shitianyi

shitianyi

Member Since 6 years ago

Experience Points
2
follower
Lessons Completed
6
follow
Lessons Completed
16
stars
Best Reply Awards
47
repos

14 contributions in the last year

Pinned
⚡ 谷歌全新开源人工智能系统TensorFlow官方文档中文版
⚡ Multifarious Scrapy examples.
⚡ Python资源大全中文版,包括:Web框架、网络爬虫、模板引擎、数据库、数据可视化、图片处理等,由伯乐在线持续更新。
⚡ C++ 资源大全中文版,标准库、Web应用框架、人工智能、数据库、图片处理、机器学习、日志、代码分析等
⚡ Some code snippets for sharing
⚡ The C++ Core Guidelines are a set of tried-and-true guidelines, rules, and best practices about coding in C++
Activity
Jan
15
4 days ago
Activity icon
issue

shitianyi issue alibaba/MNN

shitianyi
shitianyi

静态库调用MNN的vulkan后端失败

平台(如果交叉编译请再附上交叉编译目标平台):

android arm64-v8a ndk=20b api-level=21

Github版本:

1.2.2

编译方式:

cmake \
  -DCMAKE_TOOLCHAIN_FILE=${NDK_HOME}/build/cmake/android.toolchain.cmake \
  -DANDROID_STL=c++_static \
  -DCMAKE_EXPORT_COMPILE_COMMANDS=ON \
  -DANDROID_NATIVE_API_LEVEL=android-21 \
  -DANDROID_ABI=arm64-v8a \
  -DANDROID_NDK=${NDK_HOME} \
  -DMNN_BUILD_SHARED_LIBS=ON \
  -DMNN_BUILD_TOOLS=OFF \
  -DMNN_VULKAN=ON \
  -DCMAKE_BUILD_TYPE=Debug \
  -DCMAKE_INSTALL_PREFIX=${PREFIX} \
  ..

编译日志:

Build Log:

-- The ASM compiler identification is Clang
-- Found assembler: /home/zhangyuanzhen/toolchains/android-ndk-r20b/toolchains/llvm/prebuilt/linux-x86_64/bin/clang
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /home/zhangyuanzhen/toolchains/android-ndk-r20b/toolchains/llvm/prebuilt/linux-x86_64/bin/clang - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /home/zhangyuanzhen/toolchains/android-ndk-r20b/toolchains/llvm/prebuilt/linux-x86_64/bin/clang++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found PythonInterp: /usr/bin/python (found version "3.6.8") 
CMake Warning at CMakeLists.txt:86 (message):
  Close MNN_SEP_BUILD for static library


-- Use Threadpool, forbid openmp
-- >>>>>>>>>>>>>
-- MNN BUILD INFO:
--      System: Android
--      Processor: aarch64
--      Metal: OFF
--      OpenCL: OFF
--      OpenGL: OFF
--      Vulkan: ON
--      ARM82: OFF
--      oneDNN: OFF
--      TensorRT: OFF
--      CoreML: OFF
--      CUDA: OFF
--      OpenMP: OFF
--      BF16: 
--      ThreadPool: ON
--      Hidden: FALSE
--      Build Path: /home/zhangyuanzhen/3rd/mnn/build-arm64-v8a
-- Enabling AArch64 Assemblies
-- MNN_TARGETS MNNCore;MNNCV;MNNMath;MNNTransform;MNNUtils;MNNCPU;MNNARM64;MNN_Vulkan;MNN_Express;MNN
-- MNN_TARGETS MNNCore;MNNCV;MNNMath;MNNTransform;MNNUtils;MNNCPU;MNNARM64;MNN_Vulkan;MNN_Express
-- Configuring done
-- Generating done
-- Build files have been written to: /home/zhangyuanzhen/3rd/mnn/build-arm64-v8a

我在source/backend/vulkan/runtime/VulkanRuntime.cpp的gResistor的lambda表达式里添加了一行printf打印。 生成出的静态库libMNN.a无法调用vulkan后端,显示的是未找到。printf也没有打印。然而,我编出的动态库,应用程序链接是可以成功的。 我查了cmake的生成文件,在MNN.dir/flags.make里看见vulkan相关的.o文件已经打进去了。

Dec
14
1 month ago
push

shitianyi push shitianyi/MNN

shitianyi
shitianyi

feat(convert): add ROIPoolingOnnx Convert(the onnx model file export from torchvision.ops.roi_pool)

shitianyi
shitianyi

feat(test): add unit test for ROIPooling

shitianyi
shitianyi

Fix ConvDilateFuse bug for width pad not equal height pad

shitianyi
shitianyi

Merge pull request #1735 from no5-aaron-wu/dev_aaron_wu

add onnx convert and uint test for ROIPooling OP

shitianyi
shitianyi

Fix conv1d depthwise conv int8 calculation issue.

shitianyi
shitianyi

fix(op): increase compatibility of NCHW format for inputs1 in CPUROIPooling op

shitianyi
shitianyi

Merge pull request #1738 from no5-aaron-wu/dev_aaron_wu

increase compatibility of NCHW format for inputs1 in CPUROIPooling op

shitianyi
shitianyi
shitianyi
shitianyi

Fix compute small error for OnnxScatterND

shitianyi
shitianyi

[MNN:Bugfix] Fix bug for CPURaster for fuse singleConvert of dim == 3

shitianyi
shitianyi

Fix compile bug for bf16 when sse / neon is close

shitianyi
shitianyi

Merge pull request #1739 from jun-lv-17/fix-depthwiseconvint8-issue

Fix conv1d depthwise conv int8 calculation issue.

shitianyi
shitianyi

[Doc:Bugfix] Fix pic path error

shitianyi
shitianyi

Temply turn inputCount = 1's depthmulti to convolution

shitianyi
shitianyi

[MNN:Bugfix] Use fabsf instead of abs

shitianyi
shitianyi

fix: cli param keep input format.

shitianyi
shitianyi

Fix bug for onResize of CPURNNSequenceGRU

shitianyi
shitianyi

Merge pull request #1743 from Napoleon-Jm/dev

fix: cli param keep input format.

shitianyi
shitianyi

feat(convert): add ROIAlignOnnx convert(the onnx model file exported from torchvision.ops.roi_align)

shitianyi
shitianyi

fix(test): add new test case in ROIPoolingTest.cpp

commit sha: 82630820c04402e6282e7c717c4fc531c23808e2

push time in 1 month ago
Oct
29
2 months ago
Activity icon
fork

shitianyi forked goreliu/zshguide

⚡ Zsh 开发指南
shitianyi GNU General Public License v3.0 Updated
fork time in 2 months ago
Oct
25
2 months ago
Activity icon
fork

shitianyi forked nlohmann/json

⚡ JSON for Modern C++
shitianyi MIT License Updated
fork time in 2 months ago
Activity icon
fork

shitianyi forked xtensor-stack/xtl

⚡ The x template library
shitianyi BSD 3-Clause "New" or "Revised" License Updated
fork time in 2 months ago
Activity icon
fork

shitianyi forked xtensor-stack/xtensor

⚡ C++ tensors with broadcasting and lazy computing
shitianyi BSD 3-Clause "New" or "Revised" License Updated
fork time in 2 months ago
Activity icon
fork

shitianyi forked alibaba/MNN

⚡ MNN is a blazing fast, lightweight deep learning framework, battle-tested by business-critical use cases in Alibaba
shitianyi Updated
fork time in 2 months ago