络的算子表达和计算。开发请参考 贡献指南,任务列表如下:
No.95:为 OpenVINO 实现 Paddle 算子 box_coder 转换
No.96:为 OpenVINO 实现 Paddle 算子 group_norm 转换
No.97:为 OpenVINO 实现 Paddle 算子 p_norm 转换
No.98:为 OpenVINO 实现 Paddle 算子 tile 转换
No.99:为 OpenVINO 实现 Paddle 算子 ceil 转换
No.100:为 OpenVINO 实现 Paddle 算子 gather_nd 转换
No.101:为 OpenVINO 实现 Paddle 算子 sum 转换
No.102:为 OpenVINO 实现 Paddle 算子 elementwise_floordiv 转换
No.103:为 OpenVINO 实现 Paddle 算子 elementwise_mod 转换
No.104:为 OpenVINO 实现 Paddle 算子 where_index 转换
提交内容
提交地址:https://github.com/openvinotoolkit/openvino
https://www.paddlepaddle.org.cn/documentation/docs/zh/api/paddle/topk_cn.html#topk
技术要求
参考文档
OpenVINO算子库文档:
https://github.com/openvinotoolkit/openvino/blob/master/docs/ops/opset9.md
OpenVINO算子参考实现:
https://github.com/openvinotoolkit/openvino/tree/master/src/core/reference/include/ngraph/runtime/reference
PaddlePaddle算子库文档:
https://www.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html
PaddlePaddle算子参考实现:
https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/fluid/tests/unittests
Paddle2ONNX算子映射参考代码
https://github.com/PaddlePaddle/Paddle2ONNX/tree/develop/paddle2onnx/legacy/op_mapper
https://github.com/PaddlePaddle/Paddle2ONNX/tree/develop/paddle2onnx/mapper
可以先生成测试模型用Paddle VisualDL查看paddle算子的输入输出以及属性: https://www.paddlepaddle.org.cn/paddle/visualdl/demo/graph
OpenVINO源码编译方法:
参考:https://github.com/openvinotoolkit/openvino/wiki
单测测试方法:
The text was updated successfully, but these errors were encountered: