[GitHub] [incubator-mxnet] aashudwivedi removed a comment on issue #14332: MXNet static library build results in error in centos, oracle linux and similar distros
aashudwivedi removed a comment on issue #14332: MXNet static library build results in error in centos, oracle linux and similar distros URL: https://github.com/apache/incubator-mxnet/issues/14332#issuecomment-470151109 Thanks for the help @lanking520 I am not able to use the libs which are built in ubuntu, in centos. It runs into an error for missing .so files. To build in ole7 / centos, I have added the following lines in the script build_lib.sh before the line `>&2 echo "Checking linked objects on libmxnet.so..."` ``` for libname in $(ls staticdeps/lib64/*.a | xargs -n 1 basename) do cp -fL staticdeps/lib64/$libname staticdeps/lib/$libname done cp -L /usr/lib64/libgfortran.so.3 lib/libgfortran.so.3 cp -L /usr/gcc-4.8.5/release/x86_64-unknown-linux-gnu/libquadmath/.libs/libquadmath.so lib/libquadmath.so.0 ``` which resolves the previous error. However the build still fails, with the error message : ``` /usr/bin/ld: skipping incompatible /lib/librt.so when searching for -lrt /usr/bin/ld: cannot find -lgfortran /usr/bin/ld: skipping incompatible /lib/libdl.so when searching for -ldl /usr/bin/ld: skipping incompatible /lib/libm.so when searching for -lm /usr/bin/ld: skipping incompatible /lib/libpthread.so when searching for -lpthread collect2: error: ld returned 1 exit status make: *** [lib/libmxnet.so] Error 1 make: *** Waiting for unfinished jobs /usr/bin/ld: skipping incompatible /lib/librt.so when searching for -lrt /usr/bin/ld: cannot find -lgfortran /usr/bin/ld: skipping incompatible /lib/libdl.so when searching for -ldl /usr/bin/ld: skipping incompatible /lib/libm.so when searching for -lm /usr/bin/ld: skipping incompatible /lib/libpthread.so when searching for -lpthread collect2: error: ld returned 1 exit status make: *** [bin/im2rec] Error 1 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] TaoLv commented on issue #14335: [MKLDNN] Question on installation and use of MKLDNN
TaoLv commented on issue #14335: [MKLDNN] Question on installation and use of MKLDNN URL: https://github.com/apache/incubator-mxnet/issues/14335#issuecomment-471408977 @dbsxdbsx could you post the steps for reproducing the build issue? Or typical steps for building mxnet from a Windows user perspective. I also sent an email to you for details. Try to answer part of the questions: > Q2: from office tutorial of gluonCV with C++, seems it is feasible to build with MKL+MKLDNN in cmd command. But with CMake Gui, USE_MKLDNN is forbidden, as it need (NOT MSVC) . WHY? It's possible to build MXNet with MKL/MKL-DNN from source. Please take a look at what we do in CI: https://github.com/apache/incubator-mxnet/blob/master/ci/docker/runtime_functions.sh#L553 https://github.com/apache/incubator-mxnet/blob/master/ci/docker/runtime_functions.sh#L708 I'm not a CMAKE GUI user, so not sure what's the problem on it. > Q4: I googled a lot on MKl with mxnet, and I found a discussion within mxnet team, with the content of the discussion and what I found in downloadMKLML.cmake, does it mean mxnet recommand install MKLDNN with JUST submodule of MKL at present(and it is still vague to decide whether submodule of MKL is needed when building mxnet from source)? Currently, it is much easier to get MKLML compared with to get full MKL. It has more friendly license to mxnet, smaller binary size and more convenient download source. So now it's the default behavior to download MKLML and link to it when mxnet is built with USE_MKLDNN=1. But you're right, there is no need to link MKLML if full MKL is installed and linked to mxnet. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn commented on issue #14027: Julia: rename `mx.clip` to `clamp` for `NDArray`
wkcn commented on issue #14027: Julia: rename `mx.clip` to `clamp` for `NDArray` URL: https://github.com/apache/incubator-mxnet/pull/14027#issuecomment-471397243 This PR has been merged. Thanks for your contribution! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet] branch master updated: Julia: rename `mx.clip` to `clamp` for `NDArray` (#14027)
This is an automated email from the ASF dual-hosted git repository. wkcn pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new af41af5 Julia: rename `mx.clip` to `clamp` for `NDArray` (#14027) af41af5 is described below commit af41af527221ad9b2c975b3962417b5fb5e1595b Author: Iblis Lin AuthorDate: Mon Mar 11 12:02:13 2019 +0800 Julia: rename `mx.clip` to `clamp` for `NDArray` (#14027) - in order to match Julia `Base.clamp` interface - depwarn for `mx.clip` included --- julia/NEWS.md | 7 +++--- julia/src/MXNet.jl | 2 -- julia/src/deprecated.jl | 8 +-- julia/src/ndarray/arithmetic.jl | 52 + julia/src/ndarray/remap.jl | 12 ++ julia/src/optimizer.jl | 2 +- julia/test/unittest/ndarray.jl | 12 +- 7 files changed, 60 insertions(+), 35 deletions(-) diff --git a/julia/NEWS.md b/julia/NEWS.md index 4a6c1a2..3cd6162 100644 --- a/julia/NEWS.md +++ b/julia/NEWS.md @@ -19,8 +19,6 @@ * Following material from `mx` module got exported (#TBD): * `NDArray` -* `clip()` -* `clip!()` * `context()` * `expand_dims()` * `@inplace` @@ -373,11 +371,12 @@ 99.9889 100.533 100.072 ``` -* Signature of `clip` changed, it doesn't require any keyword argument now. +* Signature of `clip` changed and renamed to `clamp`. + It doesn't require any keyword argument now. (#TBD) Before: `clip(x, a_min = -4, a_max = 4)` - After: `clip(x, -4, 4)` + After: `clamp(x, -4, 4)` ### Optimizer diff --git a/julia/src/MXNet.jl b/julia/src/MXNet.jl index 68663d1..70eda96 100644 --- a/julia/src/MXNet.jl +++ b/julia/src/MXNet.jl @@ -50,8 +50,6 @@ export SymbolicNode, # ndarray.jl export NDArray, - clip, - clip!, context, expand_dims, @inplace, diff --git a/julia/src/deprecated.jl b/julia/src/deprecated.jl index 70079b8..7c49b66 100644 --- a/julia/src/deprecated.jl +++ b/julia/src/deprecated.jl @@ -72,8 +72,6 @@ end @deprecate softmax(x::NDArray; axis = ndims(x)) softmax.(x, axis) @deprecate log_softmax(x::NDArray; axis = ndims(x)) log_softmax.(x, axis) -@deprecate clip(x; a_min = 0, a_max = 0) clip(x, a_min, a_max) - function broadcast_plus(x::NDArray, y::NDArray) @warn("broadcast_plus(x, y) is deprecated, use x .+ y instead.") x .+ y @@ -194,3 +192,9 @@ function empty(dims::Int...) "use `NDArray(undef, dims...)` instead.") NDArray(undef, dims...) end + +# replaced by Base.clamp +@deprecate clip(x::NDArray, lo::Real, hi::Real) clamp(x, lo, hi) +@deprecate clip!(x::NDArray, lo::Real, hi::Real) clamp!(x, lo, hi) +@deprecate clip(x; a_min = 0, a_max = 0) clamp(x, a_min, a_max) + diff --git a/julia/src/ndarray/arithmetic.jl b/julia/src/ndarray/arithmetic.jl index 60dde6b..4c467a2 100644 --- a/julia/src/ndarray/arithmetic.jl +++ b/julia/src/ndarray/arithmetic.jl @@ -218,40 +218,52 @@ broadcasted(::typeof(^), x::NDArray{T,N}, y::NDArray{T,N}) where {T,N} = broadcasted(::typeof(^), x::NDArray{T,N}, y::NDArray{T,M}) where {T,N,M} = _broadcast_power(x, y) -_nddoc[:clip] = _nddoc[:clip!] = """ -clip(x::NDArray, min, max) -clip!(x::NDArray, min, max) +clamp(x::NDArray, lo, hi) -Clips (limits) the values in `NDArray`. +Clamps (limits) the values in `NDArray`. Given an interval, values outside the interval are clipped to the interval edges. -Clipping `x` between `min` and `x` would be: +Clamping `x` between low `lo` and high `hi` would be: ```julia -clip(x, min_, max_) = max(min(x, max_), min_)) +clamp(x, lo, hi) = max(min(x, lo), hi)) ``` +The storage type of clip output depends on storage types of inputs and the +`lo`, `hi` parameter values: + +- clamp(default) -> default +- clamp(row_sparse, lo <= 0, hi >= 0) -> row_sparse +- clamp(csr, lo <= 0, hi >= 0) -> csr +- clamp(row_sparse, lo < 0, hi < 0) -> default +- clamp(row_sparse, lo > 0, hi > 0) -> default +- clamp(csr, lo < 0, hi < 0) -> csr +- clamp(csr, lo > 0, hi > 0) -> csr + +## Examples + ```jldoctest julia> x = NDArray(1:9); -julia> mx.clip(x, 2, 8)' +julia> clamp(x, 2, 8)' 1×9 mx.NDArray{Int64,2} @ CPU0: 2 2 3 4 5 6 7 8 8 -``` -The storage type of clip output depends on storage types of inputs and the -`min`, `max` parameter values: - -- clip(default) = default -- clip(row_sparse, min <= 0, max >= 0) = row_sparse -- clip(csr, min <= 0, max >= 0) = csr -- clip(row_sparse, min < 0, max < 0) = default -- clip(row_sparse, min > 0, max > 0) = default -- clip(csr, min < 0, max < 0) = csr -- clip(csr, min > 0, max > 0) = csr +julia> clamp(x, 8, 2)' +1×9 NDArray{Int64,2} @ CPU0: + 8 8 2 2 2 2 2 2 2 + ``` +""" +Base.clamp(x::NDArray, lo::Real, hi::Real) = _clamp(x, lo, hi) +@_remap _clamp(x::NDArray, lo::Real, hi::Real) clip(x; a_min = lo,
[GitHub] [incubator-mxnet] wkcn merged pull request #14027: Julia: rename `mx.clip` to `clamp` for `NDArray`
wkcn merged pull request #14027: Julia: rename `mx.clip` to `clamp` for `NDArray` URL: https://github.com/apache/incubator-mxnet/pull/14027 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn merged pull request #14190: [Flaky Test] Python3: MKLDNN-GPU test_kvstore_gpu.test_rsp_push_pull
wkcn merged pull request #14190: [Flaky Test] Python3: MKLDNN-GPU test_kvstore_gpu.test_rsp_push_pull URL: https://github.com/apache/incubator-mxnet/pull/14190 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet] branch master updated: Flaky test https://github.com/apache/incubator-mxnet/issues/14189 (#14190)
This is an automated email from the ASF dual-hosted git repository. wkcn pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 47d4d66 Flaky test https://github.com/apache/incubator-mxnet/issues/14189 (#14190) 47d4d66 is described below commit 47d4d66ac477ae560a0d34d06e9145ce422d0e9c Author: Chance Bair AuthorDate: Mon Mar 11 05:00:37 2019 +0100 Flaky test https://github.com/apache/incubator-mxnet/issues/14189 (#14190) --- tests/python/gpu/test_kvstore_gpu.py | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/python/gpu/test_kvstore_gpu.py b/tests/python/gpu/test_kvstore_gpu.py index 8ff8752..23bab53 100644 --- a/tests/python/gpu/test_kvstore_gpu.py +++ b/tests/python/gpu/test_kvstore_gpu.py @@ -43,6 +43,7 @@ def init_kv_with_str(stype='default', kv_type='local'): # Not reproducible, so this test is back on random seeds. @with_seed() @unittest.skipIf(mx.context.num_gpus() < 2, "test_rsp_push_pull needs more than 1 GPU") +@unittest.skip("Flaky test https://github.com/apache/incubator-mxnet/issues/14189;) def test_rsp_push_pull(): def check_rsp_push_pull(kv_type, sparse_pull, is_push_cpu=True): kv = init_kv_with_str('row_sparse', kv_type)
[GitHub] [incubator-mxnet] wkcn commented on issue #14190: [Flaky Test] Python3: MKLDNN-GPU test_kvstore_gpu.test_rsp_push_pull
wkcn commented on issue #14190: [Flaky Test] Python3: MKLDNN-GPU test_kvstore_gpu.test_rsp_push_pull URL: https://github.com/apache/incubator-mxnet/pull/14190#issuecomment-471397039 The PR has been merged. Thanks for your contribution! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet] branch master updated: fix engine crash in shutdown phase (#14382)
This is an automated email from the ASF dual-hosted git repository. wkcn pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 4f5cba5 fix engine crash in shutdown phase (#14382) 4f5cba5 is described below commit 4f5cba59dd44b91bfdd59025106e3bc116600a86 Author: Wang Jiajun AuthorDate: Mon Mar 11 11:56:53 2019 +0800 fix engine crash in shutdown phase (#14382) * fix engine crash in shutdown phase * fix lint * Revert "Bypass ThreadedEngine in test_operator_gpu.py:test_convolution_multiple_streams. (#14338)" This reverts commit d6eafca2555b58746f51052fdce96a264d02a84a. --- src/engine/threaded_engine.h | 9 + tests/python/gpu/test_operator_gpu.py | 12 +--- 2 files changed, 10 insertions(+), 11 deletions(-) diff --git a/src/engine/threaded_engine.h b/src/engine/threaded_engine.h index ab06ca1..640eac4 100644 --- a/src/engine/threaded_engine.h +++ b/src/engine/threaded_engine.h @@ -30,6 +30,7 @@ #include #include #include +#include #include #include #include @@ -306,6 +307,8 @@ class ThreadedEngine : public Engine { objpool_varblk_ref_ = common::ObjectPool::_GetSharedRef(); objpool_var_ref_= common::ObjectPool::_GetSharedRef(); +storage_ref_ = Storage::_GetSharedRef(); + // Get a ref to the profiler so that it doesn't get killed before us profiler::Profiler::Get(_); } @@ -549,6 +552,12 @@ class ThreadedEngine : public Engine { std::shared_ptr > objpool_varblk_ref_; std::shared_ptr > objpool_var_ref_; + /*! + * \brief Async destruction of some objects is relied on storage, + * prevent it from being destructed too early + */ + std::shared_ptr storage_ref_; + #if MXNET_USE_CUDA /*! \brief Number of GPU devices available */ std::atomic device_count_{-1}; diff --git a/tests/python/gpu/test_operator_gpu.py b/tests/python/gpu/test_operator_gpu.py index 7d7c2ed..c12c94b 100644 --- a/tests/python/gpu/test_operator_gpu.py +++ b/tests/python/gpu/test_operator_gpu.py @@ -547,18 +547,8 @@ def _conv_with_num_streams(seed): @with_seed() def test_convolution_multiple_streams(): -engines = ['NaiveEngine', 'ThreadedEngine', 'ThreadedEnginePerDevice'] - -if os.getenv('MXNET_ENGINE_TYPE') is not None: -engines = [os.getenv('MXNET_ENGINE_TYPE'),] -print("Only running against '%s'" % engines[0], file=sys.stderr, end='') -# Remove this else clause when the ThreadedEngine can handle this test -else: -engines.remove('ThreadedEngine') -print("SKIP: 'ThreadedEngine', only running against %s" % engines, file=sys.stderr, end='') - for num_streams in [1, 2]: -for engine in engines: +for engine in ['NaiveEngine', 'ThreadedEngine', 'ThreadedEnginePerDevice']: print("Starting engine %s with %d streams." % (engine, num_streams), file=sys.stderr) run_in_spawned_process(_conv_with_num_streams, {'MXNET_GPU_WORKER_NSTREAMS' : num_streams, 'MXNET_ENGINE_TYPE' : engine})
[GitHub] [incubator-mxnet] dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471390025 @wkcn,for context of error message, the blew is the total errors: ``` 严重性 代码 说明 项目 文件 行 禁止显示状态 错误 MSB3721 命令“"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_50,code=\"sm_50,compute_50\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\x86_amd64" -x cu -I"C:\Program Files (x86)\IntelSWTools\compilers_and_libraries_2019.2.190\windows\mkl\include" -IC:\mxnet\3rdparty\mkldnn\include -IC:\mxnet\include -IC:\mxnet\src -IC:\mxnet\3rdparty\mshadow -IC:\mxnet\3rdparty\cub -IC:\mxnet\3rdparty\tvm\nnvm\include -IC:\mxnet\3rdparty\tvm\include -I"C:\mxnet\3rdparty\dmlc-core\include" -IC:\mxnet\3rdparty\dlpack\include -IC:\opencv4\build\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" --keep-dir x64\Release -maxrregcount=0 --machine 64 --compile -cudart static -Xcompiler="/MP /bigobj -openmp -Ob2 -Gy" -DNDEBUG -DWIN32_LEAN_AND_MEAN -DDMLC_USE_CXX11 -DMSHADOW_IN_CXX11 -D_SCL_SECURE_NO_WARNINGS -D_CRT_SECURE_NO_WARNINGS -DMXNET_EXPORTS -DNNVM_EXPORTS -DDMLC_STRICT_CXX11 -DNOMINMAX -DUSE_MKL=1 -DCUB_MKL=1 -DMXNET_USE_MKLDNN=1 -DMSHADOW_USE_CUDA=1 -DMXNET_USE_NCCL=0 -DUSE_MKL -DUSE_CBLAS -DMSHADOW_USE_CBLAS=0 -DMSHADOW_USE_MKL=1 -DMXNET_USE_BLAS_MKL=1 -DMXNET_USE_OPENCV=1 -DMXNET_USE_OPENMP=1 -DMXNET_USE_LAPACK=1 -DUSE_CUDNN -DMSHADOW_USE_CUDNN=1 -DMXNET_ENABLE_CUDA_RTC=1 -DMXNET_USE_CUDA=1 -D"CMAKE_INTDIR=\"Release\"" -Dmxnet_EXPORTS -DNDEBUG -DWIN32_LEAN_AND_MEAN -DDMLC_USE_CXX11 -DMSHADOW_IN_CXX11 -D_SCL_SECURE_NO_WARNINGS -D_CRT_SECURE_NO_WARNINGS -DMXNET_EXPORTS -DNNVM_EXPORTS -DDMLC_STRICT_CXX11 -DNOMINMAX -DUSE_MKL=1 -DCUB_MKL=1 -DMXNET_USE_MKLDNN=1 -DMSHADOW_USE_CUDA=1 -DMXNET_USE_NCCL=0 -DUSE_MKL -DUSE_CBLAS -DMSHADOW_USE_CBLAS=0 -DMSHADOW_USE_MKL=1 -DMXNET_USE_BLAS_MKL=1 -DMXNET_USE_OPENCV=1 -DMXNET_USE_OPENMP=1 -DMXNET_USE_LAPACK=1 -DUSE_CUDNN -DMSHADOW_USE_CUDNN=1 -DMXNET_ENABLE_CUDA_RTC=1 -DMXNET_USE_CUDA=1 -D"CMAKE_INTDIR=\"Release\"" -Dmxnet_EXPORTS -D_WINDLL -D_MBCS -Xcompiler "/EHsc /W1 /nologo /O2 /Fdmxnet.dir\Release\vc140.pdb /FS /Zi /MT " -o mxnet.dir\Release\/src/operator/image/resize.cu.obj "C:\mxnet\src\operator\image\resize.cu"”已退出,返回代码为 1。mxnet C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V140\BuildCustomizations\CUDA 10.1.targets757 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of
[GitHub] [incubator-mxnet] dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471390025 @wkcn,for context of error message, the blew is the total errors: ``` 严重性 代码 说明 项目 文件 行 禁止显示状态 错误 MSB3721 命令“"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_50,code=\"sm_50,compute_50\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\x86_amd64" -x cu -I"C:\Program Files (x86)\IntelSWTools\compilers_and_libraries_2019.2.190\windows\mkl\include" -IC:\mxnet\3rdparty\mkldnn\include -IC:\mxnet\include -IC:\mxnet\src -IC:\mxnet\3rdparty\mshadow -IC:\mxnet\3rdparty\cub -IC:\mxnet\3rdparty\tvm\nnvm\include -IC:\mxnet\3rdparty\tvm\include -I"C:\mxnet\3rdparty\dmlc-core\include" -IC:\mxnet\3rdparty\dlpack\include -IC:\opencv4\build\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" --keep-dir x64\Release -maxrregcount=0 --machine 64 --compile -cudart static -Xcompiler="/MP /bigobj -openmp -Ob2 -Gy" -DNDEBUG -DWIN32_LEAN_AND_MEAN -DDMLC_USE_CXX11 -DMSHADOW_IN_CXX11 -D_SCL_SECURE_NO_WARNINGS -D_CRT_SECURE_NO_WARNINGS -DMXNET_EXPORTS -DNNVM_EXPORTS -DDMLC_STRICT_CXX11 -DNOMINMAX -DUSE_MKL=1 -DCUB_MKL=1 -DMXNET_USE_MKLDNN=1 -DMSHADOW_USE_CUDA=1 -DMXNET_USE_NCCL=0 -DUSE_MKL -DUSE_CBLAS -DMSHADOW_USE_CBLAS=0 -DMSHADOW_USE_MKL=1 -DMXNET_USE_BLAS_MKL=1 -DMXNET_USE_OPENCV=1 -DMXNET_USE_OPENMP=1 -DMXNET_USE_LAPACK=1 -DUSE_CUDNN -DMSHADOW_USE_CUDNN=1 -DMXNET_ENABLE_CUDA_RTC=1 -DMXNET_USE_CUDA=1 -D"CMAKE_INTDIR=\"Release\"" -Dmxnet_EXPORTS -DNDEBUG -DWIN32_LEAN_AND_MEAN -DDMLC_USE_CXX11 -DMSHADOW_IN_CXX11 -D_SCL_SECURE_NO_WARNINGS -D_CRT_SECURE_NO_WARNINGS -DMXNET_EXPORTS -DNNVM_EXPORTS -DDMLC_STRICT_CXX11 -DNOMINMAX -DUSE_MKL=1 -DCUB_MKL=1 -DMXNET_USE_MKLDNN=1 -DMSHADOW_USE_CUDA=1 -DMXNET_USE_NCCL=0 -DUSE_MKL -DUSE_CBLAS -DMSHADOW_USE_CBLAS=0 -DMSHADOW_USE_MKL=1 -DMXNET_USE_BLAS_MKL=1 -DMXNET_USE_OPENCV=1 -DMXNET_USE_OPENMP=1 -DMXNET_USE_LAPACK=1 -DUSE_CUDNN -DMSHADOW_USE_CUDNN=1 -DMXNET_ENABLE_CUDA_RTC=1 -DMXNET_USE_CUDA=1 -D"CMAKE_INTDIR=\"Release\"" -Dmxnet_EXPORTS -D_WINDLL -D_MBCS -Xcompiler "/EHsc /W1 /nologo /O2 /Fdmxnet.dir\Release\vc140.pdb /FS /Zi /MT " -o mxnet.dir\Release\/src/operator/image/resize.cu.obj "C:\mxnet\src\operator\image\resize.cu"”已退出,返回代码为 1。mxnet C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V140\BuildCustomizations\CUDA 10.1.targets757 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of "std::true_type (std::integral_constant<__nv_bool, false> *)" (the meaning of a name may have changed since the template declaration -- the type of the template is "std::true_type (std::is_same))>::type, void>::type *)") mxnet c:\opencv4\build\include\opencv2\core\cvstd_wrapper.hpp 49 错误 template instantiation resulted in unexpected function type of
[GitHub] [incubator-mxnet] wkcn commented on issue #14382: fix engine crash in shutdown phase
wkcn commented on issue #14382: fix engine crash in shutdown phase URL: https://github.com/apache/incubator-mxnet/pull/14382#issuecomment-471396556 The PR has been merged. Thanks for your contribution! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn merged pull request #14382: fix engine crash in shutdown phase
wkcn merged pull request #14382: fix engine crash in shutdown phase URL: https://github.com/apache/incubator-mxnet/pull/14382 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn merged pull request #14356: support leading dimension of -1 in ravel/unravel
wkcn merged pull request #14356: support leading dimension of -1 in ravel/unravel URL: https://github.com/apache/incubator-mxnet/pull/14356 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet] branch master updated: support leading dimension of -1 in ravel/unravel (#14356)
This is an automated email from the ASF dual-hosted git repository. wkcn pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 35098b8 support leading dimension of -1 in ravel/unravel (#14356) 35098b8 is described below commit 35098b8ad27e5151fc266fbf1a766d937bd9ec8e Author: moin AuthorDate: Mon Mar 11 04:53:30 2019 +0100 support leading dimension of -1 in ravel/unravel (#14356) --- src/operator/tensor/ravel.cc | 6 -- src/operator/tensor/ravel.h| 3 ++- tests/python/unittest/test_operator.py | 7 +++ 3 files changed, 13 insertions(+), 3 deletions(-) diff --git a/src/operator/tensor/ravel.cc b/src/operator/tensor/ravel.cc index 0a66ea8..94d79c7 100644 --- a/src/operator/tensor/ravel.cc +++ b/src/operator/tensor/ravel.cc @@ -31,12 +31,13 @@ DMLC_REGISTER_PARAMETER(RavelParam); NNVM_REGISTER_OP(_ravel_multi_index) .add_alias("ravel_multi_index") -.describe(R"code(Converts a batch of index arrays into an array of flat indices. The operator follows numpy conventions so a single multi index is given by a column of the input matrix. +.describe(R"code(Converts a batch of index arrays into an array of flat indices. The operator follows numpy conventions so a single multi index is given by a column of the input matrix. The leading dimension may be left unspecified by using -1 as placeholder. Examples:: A = [[3,6,6],[4,5,1]] ravel(A, shape=(7,6)) = [22,41,37] + ravel(A, shape=(-1,6)) = [22,41,37] )code" ADD_FILELINE) .set_num_inputs(1) @@ -55,12 +56,13 @@ Examples:: NNVM_REGISTER_OP(_unravel_index) .add_alias("unravel_index") -.describe(R"code(Converts an array of flat indices into a batch of index arrays. The operator follows numpy conventions so a single multi index is given by a column of the output matrix. +.describe(R"code(Converts an array of flat indices into a batch of index arrays. The operator follows numpy conventions so a single multi index is given by a column of the output matrix. The leading dimension may be left unspecified by using -1 as placeholder. Examples:: A = [22,41,37] unravel(A, shape=(7,6)) = [[3,6,6],[4,5,1]] + unravel(A, shape=(-1,6)) = [[3,6,6],[4,5,1]] )code" ADD_FILELINE) .set_num_inputs(1) diff --git a/src/operator/tensor/ravel.h b/src/operator/tensor/ravel.h index 6d337dc..256fe33 100644 --- a/src/operator/tensor/ravel.h +++ b/src/operator/tensor/ravel.h @@ -110,11 +110,12 @@ struct unravel_index { DType *unravelled, DType *ravelled) { index_t idx(ravelled[i]); #pragma unroll -for (int j = ndim; j--; ) { +for (int j = ndim-1; j > 0; --j) { index_t tmp = idx / shape[j]; unravelled[i+j*N] = idx - tmp*shape[j]; idx = tmp; } +unravelled[i] = idx; } }; diff --git a/tests/python/unittest/test_operator.py b/tests/python/unittest/test_operator.py index 6bb8150..7169395 100644 --- a/tests/python/unittest/test_operator.py +++ b/tests/python/unittest/test_operator.py @@ -7106,6 +7106,13 @@ def test_ravel(): check_symbolic_forward(b, location={'a': data}, expected=[ravel_npy]) c = mx.sym.unravel_index(a, shape=shape) check_symbolic_forward(c, location={'a': ravel_npy}, expected=[data]) + # Test with leading dimension set to -1. + shape2 = shape + shape2 = (-1,)+shape[1:] + b = mx.sym.ravel_multi_index(a, shape=shape2) + check_symbolic_forward(b, location={'a': data}, expected=[ravel_npy]) + c = mx.sym.unravel_index(a, shape=shape2) + check_symbolic_forward(c, location={'a': ravel_npy}, expected=[data]) def test_context_num_gpus(): try:
[GitHub] [incubator-mxnet] wkcn commented on issue #14356: support leading dimension of -1 in ravel/unravel
wkcn commented on issue #14356: support leading dimension of -1 in ravel/unravel URL: https://github.com/apache/incubator-mxnet/pull/14356#issuecomment-471396126 The PR has been merged. Thanks for your contribution! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn closed issue #13862: [1.4.0] unravel_index no longer works with magic '-1' in shape parameter as in 1.3.1
wkcn closed issue #13862: [1.4.0] unravel_index no longer works with magic '-1' in shape parameter as in 1.3.1 URL: https://github.com/apache/incubator-mxnet/issues/13862 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] peterpaniff closed issue #14384: try to train ssdlite mobilenetv2, encounter the error.
peterpaniff closed issue #14384: try to train ssdlite mobilenetv2, encounter the error. URL: https://github.com/apache/incubator-mxnet/issues/14384 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] peterpaniff commented on issue #14384: try to train ssdlite mobilenetv2, encounter the error.
peterpaniff commented on issue #14384: try to train ssdlite mobilenetv2, encounter the error. URL: https://github.com/apache/incubator-mxnet/issues/14384#issuecomment-471395643 i solve it myself This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] ankkhedia commented on a change in pull request #14351: [MXNet-1348][WIP][Fit API]Adding CNN examples for fit() API
ankkhedia commented on a change in pull request #14351: [MXNet-1348][WIP][Fit API]Adding CNN examples for fit() API URL: https://github.com/apache/incubator-mxnet/pull/14351#discussion_r264084828 ## File path: example/gluon/estimator_example/alexnet.py ## @@ -0,0 +1,156 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# This example is inspired from +# https://github.com/d2l-ai/d2l-en/blob/master/chapter_convolutional-neural-networks/alexnet.md +# Model definition is from +# https://github.com/dmlc/gluon-cv/blob/master/gluoncv/model_zoo/alexnet.py + + +import os +import sys +import argparse +import mxnet as mx +from mxnet import gluon +from mxnet.gluon import nn, data +from mxnet.gluon.block import HybridBlock +from mxnet.gluon.estimator import estimator, event_handler + +def parse_args(): +''' +Command Line Interface +''' +parser = argparse.ArgumentParser(description='Train ResNet18 on Fashion-MNIST') +parser.add_argument('--batch-size', type=int, default=128, +help='training batch size per device (CPU/GPU).') +parser.add_argument('--num-epochs', type=int, default=1, +help='number of training epochs.') +parser.add_argument('--input-size', type=int, default=224, +help='size of the input image size. default is 224') +parser.add_argument('--lr', type=float, default=0.001, +help='learning rate. default is 0.001') +parser.add_argument('-j', '--num-workers', default=None, type=int, +help='number of preprocessing workers') +opt = parser.parse_args() +return opt + +class AlexNet(HybridBlock): Review comment: @abhinavs95 You can also use default arguments eg. learning rate , num_workers instead of taking them from command line arguments to make the example simple for the beginners This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] peterpaniff opened a new issue #14386: train ssdlite mobilenetv2, enconter the name error
peterpaniff opened a new issue #14386: train ssdlite mobilenetv2, enconter the name error URL: https://github.com/apache/incubator-mxnet/issues/14386 raise ValueError('Cannot find output that matches name \"%s\"' % index) ValueError: Cannot find output that matches name "mobilenetv20_features_relu61_relu6" but i visualize the pretrained mobilenetV2 structure, it does has the name. ![image](https://user-images.githubusercontent.com/6010392/54099213-0d412d80-43f3-11e9-91d6-4ccf30272969.png) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] ankkhedia commented on a change in pull request #14351: [MXNet-1348][WIP][Fit API]Adding CNN examples for fit() API
ankkhedia commented on a change in pull request #14351: [MXNet-1348][WIP][Fit API]Adding CNN examples for fit() API URL: https://github.com/apache/incubator-mxnet/pull/14351#discussion_r264084828 ## File path: example/gluon/estimator_example/alexnet.py ## @@ -0,0 +1,156 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# This example is inspired from +# https://github.com/d2l-ai/d2l-en/blob/master/chapter_convolutional-neural-networks/alexnet.md +# Model definition is from +# https://github.com/dmlc/gluon-cv/blob/master/gluoncv/model_zoo/alexnet.py + + +import os +import sys +import argparse +import mxnet as mx +from mxnet import gluon +from mxnet.gluon import nn, data +from mxnet.gluon.block import HybridBlock +from mxnet.gluon.estimator import estimator, event_handler + +def parse_args(): +''' +Command Line Interface +''' +parser = argparse.ArgumentParser(description='Train ResNet18 on Fashion-MNIST') +parser.add_argument('--batch-size', type=int, default=128, +help='training batch size per device (CPU/GPU).') +parser.add_argument('--num-epochs', type=int, default=1, +help='number of training epochs.') +parser.add_argument('--input-size', type=int, default=224, +help='size of the input image size. default is 224') +parser.add_argument('--lr', type=float, default=0.001, +help='learning rate. default is 0.001') +parser.add_argument('-j', '--num-workers', default=None, type=int, +help='number of preprocessing workers') +opt = parser.parse_args() +return opt + +class AlexNet(HybridBlock): Review comment: @abhinavs95 You can also use default arguments eg. learning rate , num_workers to make the example simple for the beginners This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] stereomatchingkiss commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10
stereomatchingkiss commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10 URL: https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-471394903 > To reproduce the problem on the Windows Any chances to add a travis build for windows(if you have enough of resources) like #14370 suggest? One with mkl + cuda Another one with mkl This way whenever you commit the codes written on x platform(mac osx, linux etc), you can ensure your new codes do not break anything on windows. It would be better if you add travis build for every platforms you plan to support, like opencv. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] ankkhedia commented on a change in pull request #14346: [MXNet-1334][WIP][Fit API]base class for estimator and eventhandler
ankkhedia commented on a change in pull request #14346: [MXNet-1334][WIP][Fit API]base class for estimator and eventhandler URL: https://github.com/apache/incubator-mxnet/pull/14346#discussion_r264084500 ## File path: python/mxnet/gluon/estimator/estimator.py ## @@ -0,0 +1,203 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# coding: utf-8 +# pylint: disable=wildcard-import +"""Gluon Estimator""" + + +import warnings + +from .event_handler import LoggingHandler +from ... import * +from ... import gluon, autograd +from ...context import cpu, gpu, num_gpus +from ...metric import EvalMetric, Loss + +__all__ = ['Estimator'] + + +class Estimator(object): +""" +Estimator Class for easy model training +TODO: update doc +""" + +def __init__(self, net, + loss=None, + metrics=None, + initializer=None, + trainers=None, + context=None): + +self.net = net +if isinstance(loss, gluon.loss.Loss): +self.loss = [loss] +else: +self.loss = loss or [] +if isinstance(metrics, EvalMetric): +self.metrics = [metrics] +else: +self.metrics = metrics or [] + +self.initializer = initializer +# store training statistics +self.train_stats = {} +self.train_stats['epochs'] = [] +self.train_stats['learning_rate'] = [] +# time used for each epoch +self.train_stats['step'] = '' +for metric in self.metrics: +# record a history of metrics over each epoch +self.train_stats['train_' + metric.name] = [] +# only record the latest metric numbers after each batch +self.train_stats['batch_' + metric.name] = 0. +self.loss_metrics = [] +# using the metric wrapper for loss to record loss value +for loss in self.loss: +self.loss_metrics.append(Loss(loss.name)) +self.train_stats['train_' + loss.name] = [] +# only record the latest loss numbers after each batch +self.train_stats['batch_' + loss.name] = 0. + +# handle context +if isinstance(context, Context): +self.context = [context] +if not context: +if num_gpus() > 0: +# only use 1 GPU by default +if num_gpus() > 1: +warnings.warn("You have multiple GPUs, gpu(0) will be used by default." + "To utilize all your GPUs, specify context as a list of gpus, e.g. context=[mx.gpu(0), mx.gpu(2)] ") +self.context = [gpu(0)] +else: +self.context = [cpu()] + +# initialize the network +if self.initializer: +if self._is_initialized(): +# if already initialized, re-init with user specified initializer +warnings.warn("You have already initialized your net, it will be forced re-initialized " + "with the initializer you speficied. You don't need to pass initializer if you alraedy initialized your net.") +self.net.initialize(init=self.initializer, ctx=self.context, force_reinit=True) +else: +# initialize with user specified initializer +self.net.initialize(init=self.initializer, ctx=self.context, force_reinit=False) +else: +if not self._is_initialized(): +self.net.initialize(ctx=self.context) + +# handle trainers +if isinstance(trainers, gluon.Trainer): +self.trainers = [trainers] +else: +self.trainers = trainers or [] +if not self.trainers: +warnings.warn("No trainer specified, default SGD optimizer with learning rate 0.001 is used.") +self.trainers = [gluon.Trainer(self.net.collect_params(), 'sgd', {'learning_rate': 0.001})] + +def _is_initialized(self): +param_dict = self.net.collect_params() +for param in param_dict: +try: +param_dict[param].list_ctx() +except RuntimeError: +
[GitHub] [incubator-mxnet] ankkhedia commented on a change in pull request #14346: [MXNet-1334][WIP][Fit API]base class for estimator and eventhandler
ankkhedia commented on a change in pull request #14346: [MXNet-1334][WIP][Fit API]base class for estimator and eventhandler URL: https://github.com/apache/incubator-mxnet/pull/14346#discussion_r264084052 ## File path: python/mxnet/gluon/estimator/estimator.py ## @@ -0,0 +1,203 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# coding: utf-8 +# pylint: disable=wildcard-import +"""Gluon Estimator""" + + +import warnings + +from .event_handler import LoggingHandler +from ... import * +from ... import gluon, autograd +from ...context import cpu, gpu, num_gpus +from ...metric import EvalMetric, Loss + +__all__ = ['Estimator'] + + +class Estimator(object): +""" +Estimator Class for easy model training +TODO: update doc +""" + +def __init__(self, net, + loss=None, + metrics=None, + initializer=None, + trainers=None, + context=None): + +self.net = net +if isinstance(loss, gluon.loss.Loss): +self.loss = [loss] +else: +self.loss = loss or [] +if isinstance(metrics, EvalMetric): +self.metrics = [metrics] +else: +self.metrics = metrics or [] + +self.initializer = initializer +# store training statistics +self.train_stats = {} +self.train_stats['epochs'] = [] +self.train_stats['learning_rate'] = [] +# time used for each epoch +self.train_stats['step'] = '' +for metric in self.metrics: +# record a history of metrics over each epoch +self.train_stats['train_' + metric.name] = [] +# only record the latest metric numbers after each batch +self.train_stats['batch_' + metric.name] = 0. +self.loss_metrics = [] +# using the metric wrapper for loss to record loss value +for loss in self.loss: +self.loss_metrics.append(Loss(loss.name)) +self.train_stats['train_' + loss.name] = [] +# only record the latest loss numbers after each batch +self.train_stats['batch_' + loss.name] = 0. + +# handle context +if isinstance(context, Context): +self.context = [context] +if not context: +if num_gpus() > 0: +# only use 1 GPU by default +if num_gpus() > 1: +warnings.warn("You have multiple GPUs, gpu(0) will be used by default." + "To utilize all your GPUs, specify context as a list of gpus, e.g. context=[mx.gpu(0), mx.gpu(2)] ") +self.context = [gpu(0)] +else: +self.context = [cpu()] + +# initialize the network +if self.initializer: +if self._is_initialized(): +# if already initialized, re-init with user specified initializer +warnings.warn("You have already initialized your net, it will be forced re-initialized " + "with the initializer you speficied. You don't need to pass initializer if you alraedy initialized your net.") +self.net.initialize(init=self.initializer, ctx=self.context, force_reinit=True) +else: +# initialize with user specified initializer +self.net.initialize(init=self.initializer, ctx=self.context, force_reinit=False) +else: +if not self._is_initialized(): +self.net.initialize(ctx=self.context) + +# handle trainers +if isinstance(trainers, gluon.Trainer): +self.trainers = [trainers] +else: +self.trainers = trainers or [] +if not self.trainers: Review comment: Are we dealing with multiple trainer case over here (e.g. Multi task classification) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With
[GitHub] [incubator-mxnet] dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471390025 @wkcn,for context of error message, not feasible at present, I would try to rebuilt it and get it soon. for code of `cvstd_wrapper.hpp` in OPENCV4.0: ``` // This file is part of OpenCV project. // It is subject to the license terms in the LICENSE file found in the top-level directory // of this distribution and at http://opencv.org/license.html. #ifndef OPENCV_CORE_CVSTD_WRAPPER_HPP #define OPENCV_CORE_CVSTD_WRAPPER_HPP #include "opencv2/core/cvdef.h" #include #include // std::shared_ptr #include // std::enable_if namespace cv { using std::nullptr_t; //! @addtogroup core_basic //! @{ #ifdef CV_DOXYGEN template using Ptr = std::shared_ptr<_Tp>; // In ideal world it should look like this, but we need some compatibility workarounds below template static inline Ptr<_Tp> makePtr(const A1&... a1) { return std::make_shared<_Tp>(a1...); } #else // cv::Ptr with compatibility workarounds // It should be defined for C-API types only. // C++ types should use regular "delete" operator. template struct DefaultDeleter; #if 0 { void operator()(Y* p) const; }; #endif namespace sfinae { template struct has_parenthesis_operator { private: template static CV_CONSTEXPR std::true_type check(typename std::is_same().operator()(std::declval()...))>::type, Ret>::type*); template static CV_CONSTEXPR std::false_type check(...); typedef decltype(check(0)) type; public: static CV_CONSTEXPR bool value = type::value; }; } // namespace sfinae template struct has_custom_delete : public std::false_type {}; template struct has_custom_delete, void, T*>::value >::type > : public std::true_type {}; template struct Ptr : public std::shared_ptr { #if 0 using std::shared_ptr::shared_ptr; // GCC 5.x can't handle this #else inline Ptr() CV_NOEXCEPT : std::shared_ptr() {} inline Ptr(nullptr_t) CV_NOEXCEPT : std::shared_ptr(nullptr) {} template inline Ptr(Y* p, D d) : std::shared_ptr(p, d) {} template inline Ptr(nullptr_t, D d) : std::shared_ptr(nullptr, d) {} template inline Ptr(const Ptr& r, T* ptr) CV_NOEXCEPT : std::shared_ptr(r, ptr) {} inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} template inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} template inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} #endif inline Ptr(const std::shared_ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(std::shared_ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} // Overload with custom DefaultDeleter: Ptr(...) template inline Ptr(const std::true_type&, Y* ptr) : std::shared_ptr(ptr, DefaultDeleter()) {} // Overload without custom deleter: Ptr(...); template inline Ptr(const std::false_type&, Y* ptr) : std::shared_ptr(ptr) {} template inline Ptr(Y* ptr) : Ptr(has_custom_delete(), ptr) {} // Overload with custom DefaultDeleter: Ptr(...) template inline void reset(const std::true_type&, Y* ptr) { std::shared_ptr::reset(ptr, DefaultDeleter()); } // Overload without custom deleter: Ptr(...); template inline void reset(const std::false_type&, Y* ptr) { std::shared_ptr::reset(ptr); } template inline void reset(Y* ptr) { Ptr::reset(has_custom_delete(), ptr); } template void reset(Y* ptr, Deleter d) { std::shared_ptr::reset(ptr, d); } void reset() CV_NOEXCEPT { std::shared_ptr::reset(); } Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } template inline Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } T* operator->() const CV_NOEXCEPT { return std::shared_ptr::get();} typename std::add_lvalue_reference::type operator*() const CV_NOEXCEPT { return *std::shared_ptr::get(); } // OpenCV 3.x methods (not a part of standart C++ library) inline void release() { std::shared_ptr::reset(); } inline operator T* () const { return std::shared_ptr::get(); } inline bool empty() const { return std::shared_ptr::get() == nullptr; } template inline Ptr staticCast() const CV_NOEXCEPT { return std::static_pointer_cast(*this); } template inline Ptr constCast() const CV_NOEXCEPT { return std::const_pointer_cast(*this); } template inline Ptr dynamicCast() const
[GitHub] [incubator-mxnet] dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx edited a comment on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471390025 @wkcn,for context of error message, not feasible at present, I would try to rebuilt it and get it soon. for code in OPENCV4.0: ``` // This file is part of OpenCV project. // It is subject to the license terms in the LICENSE file found in the top-level directory // of this distribution and at http://opencv.org/license.html. #ifndef OPENCV_CORE_CVSTD_WRAPPER_HPP #define OPENCV_CORE_CVSTD_WRAPPER_HPP #include "opencv2/core/cvdef.h" #include #include // std::shared_ptr #include // std::enable_if namespace cv { using std::nullptr_t; //! @addtogroup core_basic //! @{ #ifdef CV_DOXYGEN template using Ptr = std::shared_ptr<_Tp>; // In ideal world it should look like this, but we need some compatibility workarounds below template static inline Ptr<_Tp> makePtr(const A1&... a1) { return std::make_shared<_Tp>(a1...); } #else // cv::Ptr with compatibility workarounds // It should be defined for C-API types only. // C++ types should use regular "delete" operator. template struct DefaultDeleter; #if 0 { void operator()(Y* p) const; }; #endif namespace sfinae { template struct has_parenthesis_operator { private: template static CV_CONSTEXPR std::true_type check(typename std::is_same().operator()(std::declval()...))>::type, Ret>::type*); template static CV_CONSTEXPR std::false_type check(...); typedef decltype(check(0)) type; public: static CV_CONSTEXPR bool value = type::value; }; } // namespace sfinae template struct has_custom_delete : public std::false_type {}; template struct has_custom_delete, void, T*>::value >::type > : public std::true_type {}; template struct Ptr : public std::shared_ptr { #if 0 using std::shared_ptr::shared_ptr; // GCC 5.x can't handle this #else inline Ptr() CV_NOEXCEPT : std::shared_ptr() {} inline Ptr(nullptr_t) CV_NOEXCEPT : std::shared_ptr(nullptr) {} template inline Ptr(Y* p, D d) : std::shared_ptr(p, d) {} template inline Ptr(nullptr_t, D d) : std::shared_ptr(nullptr, d) {} template inline Ptr(const Ptr& r, T* ptr) CV_NOEXCEPT : std::shared_ptr(r, ptr) {} inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} template inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} template inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} #endif inline Ptr(const std::shared_ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(std::shared_ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} // Overload with custom DefaultDeleter: Ptr(...) template inline Ptr(const std::true_type&, Y* ptr) : std::shared_ptr(ptr, DefaultDeleter()) {} // Overload without custom deleter: Ptr(...); template inline Ptr(const std::false_type&, Y* ptr) : std::shared_ptr(ptr) {} template inline Ptr(Y* ptr) : Ptr(has_custom_delete(), ptr) {} // Overload with custom DefaultDeleter: Ptr(...) template inline void reset(const std::true_type&, Y* ptr) { std::shared_ptr::reset(ptr, DefaultDeleter()); } // Overload without custom deleter: Ptr(...); template inline void reset(const std::false_type&, Y* ptr) { std::shared_ptr::reset(ptr); } template inline void reset(Y* ptr) { Ptr::reset(has_custom_delete(), ptr); } template void reset(Y* ptr, Deleter d) { std::shared_ptr::reset(ptr, d); } void reset() CV_NOEXCEPT { std::shared_ptr::reset(); } Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } template inline Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } T* operator->() const CV_NOEXCEPT { return std::shared_ptr::get();} typename std::add_lvalue_reference::type operator*() const CV_NOEXCEPT { return *std::shared_ptr::get(); } // OpenCV 3.x methods (not a part of standart C++ library) inline void release() { std::shared_ptr::reset(); } inline operator T* () const { return std::shared_ptr::get(); } inline bool empty() const { return std::shared_ptr::get() == nullptr; } template inline Ptr staticCast() const CV_NOEXCEPT { return std::static_pointer_cast(*this); } template inline Ptr constCast() const CV_NOEXCEPT { return std::const_pointer_cast(*this); } template inline Ptr dynamicCast() const CV_NOEXCEPT { return
[GitHub] [incubator-mxnet] dbsxdbsx commented on issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx commented on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471390025 for code in OPENCV4.0 ``` // This file is part of OpenCV project. // It is subject to the license terms in the LICENSE file found in the top-level directory // of this distribution and at http://opencv.org/license.html. #ifndef OPENCV_CORE_CVSTD_WRAPPER_HPP #define OPENCV_CORE_CVSTD_WRAPPER_HPP #include "opencv2/core/cvdef.h" #include #include // std::shared_ptr #include // std::enable_if namespace cv { using std::nullptr_t; //! @addtogroup core_basic //! @{ #ifdef CV_DOXYGEN template using Ptr = std::shared_ptr<_Tp>; // In ideal world it should look like this, but we need some compatibility workarounds below template static inline Ptr<_Tp> makePtr(const A1&... a1) { return std::make_shared<_Tp>(a1...); } #else // cv::Ptr with compatibility workarounds // It should be defined for C-API types only. // C++ types should use regular "delete" operator. template struct DefaultDeleter; #if 0 { void operator()(Y* p) const; }; #endif namespace sfinae { template struct has_parenthesis_operator { private: template static CV_CONSTEXPR std::true_type check(typename std::is_same().operator()(std::declval()...))>::type, Ret>::type*); template static CV_CONSTEXPR std::false_type check(...); typedef decltype(check(0)) type; public: static CV_CONSTEXPR bool value = type::value; }; } // namespace sfinae template struct has_custom_delete : public std::false_type {}; template struct has_custom_delete, void, T*>::value >::type > : public std::true_type {}; template struct Ptr : public std::shared_ptr { #if 0 using std::shared_ptr::shared_ptr; // GCC 5.x can't handle this #else inline Ptr() CV_NOEXCEPT : std::shared_ptr() {} inline Ptr(nullptr_t) CV_NOEXCEPT : std::shared_ptr(nullptr) {} template inline Ptr(Y* p, D d) : std::shared_ptr(p, d) {} template inline Ptr(nullptr_t, D d) : std::shared_ptr(nullptr, d) {} template inline Ptr(const Ptr& r, T* ptr) CV_NOEXCEPT : std::shared_ptr(r, ptr) {} inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} template inline Ptr(const Ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} template inline Ptr(Ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} #endif inline Ptr(const std::shared_ptr& o) CV_NOEXCEPT : std::shared_ptr(o) {} inline Ptr(std::shared_ptr&& o) CV_NOEXCEPT : std::shared_ptr(std::move(o)) {} // Overload with custom DefaultDeleter: Ptr(...) template inline Ptr(const std::true_type&, Y* ptr) : std::shared_ptr(ptr, DefaultDeleter()) {} // Overload without custom deleter: Ptr(...); template inline Ptr(const std::false_type&, Y* ptr) : std::shared_ptr(ptr) {} template inline Ptr(Y* ptr) : Ptr(has_custom_delete(), ptr) {} // Overload with custom DefaultDeleter: Ptr(...) template inline void reset(const std::true_type&, Y* ptr) { std::shared_ptr::reset(ptr, DefaultDeleter()); } // Overload without custom deleter: Ptr(...); template inline void reset(const std::false_type&, Y* ptr) { std::shared_ptr::reset(ptr); } template inline void reset(Y* ptr) { Ptr::reset(has_custom_delete(), ptr); } template void reset(Y* ptr, Deleter d) { std::shared_ptr::reset(ptr, d); } void reset() CV_NOEXCEPT { std::shared_ptr::reset(); } Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } template inline Ptr& operator=(const Ptr& o) { std::shared_ptr::operator =(o); return *this; } T* operator->() const CV_NOEXCEPT { return std::shared_ptr::get();} typename std::add_lvalue_reference::type operator*() const CV_NOEXCEPT { return *std::shared_ptr::get(); } // OpenCV 3.x methods (not a part of standart C++ library) inline void release() { std::shared_ptr::reset(); } inline operator T* () const { return std::shared_ptr::get(); } inline bool empty() const { return std::shared_ptr::get() == nullptr; } template inline Ptr staticCast() const CV_NOEXCEPT { return std::static_pointer_cast(*this); } template inline Ptr constCast() const CV_NOEXCEPT { return std::const_pointer_cast(*this); } template inline Ptr dynamicCast() const CV_NOEXCEPT { return std::dynamic_pointer_cast(*this); } }; template static inline Ptr<_Tp> makePtr(const A1&... a1) {
[GitHub] [incubator-mxnet] arcadiaphy commented on issue #14058: add backgroud class in box_nms
arcadiaphy commented on issue #14058: add backgroud class in box_nms URL: https://github.com/apache/incubator-mxnet/pull/14058#issuecomment-471387717 @zhreshold OK to go now. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn commented on issue #14385: OpenCV 4.0 is currently not support on win10
wkcn commented on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471386554 Could you please provide the context of error message, and the code of c:\opencv\build\include\opencv2\core\cvstd_wrapper.hpp45? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] hengdos edited a comment on issue #13870: [C++] Linking static library (error or bug?) to load python trained model
hengdos edited a comment on issue #13870: [C++] Linking static library (error or bug?) to load python trained model URL: https://github.com/apache/incubator-mxnet/issues/13870#issuecomment-471385323 > @hengdos I was trying to reproduce this issue on Mac and Ubuntu. But I was not able to invoke 'make' while building the 'transformer' with statically linked libraries. Can you please ensure that CMakelist.txt for static library is correct? > I will try it again as well. @leleamol hi, I have tested the following CMakelist.txt for static library. It works well on my macbook (macOS Mojave 10.14.3). ``` cmake_minimum_required(VERSION 3.9) set (CMAKE_CXX_STANDARD 11) add_executable(transformer main.cc ) target_link_libraries(transformer ${PROJECT_SOURCE_DIR}/lib/libmxnet.a ${PROJECT_SOURCE_DIR}/lib/libdmlc.a ${PROJECT_SOURCE_DIR}/lib/libnnvm.a ) target_include_directories(transformer PUBLIC ${PROJECT_SOURCE_DIR}/include ) ``` The output is like this: ```shell $ ./transformer net-symbol.json ... 954 bytes net-.params ... 164 bytes Assertion failed: (pred_hnd), function main, file /Users/heng/Projects/lrcplus/main.cc, line 101. [1]30126 abort ./transformer ``` Thanks for your reply. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] hengdos commented on issue #13870: [C++] Linking static library (error or bug?) to load python trained model
hengdos commented on issue #13870: [C++] Linking static library (error or bug?) to load python trained model URL: https://github.com/apache/incubator-mxnet/issues/13870#issuecomment-471385323 > @hengdos I was trying to reproduce this issue on Mac and Ubuntu. But I was not able to invoke 'make' while building the 'transformer' with statically linked libraries. Can you please ensure that CMakelist.txt for static library is correct? > I will try it again as well. @leleamol hi, I have tested the following CMakelist.txt for static library on macOS Mojave 10.14.3. ``` cmake_minimum_required(VERSION 3.9) set (CMAKE_CXX_STANDARD 11) add_executable(transformer main.cc ) target_link_libraries(transformer ${PROJECT_SOURCE_DIR}/lib/libmxnet.a ${PROJECT_SOURCE_DIR}/lib/libdmlc.a ${PROJECT_SOURCE_DIR}/lib/libnnvm.a ) target_include_directories(transformer PUBLIC ${PROJECT_SOURCE_DIR}/include ) ``` The output is like this: ```shell $ ./transformer net-symbol.json ... 954 bytes net-.params ... 164 bytes Assertion failed: (pred_hnd), function main, file /Users/heng/Projects/lrcplus/main.cc, line 101. [1]30126 abort ./transformer ``` Thanks for your reply. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] dbsxdbsx commented on issue #14313: compatibility with opencv4
dbsxdbsx commented on issue #14313: compatibility with opencv4 URL: https://github.com/apache/incubator-mxnet/pull/14313#issuecomment-471385029 @wkcn, thanks. I posted it #14385. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #14385: OpenCV 4.0 is currently not support on win10
mxnet-label-bot commented on issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385#issuecomment-471384938 Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it. Here are my recommended labels: Build This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] dbsxdbsx opened a new issue #14385: OpenCV 4.0 is currently not support on win10
dbsxdbsx opened a new issue #14385: OpenCV 4.0 is currently not support on win10 URL: https://github.com/apache/incubator-mxnet/issues/14385 I try to build mxnet on win10_64. With the latest mxnet source code, and vs2017(changed to v14) cuda+cudnn+mkl+mkldnn+opencv4(prebuilt), then it raises an error after compiling about 15 minutes. ``` incomplete type is not allowed mxnet c:\opencv\build\include\opencv2\core\cvstd_wrapper.hpp 45 ``` But everything is ok with opencv3.4.4(prebuilt). This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10
pengzhao-intel commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10 URL: https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-471383996 To reproduce the problem on the Windows This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wufengqian commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10
wufengqian commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10 URL: https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-471382917 Building the windows system inside? I'm sorry but I am curious about that. What is the purpose ? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10
pengzhao-intel commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10 URL: https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-471382333 Sorry, we don't check the forum frequently :( BTW, please wait a moment because we are building the windows system inside now. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wufengqian commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10
wufengqian commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10 URL: https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-471381106 Hi, I also asked the same question in [discuss.gluon.ai](url) ,but it seems that no one know how to solve the problem in detail. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14375: 新版官网没有验证安装过程
pengzhao-intel commented on issue #14375: 新版官网没有验证安装过程 URL: https://github.com/apache/incubator-mxnet/issues/14375#issuecomment-471380094 @juliusshufan This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14375: 新版官网没有验证安装过程
pengzhao-intel commented on issue #14375: 新版官网没有验证安装过程 URL: https://github.com/apache/incubator-mxnet/issues/14375#issuecomment-471380005 @wufengqian thanks for the great suggestions. @TaoLv is working on the MKL documentation https://github.com/apache/incubator-mxnet/pull/14202 @mxnet-label-bot add [doc, questions] This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] wkcn commented on issue #14313: compatibility with opencv4
wkcn commented on issue #14313: compatibility with opencv4 URL: https://github.com/apache/incubator-mxnet/pull/14313#issuecomment-471378485 @dbsxdbsx Could you please open an issue and provide more error message? I only fix the compatibility on Linux in this PR. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] dbsxdbsx commented on issue #14313: compatibility with opencv4
dbsxdbsx commented on issue #14313: compatibility with opencv4 URL: https://github.com/apache/incubator-mxnet/pull/14313#issuecomment-471375630 As far as I know, when I tried to build mxnet on win10 with cuda 10.1. It failed with compling error. ``` incomplete type is not allowed mxnet c:\opencv\build\include\opencv2\core\cvstd_wrapper.hpp 45 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
This is an automated email from the ASF dual-hosted git repository. zhasheng pushed a commit to branch asf-site in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git The following commit(s) were added to refs/heads/asf-site by this push: new d650bd1 Bump the publish timestamp. d650bd1 is described below commit d650bd1181a77025932ddff20620326846a63fa7 Author: mxnet-ci AuthorDate: Mon Mar 11 01:18:19 2019 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..3fb55ea --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Mon Mar 11 01:18:19 UTC 2019
[GitHub] [incubator-mxnet] zboldyga edited a comment on issue #14360: supporting matrix inversion and determinant
zboldyga edited a comment on issue #14360: supporting matrix inversion and determinant URL: https://github.com/apache/incubator-mxnet/issues/14360#issuecomment-471338317 @ketranm Looks like inversion via Cholesky factorization is supported, and there's also an API handle for getting the inversion using that factorization: https://mxnet.apache.org/api/python/ndarray/linalg.html#linear-algebra potrf - get the Cholesky factorization (triangular matrix) potri - calculate inversion (edit: using the Cholesky factorization from potrf) sumlogdiag - *may* be useful for calculating logdeterminant (my linear algebra is a little rusty) There's not a shortcut for getting the determinant or log determinant, but these are simple ops using the Cholesky factorization. It seems to me that all of this should be clarified in the documentation, at the minimum, and we should probably add API calls for det and logdet. I've also made a request to have a single 'inverse' operation as with Torch. I opened a JIRA ticket and will start implementing these as soon as someone more internal to the MXNet project signs-off! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] gigasquid commented on a change in pull request #14305: [Clojure] Helper function for n-dim vector to ndarray
gigasquid commented on a change in pull request #14305: [Clojure] Helper function for n-dim vector to ndarray URL: https://github.com/apache/incubator-mxnet/pull/14305#discussion_r264014076 ## File path: contrib/clojure-package/src/org/apache/clojure_mxnet/util.clj ## @@ -218,15 +218,25 @@ (throw (ex-info error-msg (s/explain-data spec value) -(s/def ::non-empty-seq sequential?) +(s/def ::non-empty-seq (s/and sequential? not-empty)) (defn to-array-nd "Converts any N-D sequential structure to an array with the same dimensions." - [s] - (validate! ::non-empty-seq s "Invalid N-D sequence") - (if (sequential? (first s)) -(to-array (mapv to-array-nd s)) -(to-array s))) + [nd-seq] + (validate! ::non-empty-seq nd-seq "Invalid N-D sequence") + (if (sequential? (first nd-seq)) +(to-array (mapv to-array-nd nd-seq)) +(to-array nd-seq))) + +(defn nd-seq-shape Review comment: Nice! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet] branch master updated: [clojure-package][wip] add `->nd-vec` function in `ndarray.clj` (#14308)
This is an automated email from the ASF dual-hosted git repository. cmeier pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 8be97d7 [clojure-package][wip] add `->nd-vec` function in `ndarray.clj` (#14308) 8be97d7 is described below commit 8be97d7a79f9ea9815e41956e5f15ddcf25026b6 Author: Arthur Caillau AuthorDate: Mon Mar 11 00:46:50 2019 +0100 [clojure-package][wip] add `->nd-vec` function in `ndarray.clj` (#14308) * [clojure-package][wip] add `->nd-vec` function in `ndarray.clj` * WIP * Unit tests need to be added * [clojure-package][ndarray] add unit tests for `->nd-vec` util fn --- .../src/org/apache/clojure_mxnet/ndarray.clj | 58 +++--- .../test/org/apache/clojure_mxnet/ndarray_test.clj | 12 + 2 files changed, 64 insertions(+), 6 deletions(-) diff --git a/contrib/clojure-package/src/org/apache/clojure_mxnet/ndarray.clj b/contrib/clojure-package/src/org/apache/clojure_mxnet/ndarray.clj index 651bdcb..151e18b 100644 --- a/contrib/clojure-package/src/org/apache/clojure_mxnet/ndarray.clj +++ b/contrib/clojure-package/src/org/apache/clojure_mxnet/ndarray.clj @@ -16,15 +16,18 @@ ;; (ns org.apache.clojure-mxnet.ndarray + "NDArray API for Clojure package." (:refer-clojure :exclude [* - + > >= < <= / cast concat flatten identity load max min repeat reverse set sort take to-array empty shuffle ref]) - (:require [org.apache.clojure-mxnet.base :as base] -[org.apache.clojure-mxnet.context :as mx-context] -[org.apache.clojure-mxnet.shape :as mx-shape] -[org.apache.clojure-mxnet.util :as util] -[clojure.reflect :as r] -[t6.from-scala.core :refer [$] :as $]) + (:require +[clojure.spec.alpha :as s] + +[org.apache.clojure-mxnet.base :as base] +[org.apache.clojure-mxnet.context :as mx-context] +[org.apache.clojure-mxnet.shape :as mx-shape] +[org.apache.clojure-mxnet.util :as util] +[t6.from-scala.core :refer [$] :as $]) (:import (org.apache.mxnet NDArray))) ;; loads the generated functions into the namespace @@ -167,3 +170,46 @@ (defn shape-vec [ndarray] (mx-shape/->vec (shape ndarray))) + +(s/def ::ndarray #(instance? NDArray %)) +(s/def ::vector vector?) +(s/def ::sequential sequential?) +(s/def ::shape-vec-match-vec + (fn [[v vec-shape]] (= (count v) (reduce clojure.core/* 1 vec-shape + +(s/fdef vec->nd-vec +:args (s/cat :v ::sequential :shape-vec ::sequential) +:ret ::vector) + +(defn- vec->nd-vec + "Convert a vector `v` into a n-dimensional vector given the `shape-vec` + Ex: +(vec->nd-vec [1 2 3] [1 1 3]) ;[[[1 2 3]]] +(vec->nd-vec [1 2 3 4 5 6] [2 3 1]) ;[[[1] [2] [3]] [[4] [5] [6]]] +(vec->nd-vec [1 2 3 4 5 6] [1 2 3]) ;[[[1 2 3]] [4 5 6]]] +(vec->nd-vec [1 2 3 4 5 6] [3 1 2]) ;[[[1 2]] [[3 4]] [[5 6]]] +(vec->nd-vec [1 2 3 4 5 6] [3 2]) ;[[1 2] [3 4] [5 6]]" + [v [s1 & ss :as shape-vec]] + (util/validate! ::sequential v "Invalid input vector `v`") + (util/validate! ::sequential shape-vec "Invalid input vector `shape-vec`") + (util/validate! ::shape-vec-match-vec + [v shape-vec] + "Mismatch between vector `v` and vector `shape-vec`") + (if-not (seq ss) +(vec v) +(->> v + (partition (clojure.core// (count v) s1)) + vec + (mapv #(vec->nd-vec % ss) + +(s/fdef ->nd-vec :args (s/cat :ndarray ::ndarray) :ret ::vector) + +(defn ->nd-vec + "Convert an ndarray `ndarray` into a n-dimensional Clojure vector. + Ex: +(->nd-vec (array [1] [1 1 1])) ;[[[1.0]]] +(->nd-vec (array [1 2 3] [3 1 1])) ;[[[1.0]] [[2.0]] [[3.0]]] +(->nd-vec (array [1 2 3 4 5 6]) [3 1 2]) ;[[[1.0 2.0]] [[3.0 4.0]] [[5.0 6.0]]]" + [ndarray] + (util/validate! ::ndarray ndarray "Invalid input array") + (vec->nd-vec (->vec ndarray) (shape-vec ndarray))) diff --git a/contrib/clojure-package/test/org/apache/clojure_mxnet/ndarray_test.clj b/contrib/clojure-package/test/org/apache/clojure_mxnet/ndarray_test.clj index 9ffd3ab..a9ae296 100644 --- a/contrib/clojure-package/test/org/apache/clojure_mxnet/ndarray_test.clj +++ b/contrib/clojure-package/test/org/apache/clojure_mxnet/ndarray_test.clj @@ -473,3 +473,15 @@ (is (= [2 2] (ndarray/->int-vec nda))) (is (= [2.0 2.0] (ndarray/->double-vec nda))) (is (= [(byte 2) (byte 2)] (ndarray/->byte-vec nda) + +(deftest test->nd-vec + (is (= [[[1.0]]] + (ndarray/->nd-vec (ndarray/array [1] [1 1 1] + (is (= [[[1.0]] [[2.0]] [[3.0]]] + (ndarray/->nd-vec (ndarray/array [1 2 3] [3 1 1] + (is (= [[[1.0 2.0]] [[3.0 4.0]] [[5.0 6.0]]] + (ndarray/->nd-vec (ndarray/array [1 2 3 4 5 6] [3 1 2] + (is (= [[[1.0] [2.0]] [[3.0] [4.0]] [[5.0] [6.0]]] +
[GitHub] [incubator-mxnet] gigasquid merged pull request #14308: [clojure-package] add `->nd-vec` function in `ndarray.clj`
gigasquid merged pull request #14308: [clojure-package] add `->nd-vec` function in `ndarray.clj` URL: https://github.com/apache/incubator-mxnet/pull/14308 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] gigasquid commented on issue #14308: [clojure-package] add `->nd-vec` function in `ndarray.clj`
gigasquid commented on issue #14308: [clojure-package] add `->nd-vec` function in `ndarray.clj` URL: https://github.com/apache/incubator-mxnet/pull/14308#issuecomment-471366487 Thanks for your contribution @Chouffe This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] perdasilva commented on issue #14336: CI Changes for Codified Windows AMIs
perdasilva commented on issue #14336: CI Changes for Codified Windows AMIs URL: https://github.com/apache/incubator-mxnet/pull/14336#issuecomment-471356456 @marcoabreu please review and merge if it's ok. Thank you! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
This is an automated email from the ASF dual-hosted git repository. zhasheng pushed a commit to branch asf-site in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git The following commit(s) were added to refs/heads/asf-site by this push: new a645d3b Bump the publish timestamp. a645d3b is described below commit a645d3b810cb5f80dbd6b402694fe70fde3831ef Author: mxnet-ci AuthorDate: Sun Mar 10 20:50:41 2019 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..3bc06f9 --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Sun Mar 10 20:50:41 UTC 2019
[GitHub] [incubator-mxnet] zboldyga edited a comment on issue #14360: supporting matrix inversion and determinant
zboldyga edited a comment on issue #14360: supporting matrix inversion and determinant URL: https://github.com/apache/incubator-mxnet/issues/14360#issuecomment-471338317 @ketranm Looks like inversion via Cholesky factorization is supported, and there's also an API handle for getting the inversion using that factorization: https://mxnet.apache.org/api/python/ndarray/linalg.html#linear-algebra potrf - get the Cholesky factorization (triangular matrix) potri - calculate inversion sumlogdiag - *may* be useful for calculating logdeterminant (my linear algebra is a little rusty) There's not a shortcut for getting the determinant or log determinant, but these are simple ops using the Cholesky factorization. It seems to me that all of this should be clarified in the documentation, at the minimum, and we should probably add API calls for det and logdet. I've also made a request to have a single 'inverse' operation as with Torch. I opened a JIRA ticket and will start implementing these as soon as someone more internal to the MXNet project signs-off! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] zboldyga commented on issue #14360: supporting matrix inversion and determinant
zboldyga commented on issue #14360: supporting matrix inversion and determinant URL: https://github.com/apache/incubator-mxnet/issues/14360#issuecomment-471338317 @ketranm Looks like inversion via Cholesky factorization is supported, and there's also an API handle for getting the inversion using that factorization: https://mxnet.apache.org/api/python/ndarray/linalg.html#linear-algebra potrf - get the Cholesky factorization (triangular matrix) potri - calculate inversion sumlogdiag - *may* be useful for calculating logdeterminant (my linear algebra is a little rusty) There's not a shortcut for getting the determinant or log determinant, but these are simple ops using the Cholesky factorization. It seems to me that all of this should be clarified in the documentation, at the minimum, and we should probably add API calls for det and logdet. I opened a JIRA ticket and will start implementing these as soon as someone more internal to the MXNet project signs-off! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
This is an automated email from the ASF dual-hosted git repository. zhasheng pushed a commit to branch asf-site in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git The following commit(s) were added to refs/heads/asf-site by this push: new d21eb5b Bump the publish timestamp. d21eb5b is described below commit d21eb5bf919345697544e399dc31e55c73d5ad48 Author: mxnet-ci AuthorDate: Sun Mar 10 19:18:15 2019 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..31a66cb --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Sun Mar 10 19:18:15 UTC 2019
[GitHub] [incubator-mxnet] Chouffe commented on issue #14308: [clojure-package][wip] add `->nd-vec` function in `ndarray.clj`
Chouffe commented on issue #14308: [clojure-package][wip] add `->nd-vec` function in `ndarray.clj` URL: https://github.com/apache/incubator-mxnet/pull/14308#issuecomment-471333656 Unit tests added and all checks have passed! This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx
turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx URL: https://github.com/apache/incubator-mxnet/issues/13949#issuecomment-471307657 > @wangliye00 @Con-Mi However, I do see the error that you facing with MXNet v1.3.1. For fixing this, could you try pulling in the commit #13413 and checking if you are able to proceed further? I am facing similar issue when loading pytorch-densenet onnx model into mxnet. The error message reads: ``` /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/import_onnx.py in _convert_operator(self, node_name, op_name, attrs, inputs) 59 """ 60 if op_name in convert_map: ---> 61 op_name, new_attrs, inputs = convert_map[op_name](attrs, inputs, self) 62 else: 63 raise NotImplementedError("Operator {} not implemented.".format(op_name)) /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/_op_translations.py in reshape(attrs, inputs, proto_obj) 432 if len(inputs) == 1: 433 return 'reshape', attrs, inputs[0] --> 434 reshape_shape = list(proto_obj._params[inputs[1].name].asnumpy()) 435 reshape_shape = [int(i) for i in reshape_shape] 436 new_attrs = {'shape': reshape_shape} KeyError: 'concat51' ``` I tried mxnet 1.3.1 (after patched `import_onnx.py` following your suggestion) and 1.4.0 with no luck, both raised similar exception. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx
turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx URL: https://github.com/apache/incubator-mxnet/issues/13949#issuecomment-471307657 > @wangliye00 @Con-Mi However, I do see the error that you facing with MXNet v1.3.1. For fixing this, could you try pulling in the commit #13413 and checking if you are able to proceed further? I am facing similar issue when loading pytorch-densenet onnx model into mxnet. The error message reads: ``` /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/import_onnx.py in _convert_operator(self, node_name, op_name, attrs, inputs) 59 """ 60 if op_name in convert_map: ---> 61 op_name, new_attrs, inputs = convert_map[op_name](attrs, inputs, self) 62 else: 63 raise NotImplementedError("Operator {} not implemented.".format(op_name)) /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/_op_translations.py in reshape(attrs, inputs, proto_obj) 432 if len(inputs) == 1: 433 return 'reshape', attrs, inputs[0] --> 434 reshape_shape = list(proto_obj._params[inputs[1].name].asnumpy()) 435 reshape_shape = [int(i) for i in reshape_shape] 436 new_attrs = {'shape': reshape_shape} KeyError: 'concat51' ``` I tried mxnet 1.3.1 (after patched `import_onnx.py` following your suggestion) and 1.4.0 with no luck, both raised similar Exception. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] turtleizzy commented on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx
turtleizzy commented on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx URL: https://github.com/apache/incubator-mxnet/issues/13949#issuecomment-471307657 > @wangliye00 @Con-Mi However, I do see the error that you facing with MXNet v1.3.1. For fixing this, could you try pulling in the commit #13413 and checking if you are able to proceed further? I am facing similar issue when loading pytorch-densenet onnx model into mxnet. The error message reads: `/usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/import_onnx.py in _convert_operator(self, node_name, op_name, attrs, inputs) 59 """ 60 if op_name in convert_map: ---> 61 op_name, new_attrs, inputs = convert_map[op_name](attrs, inputs, self) 62 else: 63 raise NotImplementedError("Operator {} not implemented.".format(op_name)) /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/_op_translations.py in reshape(attrs, inputs, proto_obj) 432 if len(inputs) == 1: 433 return 'reshape', attrs, inputs[0] --> 434 reshape_shape = list(proto_obj._params[inputs[1].name].asnumpy()) 435 reshape_shape = [int(i) for i in reshape_shape] 436 new_attrs = {'shape': reshape_shape} KeyError: 'concat51'` I tried mxnet 1.3.1 (after patched `import_onnx.py` following your suggestion) and 1.4.0 with no luck, both raised similar Exception. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
This is an automated email from the ASF dual-hosted git repository. zhasheng pushed a commit to branch asf-site in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git The following commit(s) were added to refs/heads/asf-site by this push: new eae7471 Bump the publish timestamp. eae7471 is described below commit eae74713a499093dffe39eaad39cf557d2efb583 Author: mxnet-ci AuthorDate: Sun Mar 10 13:18:37 2019 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..0110165 --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Sun Mar 10 13:18:37 UTC 2019
[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14335: [MKLDNN] Question on installation and use of MKLDNN
pengzhao-intel commented on issue #14335: [MKLDNN] Question on installation and use of MKLDNN URL: https://github.com/apache/incubator-mxnet/issues/14335#issuecomment-471278086 + @juliusshufan to track the windows related building issues @mxnet-label-bot Add [Windows] This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] stereomatchingkiss commented on issue #14380: fix type mismatch bugs
stereomatchingkiss commented on issue #14380: fix type mismatch bugs URL: https://github.com/apache/incubator-mxnet/pull/14380#issuecomment-471269805 > And I think you could merge the PR into the master branch, thank you! There are two problems stopping me to do that 1. Since mxnet 1.4.0, mxnet got even more bugs when you try to compile it on windows, I can't specify those bugs are caused by this pull request or not because of the new bugs exist in master branch(please check #14203) 2. Could you tell me why jenkins complain? Thanks This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] stereomatchingkiss commented on issue #14370: Add travis build for different platforms
stereomatchingkiss commented on issue #14370: Add travis build for different platforms URL: https://github.com/apache/incubator-mxnet/issues/14370#issuecomment-471269478 > Or are the Jenkins CI testing out various platform sufficient for now ? I hope it is enough too, but it is not, there are many bugs when you try to build mxnet on windows, especially when you try to build with intel mkl, please check issues #14343, #14364 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] peterpaniff opened a new issue #14384: try to train ssdlite mobilenetv2, encounter the error.
peterpaniff opened a new issue #14384: try to train ssdlite mobilenetv2, encounter the error. URL: https://github.com/apache/incubator-mxnet/issues/14384 i follow the instrctions. The code below was added to example/ssd/symbol/symbol_factory.py elif network == 'mobilenet_v2': image_shape = '3,224,224' network = 'mobilenet_v2' from_layers = ['relu6_1_expand', 'relu6_4', '', '', '', ''] num_filters = [-1, -1, 512, 256, 256, 128] strides = [-1, -1, 2, 2, 2, 2] pads = [-1, -1, 1, 1, 1, 1] sizes = [[.1, .141], [.2, .272], [.37, .447], [.54, .619], [.71, .79], [.88, .961]] ratios = [[1, 2, .5], [1, 2, .5, 3, 1. / 3], [1, 2, .5, 3, 1. / 3], [1, 2, .5, 3, 1. / 3], \ [1, 2, .5], [1, 2, .5]] normalizations = -1 steps = [] return locals() when i train the model, but i enconter the error below: TypeError: get_symbol() got an unexpected keyword argument 'image_shape' This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services