[GitHub] [incubator-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309496476 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: ``` The build log is here: ubuntu@ip-172-31-18-113:~/incubator-singa/build$ rm -rf * ubuntu@ip-172-31-18-113:~/incubator-singa/build$ cmake -D CMAKE_PREFIX_PATH="/usr/local/cuda/lib64;/usr/local/cuda/" -DENABLE_TEST=OFF -DUSE_CUDA=ON -DUSE_PYTHON3=ON -DUSE_MKLDNN=ON -DUSE_MODULES=OFF -DUSE_DIST=ON .. -- The C compiler identification is GNU 5.4.0 -- The CXX compiler identification is GNU 5.4.0 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- Looking for pthread.h -- Looking for pthread.h - found -- Looking for pthread_create -- Looking for pthread_create - not found -- Looking for pthread_create in pthreads -- Looking for pthread_create in pthreads - not found -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE -- Found Protobuf: /usr/local/lib/libprotobuf.so;-lpthread (found suitable version "3.0.0", minimum required is "3.0") -- Found CBLAS: /usr/local/include -- Found GLOG: /usr/include -- Found cuda_v10.0 -- Found CUDNN: /usr/local/cuda/include -- Found Cudnn_7401 at /usr/local/cuda/include /usr/local/cuda/lib64/libcudnn.so -- Found PythonInterp: /usr/bin/python3 (found suitable version "3.5.2", minimum required is "3") -- Found PythonLibs: /usr/lib/x86_64-linux-gnu/libpython3.5m.so (found suitable version "3.5.2", minimum required is "3") -- Found SWIG: /usr/local/bin/swig (found suitable version "3.0.12", minimum required is "3.0.10") -- Found MKLDNN at /usr/local/include -- Found MPI at /home/ubuntu/mpich-3.3/build/include -- Found MPI lib at /home/ubuntu/mpich-3.3/build/lib/libmpi.so -- Found all lib at /usr/local/lib/libprotobuf.so;/usr/local/lib/libopenblas.so;/usr/lib/x86_64-linux-gnu/libglog.so;/usr/local/cuda/lib64/libcudnn.so;/usr/local/cuda/lib64/libcudart.so;/usr/local/cuda/lib64/libcurand.so;/usr/local/cuda/lib64/libcublas.so;/home/ubuntu/incubator-singa/build/lib/libcnmem.a;/usr/local/lib/libmkldnn.so;/home/ubuntu/mpich-3.3/build/lib/libmpi.so;/home/ubuntu/mpich-3.3/build/lib/libmpicxx.so -- Found NCCL at /usr/local/cuda/include -- Found NCCL lib at /usr/local/cuda/lib/libnccl.so -- Configuring done -- Generating done -- Build files have been written to: /home/ubuntu/incubator-singa/build ubuntu@ip-172-31-18-113:~/incubator-singa/build$ make -j2 Scanning dependencies of target cnmem Scanning dependencies of target copy_protobuf [ 1%] Creating directories for 'cnmem' [ 2%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/model.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: model.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 3%] Performing download step (git clone) for 'cnmem' Cloning into 'cnmem'... [ 4%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/caffe.proto [ 5%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/core.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: core.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 6%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/io.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: io.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 7%] Copying Protobuf headers [ 7%] Built target copy_protobuf [ 8%] Building NVCC (Device) object src/CMakeFiles/cuda_compile_1.dir/core/tensor/cuda_compile_1_generated_math_kernel.cu.o Scanning dependencies of target singa_objects [ 9%] Building CXX object src/CMakeFiles/singa_objects.dir/caffe.pb.cc.o Already on 'master' Your branch is up-to-date with 'origin/master'. [ 10%] No
[GitHub] [incubator-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309496476 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: The build log is here: ubuntu@ip-172-31-18-113:~/incubator-singa/build$ rm -rf * ubuntu@ip-172-31-18-113:~/incubator-singa/build$ cmake -D CMAKE_PREFIX_PATH="/usr/local/cuda/lib64;/usr/local/cuda/" -DENABLE_TEST=OFF -DUSE_CUDA=ON -DUSE_PYTHON3=ON -DUSE_MKLDNN=ON -DUSE_MODULES=OFF -DUSE_DIST=ON .. -- The C compiler identification is GNU 5.4.0 -- The CXX compiler identification is GNU 5.4.0 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- Looking for pthread.h -- Looking for pthread.h - found -- Looking for pthread_create -- Looking for pthread_create - not found -- Looking for pthread_create in pthreads -- Looking for pthread_create in pthreads - not found -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE -- Found Protobuf: /usr/local/lib/libprotobuf.so;-lpthread (found suitable version "3.0.0", minimum required is "3.0") -- Found CBLAS: /usr/local/include -- Found GLOG: /usr/include -- Found cuda_v10.0 -- Found CUDNN: /usr/local/cuda/include -- Found Cudnn_7401 at /usr/local/cuda/include /usr/local/cuda/lib64/libcudnn.so -- Found PythonInterp: /usr/bin/python3 (found suitable version "3.5.2", minimum required is "3") -- Found PythonLibs: /usr/lib/x86_64-linux-gnu/libpython3.5m.so (found suitable version "3.5.2", minimum required is "3") -- Found SWIG: /usr/local/bin/swig (found suitable version "3.0.12", minimum required is "3.0.10") -- Found MKLDNN at /usr/local/include -- Found MPI at /home/ubuntu/mpich-3.3/build/include -- Found MPI lib at /home/ubuntu/mpich-3.3/build/lib/libmpi.so -- Found all lib at /usr/local/lib/libprotobuf.so;/usr/local/lib/libopenblas.so;/usr/lib/x86_64-linux-gnu/libglog.so;/usr/local/cuda/lib64/libcudnn.so;/usr/local/cuda/lib64/libcudart.so;/usr/local/cuda/lib64/libcurand.so;/usr/local/cuda/lib64/libcublas.so;/home/ubuntu/incubator-singa/build/lib/libcnmem.a;/usr/local/lib/libmkldnn.so;/home/ubuntu/mpich-3.3/build/lib/libmpi.so;/home/ubuntu/mpich-3.3/build/lib/libmpicxx.so -- Found NCCL at /usr/local/cuda/include -- Found NCCL lib at /usr/local/cuda/lib/libnccl.so -- Configuring done -- Generating done -- Build files have been written to: /home/ubuntu/incubator-singa/build ubuntu@ip-172-31-18-113:~/incubator-singa/build$ make -j2 Scanning dependencies of target cnmem Scanning dependencies of target copy_protobuf [ 1%] Creating directories for 'cnmem' [ 2%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/model.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: model.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 3%] Performing download step (git clone) for 'cnmem' Cloning into 'cnmem'... [ 4%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/caffe.proto [ 5%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/core.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: core.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 6%] Running C++ protocol buffer compiler on /home/ubuntu/incubator-singa/src/proto/io.proto [libprotobuf WARNING google/protobuf/compiler/parser.cc:547] No syntax specified for the proto file: io.proto. Please use 'syntax = "proto2";' or 'syntax = "proto3";' to specify a syntax version. (Defaulted to proto2 syntax.) [ 7%] Copying Protobuf headers [ 7%] Built target copy_protobuf [ 8%] Building NVCC (Device) object src/CMakeFiles/cuda_compile_1.dir/core/tensor/cuda_compile_1_generated_math_kernel.cu.o Scanning dependencies of target singa_objects [ 9%] Building CXX object src/CMakeFiles/singa_objects.dir/caffe.pb.cc.o Already on 'master' Your branch is up-to-date with 'origin/master'. [ 10%] No patch
[GitHub] [incubator-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309329911 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: I updated also some files to include USE_DIST, see the following grep result on USE_DIST: ubuntu@ip-172-31-18-113:~/incubator-singa$ git grep USE_DIST CMakeLists.txt:OPTION(USE_DIST "Use nccl distributed module" OFF) cmake/Dependencies.cmake:IF(USE_DIST) cmake/Templates/singa_config.h.in:#cmakedefine USE_DIST include/singa/dist/communicator.h:#ifdef USE_DIST include/singa/dist/communicator.h:#endif // USE_DIST src/CMakeLists.txt:IF (USE_DIST) src/CMakeLists.txt:ENDIF (USE_DIST) src/api/config.i:#define USE_DIST 0 src/api/config.i.in:#cmakedefine01 USE_DIST src/api/dist_communicator.i:#if USE_DIST src/api/dist_communicator.i:#endif // USE_DIST src/dist/communicator.cc:#ifdef USE_DIST src/dist/communicator.cc:#endif // USE_DIST Note that the default is OFF if we do not set -DUSE_DIST=ON The test was on version 1.2 although I set the displayed value in CMakeLists to be version 2.0. I will still need to test the dist module on singa version 2.0 and add partitioning of dataset according to MPI rank, etc. 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-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309329911 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: I updated also some files to include USE_DIST, see the following grep result on USE_DIST: ubuntu@ip-172-31-18-113:~/incubator-singa$ git grep USE_DIST CMakeLists.txt:OPTION(USE_DIST "Use nccl distributed module" OFF) cmake/Dependencies.cmake:IF(USE_DIST) cmake/Templates/singa_config.h.in:#cmakedefine USE_DIST include/singa/dist/communicator.h:#ifdef USE_DIST include/singa/dist/communicator.h:#endif // USE_DIST src/CMakeLists.txt:IF (USE_DIST) src/CMakeLists.txt:ENDIF (USE_DIST) src/api/config.i:#define USE_DIST 1 src/api/config.i.in:#cmakedefine01 USE_DIST src/api/dist_communicator.i:#if USE_DIST src/api/dist_communicator.i:#endif // USE_DIST src/dist/communicator.cc:#ifdef USE_DIST src/dist/communicator.cc:#endif // USE_DIST Note that the default is OFF if we do not set -DUSE_DIST=ON The test was on version 1.2 although I set the displayed value in CMakeLists to be version 2.0. I will still need to test the dist module on singa version 2.0 and add partitioning of dataset according to MPI rank, etc. 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-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309329911 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: I updated also some files to include USE_DIST, see the following grep result on USE_DIST: ubuntu@ip-172-31-18-113:~/incubator-singa$ git grep USE_DIST CMakeLists.txt:OPTION(USE_DIST "Use nccl distributed module" OFF) cmake/Dependencies.cmake:IF(USE_DIST) cmake/Templates/singa_config.h.in:#cmakedefine USE_DIST include/singa/dist/communicator.h:#ifdef USE_DIST include/singa/dist/communicator.h:#endif // USE_DIST src/CMakeLists.txt:IF (USE_DIST) src/CMakeLists.txt:ENDIF (USE_DIST) src/api/config.i:#define USE_DIST 1 src/api/config.i.in:#cmakedefine01 USE_DIST src/api/dist_communicator.i:#if USE_DIST src/api/dist_communicator.i:#endif // USE_DIST src/dist/communicator.cc:#ifdef USE_DIST src/dist/communicator.cc:#endif // USE_DIST Note that the default is OFF if we do not set -DUSE_DIST=ON The test was on version 1.2 although I set the displayed value in CMakeLists to be version 2.0. I will still need to test version 2.0 and add partitioning of dataset according to MPI rank, etc. 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-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309144703 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: Changed the files cmake/Dependencies.cmake and src/CMakeLists.txt Can use cmake -DUSE_DIST=ON to turn on the distributed module 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-singa] ShichengChen opened a new pull request #496: Mean
ShichengChen opened a new pull request #496: Mean URL: https://github.com/apache/incubator-singa/pull/496 Implement ONNX Operators following https://github.com/onnx/onnx/blob/master/docs/Operators.md 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-singa] chrishkchris commented on a change in pull request #468: Distributted module
chrishkchris commented on a change in pull request #468: Distributted module URL: https://github.com/apache/incubator-singa/pull/468#discussion_r309144703 ## File path: src/CMakeLists.txt ## @@ -36,6 +36,9 @@ AUX_SOURCE_DIRECTORY(core/scheduler core_source) AUX_SOURCE_DIRECTORY(core/tensor core_source) LIST(APPEND singa_sources ${core_source}) Review comment: Changed the files cmake/Dependencies.cmake and src/CMakeLists.txt Can use cmake -DUSE_DIST=ON to turn on the distributed module However, there are some bugs (mainly segmentation fault) if I add the #ifdef USE_DIST in the files communicator.h and communicator.cc I will update other files as well (e.g. #cmakedefine and #if USE_DIST etc. in many files) when I successfully remove the bug. 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-singa] chrishkchris commented on a change in pull request #493: SINGA-473 Autograd Trigonometry: Backward Test
chrishkchris commented on a change in pull request #493: SINGA-473 Autograd Trigonometry: Backward Test URL: https://github.com/apache/incubator-singa/pull/493#discussion_r309137122 ## File path: test/python/test_operation.py ## @@ -65,6 +65,17 @@ def prepare_inputs_targets_for_rnn_test(): targets = [t0, t1, t2] return inputs, targets, h0 +def numpy_unary_ops_backward(func, x, dy, h=0.0005): Review comment: Changed the code by computing the gradient explicitly, the accuracy check is up to 5 decimals. 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-singa] pinpom commented on a change in pull request #494: SINGA-475 add SoftPlus operator
pinpom commented on a change in pull request #494: SINGA-475 add SoftPlus operator URL: https://github.com/apache/incubator-singa/pull/494#discussion_r309088705 ## File path: test/python/test_operation.py ## @@ -610,6 +610,17 @@ def test_Atanh_gpu(self): np.testing.assert_array_almost_equal(tensor.to_numpy(result), XT, decimal=5) self.check_shape(dx.shape(), (3, 2)) +def test_SoftPlus(self): +X=np.array([1.0,2.0,3.0,4.0,5.0,6.0]).reshape(3,2).astype(np.float32) +XT=np.log(np.exp(X) + 1) +x=tensor.from_numpy(X) +x.to_device(gpu_dev) + +result=autograd.softplus(x) +dx=result.creator.backward(x.data) Review comment: yes. just added gradient test 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-singa] pinpom commented on a change in pull request #494: SINGA-475 add SoftPlus operator
pinpom commented on a change in pull request #494: SINGA-475 add SoftPlus operator URL: https://github.com/apache/incubator-singa/pull/494#discussion_r309088705 ## File path: test/python/test_operation.py ## @@ -610,6 +610,17 @@ def test_Atanh_gpu(self): np.testing.assert_array_almost_equal(tensor.to_numpy(result), XT, decimal=5) self.check_shape(dx.shape(), (3, 2)) +def test_SoftPlus(self): +X=np.array([1.0,2.0,3.0,4.0,5.0,6.0]).reshape(3,2).astype(np.float32) +XT=np.log(np.exp(X) + 1) +x=tensor.from_numpy(X) +x.to_device(gpu_dev) + +result=autograd.softplus(x) +dx=result.creator.backward(x.data) Review comment: I added gradient test 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
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Description: For the demo purpose, we need to implement these three models and their components as following: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and there're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas| [https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. was: For the demo purpose, we need to implement these three models and their components as following: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and there're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, bidaf.png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models and their > components as following: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > Add > BatchNormalization > Conv > LeakyRelu > MaxPool > Mul > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Acos > Add > BatchNormalization > Conv > Cos > Dropout > Flatten > Gemm > Identity > InstanceNormalization > LpNormalization > Mul > PRelu > Reshape > Sub > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > Abs > Add > Add > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Conv > Dropout > Gather > Hardmax > Log > LSTM > MatMul > ReduceMax > ReduceSum > Relu > Shape > Sigmoid > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > > In summary, we already implemented 13 ops, and there're still 27 ops needed > to be implemented: > h2. Already implemented: > -Acos- > -BatchNormalization- > -Cos- > -Conv- > -LeakyRelu- > -LSTM- > -Abs- > -MaxPool- > -Flatten- > -Add- > -MatMul- > -Relu- > -Sigmoid- > h2. To be implemented: > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Dropout > Gather > Gemm > Hardmax > Identity > InstanceNormalization > Log > LpNormalization > Mul > PRelu > ReduceMax > ReduceSum > Reshape > Shape > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > Please refer to the [ONNX Operator Schemas| > [https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more > detailed information. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Description: For the demo purpose, we need to implement these three models and their components as following: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and there're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas| https://github.com/onnx/onnx/blob/master/docs/Operators.md] for more detailed information. was: For the demo purpose, we need to implement these three models and their components as following: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and there're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas| [https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, bidaf.png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models and their > components as following: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > Add > BatchNormalization > Conv > LeakyRelu > MaxPool > Mul > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Acos > Add > BatchNormalization > Conv > Cos > Dropout > Flatten > Gemm > Identity > InstanceNormalization > LpNormalization > Mul > PRelu > Reshape > Sub > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > Abs > Add > Add > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Conv > Dropout > Gather > Hardmax > Log > LSTM > MatMul > ReduceMax > ReduceSum > Relu > Shape > Sigmoid > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > > In summary, we already implemented 13 ops, and there're still 27 ops needed > to be implemented: > h2. Already implemented: > -Acos- > -BatchNormalization- > -Cos- > -Conv- > -LeakyRelu- > -LSTM- > -Abs- > -MaxPool- > -Flatten- > -Add- > -MatMul- > -Relu- > -Sigmoid- > h2. To be implemented: > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Dropout > Gather > Gemm > Hardmax > Identity > InstanceNormalization > Log > LpNormalization > Mul > PRelu > ReduceMax > ReduceSum > Reshape > Shape > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > Please refer to the [ONNX Operator Schemas| > https://github.com/onnx/onnx/blob/master/docs/Operators.md] for more detailed > information. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Description: For the demo purpose, we need to implement these three models and their components as following: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and there're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. was: For the demo purpose, we need to implement these three models, and these are their components: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and they're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, bidaf.png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models and their > components as following: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > Add > BatchNormalization > Conv > LeakyRelu > MaxPool > Mul > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Acos > Add > BatchNormalization > Conv > Cos > Dropout > Flatten > Gemm > Identity > InstanceNormalization > LpNormalization > Mul > PRelu > Reshape > Sub > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > Abs > Add > Add > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Conv > Dropout > Gather > Hardmax > Log > LSTM > MatMul > ReduceMax > ReduceSum > Relu > Shape > Sigmoid > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > > In summary, we already implemented 13 ops, and there're still 27 ops needed > to be implemented: > h2. Already implemented: > -Acos- > -BatchNormalization- > -Cos- > -Conv- > -LeakyRelu- > -LSTM- > -Abs- > -MaxPool- > -Flatten- > -Add- > -MatMul- > -Relu- > -Sigmoid- > h2. To be implemented: > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Dropout > Gather > Gemm > Hardmax > Identity > InstanceNormalization > Log > LpNormalization > Mul > PRelu > ReduceMax > ReduceSum > Reshape > Shape > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > Please refer to the [ONNX Operator > Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for > more detailed information. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Description: For the demo purpose, we need to implement these three models, and these are their components: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] Add BatchNormalization Conv LeakyRelu MaxPool Mul h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Acos Add BatchNormalization Conv Cos Dropout Flatten Gemm Identity InstanceNormalization LpNormalization Mul PRelu Reshape Sub h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] Abs Add Add ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Conv Dropout Gather Hardmax Log LSTM MatMul ReduceMax ReduceSum Relu Shape Sigmoid Slice Squeeze Sub Sum Transpose Unsqueeze In summary, we already implemented 13 ops, and they're still 27 ops needed to be implemented: h2. Already implemented: -Acos- -BatchNormalization- -Cos- -Conv- -LeakyRelu- -LSTM- -Abs- -MaxPool- -Flatten- -Add- -MatMul- -Relu- -Sigmoid- h2. To be implemented: ArgMax Cast Ceil Clip Compress Concat ConstantOfShape Dropout Gather Gemm Hardmax Identity InstanceNormalization Log LpNormalization Mul PRelu ReduceMax ReduceSum Reshape Shape Slice Squeeze Sub Sum Transpose Unsqueeze Please refer to the [ONNX Operator Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for more detailed information. was: For the demo purpose, we need to implement these three models, and these are their components: h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] MaxPooling2D Conv2D BatchNormalization LeakyReLU Reshape h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] Conv2D BatchNormalization relu MaxPooling2D Dropout Flatten Dense Softmax l2_normalize acos cos h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] K.stack Softmax K.expand_dims K.sum Constant Dense Lambda(lambda x: 1.0 - x, output_shape=(dim,)) Multiply Add K.concatenate K.shape K.max K.tile K.squeeze linear TimeDistributed Bidirectional(LSTM In summary, we already implemented 12 ops, and there still are 16 ops needed to be implemented: h2. Already implemented: -LSTM- -Multiply- -Add- -linear- -relu- -acos- -cos- -LeakyReLU- -Softmax- -MaxPooling2D- -Conv2D- -BatchNormalization- h2. To be implemented: Reshape Flatten Dropout max shape concatenate Constant L2Normalization Expand tile squeeze Dense* TimeDistributed* Bidirectional* Stack* Lambda* *means this op doesn't have a corresponding one at ONNX op sets, therefore, it needs a converter function by using basic op sets. > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, bidaf.png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models, and these are > their components: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > Add > BatchNormalization > Conv > LeakyRelu > MaxPool > Mul > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Acos > Add > BatchNormalization > Conv > Cos > Dropout > Flatten > Gemm > Identity > InstanceNormalization > LpNormalization > Mul > PRelu > Reshape > Sub > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > Abs > Add > Add > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Conv > Dropout > Gather > Hardmax > Log > LSTM > MatMul > ReduceMax > ReduceSum > Relu > Shape > Sigmoid > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > > In summary, we already implemented 13 ops, and they're still 27 ops needed to > be implemented: > h2. Already implemented: > -Acos- > -BatchNormalization- > -Cos- > -Conv- > -LeakyRelu- > -LSTM- > -Abs- > -MaxPool- > -Flatten- > -Add- > -MatMul- > -Relu- > -Sigmoid- > h2. To be implemented: > ArgMax > Cast > Ceil > Clip > Compress > Concat > ConstantOfShape > Dropout > Gather > Gemm > Hardmax > Identity > InstanceNormalization > Log > LpNormalization > Mul > PRelu > ReduceMax > ReduceSum > Reshape > Shape > Slice > Squeeze > Sub > Sum > Transpose > Unsqueeze > Please refer to the [ONNX Operator > Schemas|[https://github.com/onnx/onnx/blob/master/docs/Operators.md]] for > more detailed information. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Attachment: bidaf.png > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, bidaf.png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models, and these are > their components: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > MaxPooling2D > Conv2D > BatchNormalization > LeakyReLU > Reshape > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Conv2D > BatchNormalization > relu > MaxPooling2D > Dropout > Flatten > Dense > Softmax > l2_normalize > acos > cos > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > K.stack > Softmax > K.expand_dims > K.sum > Constant > Dense > Lambda(lambda x: 1.0 - x, output_shape=(dim,)) > Multiply > Add > K.concatenate > K.shape > K.max > K.tile > K.squeeze > linear > TimeDistributed > Bidirectional(LSTM > > > In summary, we already implemented 12 ops, and there still are 16 ops needed > to be implemented: > h2. Already implemented: > -LSTM- > -Multiply- > -Add- > -linear- > -relu- > -acos- > -cos- > -LeakyReLU- > -Softmax- > -MaxPooling2D- > -Conv2D- > -BatchNormalization- > h2. To be implemented: > Reshape > Flatten > Dropout > max > shape > concatenate > Constant > L2Normalization > Expand > tile > squeeze > Dense* > TimeDistributed* > Bidirectional* > Stack* > Lambda* > *means this op doesn't have a corresponding one at ONNX op sets, therefore, > it needs a converter function by using basic op sets. > -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Attachment: arcface(based resnet100).png > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models, and these are > their components: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > MaxPooling2D > Conv2D > BatchNormalization > LeakyReLU > Reshape > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Conv2D > BatchNormalization > relu > MaxPooling2D > Dropout > Flatten > Dense > Softmax > l2_normalize > acos > cos > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > K.stack > Softmax > K.expand_dims > K.sum > Constant > Dense > Lambda(lambda x: 1.0 - x, output_shape=(dim,)) > Multiply > Add > K.concatenate > K.shape > K.max > K.tile > K.squeeze > linear > TimeDistributed > Bidirectional(LSTM > > > In summary, we already implemented 12 ops, and there still are 16 ops needed > to be implemented: > h2. Already implemented: > -LSTM- > -Multiply- > -Add- > -linear- > -relu- > -acos- > -cos- > -LeakyReLU- > -Softmax- > -MaxPooling2D- > -Conv2D- > -BatchNormalization- > h2. To be implemented: > Reshape > Flatten > Dropout > max > shape > concatenate > Constant > L2Normalization > Expand > tile > squeeze > Dense* > TimeDistributed* > Bidirectional* > Stack* > Lambda* > *means this op doesn't have a corresponding one at ONNX op sets, therefore, > it needs a converter function by using basic op sets. > -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Updated] (SINGA-476) Autograd operators for ONNX
[ https://issues.apache.org/jira/browse/SINGA-476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] zhangzhaoqi updated SINGA-476: -- Attachment: tiny_yolov2.png > Autograd operators for ONNX > --- > > Key: SINGA-476 > URL: https://issues.apache.org/jira/browse/SINGA-476 > Project: Singa > Issue Type: New Feature >Reporter: zhangzhaoqi >Priority: Critical > Attachments: arcface(based resnet100).png, tiny_yolov2.png > > > For the demo purpose, we need to implement these three models, and these are > their components: > h2. [Tiny yolov2|https://arxiv.org/pdf/1612.08242.pdf] > MaxPooling2D > Conv2D > BatchNormalization > LeakyReLU > Reshape > h2. [Arcface|https://arxiv.org/pdf/1801.07698.pdf] > Conv2D > BatchNormalization > relu > MaxPooling2D > Dropout > Flatten > Dense > Softmax > l2_normalize > acos > cos > h2. [BIDAF|https://arxiv.org/pdf/1611.01603.pdf] > K.stack > Softmax > K.expand_dims > K.sum > Constant > Dense > Lambda(lambda x: 1.0 - x, output_shape=(dim,)) > Multiply > Add > K.concatenate > K.shape > K.max > K.tile > K.squeeze > linear > TimeDistributed > Bidirectional(LSTM > > > In summary, we already implemented 12 ops, and there still are 16 ops needed > to be implemented: > h2. Already implemented: > -LSTM- > -Multiply- > -Add- > -linear- > -relu- > -acos- > -cos- > -LeakyReLU- > -Softmax- > -MaxPooling2D- > -Conv2D- > -BatchNormalization- > h2. To be implemented: > Reshape > Flatten > Dropout > max > shape > concatenate > Constant > L2Normalization > Expand > tile > squeeze > Dense* > TimeDistributed* > Bidirectional* > Stack* > Lambda* > *means this op doesn't have a corresponding one at ONNX op sets, therefore, > it needs a converter function by using basic op sets. > -- This message was sent by Atlassian JIRA (v7.6.14#76016)