[GitHub] andrewfayres commented on issue #12772: [MXNET-984] Add Java NDArray and introduce Java Operator Builder class
andrewfayres commented on issue #12772: [MXNET-984] Add Java NDArray and introduce Java Operator Builder class URL: https://github.com/apache/incubator-mxnet/pull/12772#issuecomment-429208229 A few comments, what this is doing is essentially spark's approach (see their [JavaDoubleRDD](https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/api/java/JavaDoubleRDD.scala)). They have a JavaDoubleRDD class that they use to just call methods of their Scala RDD and internally handle the conversions. You can also see this in their java [examples](https://github.com/apache/spark/tree/master/examples/src/main/java/org/apache/spark/examples), throughout all of the examples that I've looked at the Java examples never use any part of the core library outside of the api.java package. If we don't provide a Java friendly wrapper for all of the classes we want to expose to Java developers it'll lead to a rather confusing experience. They'll be expected to learn what is and is not meant for them to use from Scala's package while also having to know when to use the javaapi package instead. Take the IO wrapper for example, all the Java wrapper is exposing is a Java friendly version of DataDesc. If we don't wrap that and instead expect them to use the Scala one then they'll also get exposed to all the other classes in IO that aren't intended for them to use (aren't Java friendly). Any methods that we add there to make it Java friendly will pollute the Scala API with stuff that scala developers don't need. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sxjscience edited a comment on issue #12798: Fix Flaky Topk
sxjscience edited a comment on issue #12798: Fix Flaky Topk URL: https://github.com/apache/incubator-mxnet/pull/12798#issuecomment-429206097 @ankkhedia The test should have passed. The error is in batchnorm and is not related to this PR. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sxjscience commented on issue #12798: Fix Flaky Topk
sxjscience commented on issue #12798: Fix Flaky Topk URL: https://github.com/apache/incubator-mxnet/pull/12798#issuecomment-429206097 @ankkhedia The test should pass now. The error is in batchnorm and is not related to this PR. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] reminisce edited a comment on issue #12502: [NGRAPH] MXNet - nGraph initial integration
reminisce edited a comment on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-429199260 In principle, symbolic graph should not have duplicate names for different nodes, because it will introduce ambiguity into some APIs, such as `Symbol.simple_bind`, `Executor.arg_dict`, etc. I don't know what the intention was to specify the same name for two data nodes in that unit test, but I think it should be changed to using different names even though using the same name in that test had no negative effect . This is an automated message from the Apache Git Service. To respond to the message, please log on 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] reminisce commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration
reminisce commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-429199260 In principle, symbolic graph should not have duplicate names for different nodes, because it will introduce ambiguity into some APIs, such as `Symbol.simple_bind`, `Executor.arg_dict`, etc. I don't know what the intention was to specify the same name for two data nodes in that unit test, but I think it should be changed to using different names even using the same name in that test had no negative effect . This is an automated message from the Apache Git Service. To respond to the message, please log on 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] yzhliu commented on a change in pull request #12772: [MXNET-984] Add Java NDArray and introduce Java Operator Builder class
yzhliu commented on a change in pull request #12772: [MXNET-984] Add Java NDArray and introduce Java Operator Builder class URL: https://github.com/apache/incubator-mxnet/pull/12772#discussion_r224664357 ## File path: Makefile ## @@ -606,7 +606,7 @@ scalaclean: scalapkg: (cd $(ROOTDIR)/scala-package; \ - mvn package -P$(SCALA_PKG_PROFILE),$(SCALA_VERSION_PROFILE) -Dcxx="$(CXX)" \ + mvn package -P$(SCALA_PKG_PROFILE),$(SCALA_VERSION_PROFILE),integrationtest -Dcxx="$(CXX)" \ Review comment: why running test here? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] gaurav-gireesh commented on a change in pull request #12697: [MXNET -1004] Poisson NegativeLog Likelihood loss
gaurav-gireesh commented on a change in pull request #12697: [MXNET -1004] Poisson NegativeLog Likelihood loss URL: https://github.com/apache/incubator-mxnet/pull/12697#discussion_r224663176 ## File path: tests/python/unittest/test_loss.py ## @@ -348,6 +348,44 @@ def test_triplet_loss(): optimizer='adam') assert mod.score(data_iter, eval_metric=mx.metric.Loss())[0][1] < 0.05 +@with_seed() +def test_poisson_nllloss(): +pred = mx.nd.random.normal(shape=(3, 4)) +min_pred = mx.nd.min(pred) +#This is necessary to ensure only positive random values are generated for prediction, +# to avoid ivalid log calculation +pred[:] = pred + mx.nd.abs(min_pred) +target = mx.nd.random.normal(shape=(3, 4)) +min_target = mx.nd.min(target) +#This is necessary to ensure only positive random values are generated for prediction, +# to avoid ivalid log calculation +target[:] += mx.nd.abs(min_target) + +Loss = gluon.loss.PoissonNLLLoss(from_logits=True) +Loss_no_logits = gluon.loss.PoissonNLLLoss(from_logits=False) +#Calculating by brute formula for default value of from_logits = True + +# 1) Testing for flag logits = True +brute_loss = np.mean(np.exp(pred.asnumpy()) - target.asnumpy() * pred.asnumpy()) +loss_withlogits = Loss(pred, target) +assert_almost_equal(brute_loss, loss_withlogits.asscalar()) + +#2) Testing for flag logits = False +loss_no_logits = Loss_no_logits(pred, target) +np_loss_no_logits = np.mean(pred.asnumpy() - target.asnumpy() * np.log(pred.asnumpy() + 1e-08)) +if np.isnan(loss_no_logits.asscalar()): +assert_almost_equal(np.isnan(np_loss_no_logits), np.isnan(loss_no_logits.asscalar())) +else: +assert_almost_equal(np_loss_no_logits, loss_no_logits.asscalar()) + +#3) Testing for Sterling approximation +np_pred = np.random.uniform(1, 5, (2, 3)) Review comment: > Please add tests for hybridized version as well. Added the unit test for it. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] gaurav-gireesh commented on a change in pull request #12697: [MXNET -1004] Poisson NegativeLog Likelihood loss
gaurav-gireesh commented on a change in pull request #12697: [MXNET -1004] Poisson NegativeLog Likelihood loss URL: https://github.com/apache/incubator-mxnet/pull/12697#discussion_r224663128 ## File path: tests/python/unittest/test_loss.py ## @@ -348,6 +348,44 @@ def test_triplet_loss(): optimizer='adam') assert mod.score(data_iter, eval_metric=mx.metric.Loss())[0][1] < 0.05 +@with_seed() +def test_poisson_nllloss(): +pred = mx.nd.random.normal(shape=(3, 4)) +min_pred = mx.nd.min(pred) +#This is necessary to ensure only positive random values are generated for prediction, +# to avoid ivalid log calculation +pred[:] = pred + mx.nd.abs(min_pred) +target = mx.nd.random.normal(shape=(3, 4)) +min_target = mx.nd.min(target) +#This is necessary to ensure only positive random values are generated for prediction, +# to avoid ivalid log calculation +target[:] += mx.nd.abs(min_target) + +Loss = gluon.loss.PoissonNLLLoss(from_logits=True) +Loss_no_logits = gluon.loss.PoissonNLLLoss(from_logits=False) +#Calculating by brute formula for default value of from_logits = True + +# 1) Testing for flag logits = True +brute_loss = np.mean(np.exp(pred.asnumpy()) - target.asnumpy() * pred.asnumpy()) +loss_withlogits = Loss(pred, target) +assert_almost_equal(brute_loss, loss_withlogits.asscalar()) Review comment: Unit test added. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] pengzhao-intel commented on issue #12798: Fix Flaky Topk
pengzhao-intel commented on issue #12798: Fix Flaky Topk URL: https://github.com/apache/incubator-mxnet/pull/12798#issuecomment-429187361 Thanks, @sxjscience, please help verify the performance after topK code changes. https://github.com/apache/incubator-mxnet/pull/12085#issuecomment-411657593 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] pengzhao-intel commented on issue #12710: Subprocess Deadlock with mxnet-mkl
pengzhao-intel commented on issue #12710: Subprocess Deadlock with mxnet-mkl URL: https://github.com/apache/incubator-mxnet/issues/12710#issuecomment-429186779 We start to look at the issue and will back soon :) This is an automated message from the Apache Git Service. To respond to the message, please log on 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] pengzhao-intel commented on issue #12031: Fix CPUPinned unexpected behaviour
pengzhao-intel commented on issue #12031: Fix CPUPinned unexpected behaviour URL: https://github.com/apache/incubator-mxnet/pull/12031#issuecomment-429185772 @eric-haibin-lin Any update on this PR? It opens for a long time and we prefer to leverage the feature in this PR :) This is an automated message from the Apache Git Service. To respond to the message, please log on 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] DickJC123 opened a new pull request #12804: CudnnFind() usage improvements
DickJC123 opened a new pull request #12804: CudnnFind() usage improvements URL: https://github.com/apache/incubator-mxnet/pull/12804 ## Description ## This PR improves the MXNet's use of cudnnFind() to address a few issues: 1. With the gluon imperative style, cudnnFind() is called during forward(), and so might have its timings perturbed by other GPU activity (including potentially other cudnnFind() calls). 2. With some cuda drivers versions, care is needed to ensure that the large I/O and workspace cudaMallocs() performed by cudnnFind() are immediately released and available to MXNet. 3. cudnnFind() makes both conv I/O and workspace allocations that must be covered by the GPU global memory headroom defined by MXNET_GPU_MEM_POOL_RESERVE. Per issue #12662, large convolutions can result in out-of-memory errors, even when MXNet's storage allocator has free memory in its pool. This PR addresses these issues, providing the following benefits: 1. Consistent algo choice for a given convolution type in a model, both for instances in the same GPU and in other GPUs in a multi-GPU training setting. 2. Consistent algo choice from run to run, based on eliminating sources of interference of the cudnnFind() timing process. 3. Consistent model global memory footprint, both because of the consistent algo choice (algo's can have markedly different workspace requirements) and changes to MXNet's use of cudaMalloc. 4. Increased training performance based on being able to consistently run with models that approach the GPU's full global memory footprint. 5. Adds a unittest for and solves issue #12662. Now specifically, the PR makes the following changes/additions: 1. Merges the cudnn algo registry separate Find() and Register() methods into a single FindOrElseRegister() call protected by a mutex. Before, there was a window of time between when a thread found no algo entry and when it registered a new entry. This window could allow a different thread to call cudnnFind() on the same convolution type, getting a different answer. FindOrElseRegister() takes a callback that determines the algo via cudnnFind() and makes access to the cudnn algo registry effectively atomic. The callbacks provided with cudnn_convolution and cudnn_deconvolution also grab the storage allocator's mutex to quiet the GPU activity from other threads during the running of cudnnFind(). 2. Changes MXNet's storage allocator to round large allocations to a multiple of 2MB. This value ensures that small permanent allocations that might have shared a cuda memory allocation slab with a prior large allocation don't prevent the reclaiming of that slab area. 3. Includes use of the MXNet storage allocator's Alloc() and DirectFree() API to coax a convolution's I/O storage needs away from MXNet's storage allocator prior to calling cudnnFind(). The first commit introduces a unittest that demonstrates issue #12662 in the current codebase. In order to make the test adaptable to GPUs of varying global memory, a C-level API was brought out to the python API by introducing mx.context.gpu_memory_info(dev_id). Follow-up commits supply the fix and the rest of the PR. Hopefully Interested parties: @KellenSunderland @piiswrong @eric-haibin-lin @zheng-da ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) created (except PRs with tiny changes) - [ X] Changes are complete (i.e. I finished coding on this PR) - [X ] All changes have test coverage: - Unit tests are added for small changes to verify correctness (e.g. adding a new operator) - Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore) - Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL) - [X ] Code is well-documented: - For user-facing API changes, API doc string has been updated. - For new C++ functions in header files, their functionalities and arguments are documented. - For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable - Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html - [X ] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change ### Changes ### - [ ] Feature1, tests, (and when applicable, API doc) - [ ] Feature2, tests, (and when applicable, API doc) ## Comments ## The following command was run on an 8-GPU system to demonstrate the new
[GitHub] xinyu-intel commented on issue #12780: Add more models to benchmark_score
xinyu-intel commented on issue #12780: Add more models to benchmark_score URL: https://github.com/apache/incubator-mxnet/pull/12780#issuecomment-429177958 @sandeep-krishnamurthy Thanks for suggestions. It can handle multi-GPUs now:) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] xinyu-intel commented on a change in pull request #12724: simplify mac mkldnn build
xinyu-intel commented on a change in pull request #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#discussion_r224648886 ## File path: MKLDNN_README.md ## @@ -70,24 +69,10 @@ git clone --recursive https://github.com/apache/incubator-mxnet.git cd incubator-mxnet ``` -### Enable OpenMP for MacOS - -If you want to enable OpenMP for better performance, you should modify the Makefile in MXNet root dictionary: - -Add CFLAGS '-fopenmp' for Darwin. - -``` -ifeq ($(USE_OPENMP), 1) -# ifneq ($(UNAME_S), Darwin) -CFLAGS += -fopenmp -# endif -endif -``` - ### Build MXNet with MKL-DNN ``` -make -j $(sysctl -n hw.ncpu) CC=gcc-4.9 CXX=g++-4.9 USE_OPENCV=0 USE_OPENMP=1 USE_MKLDNN=1 USE_BLAS=apple USE_PROFILER=1 +LIBRARY_PATH=$(brew --prefix llvm)/lib/ make -j $(sysctl -n hw.ncpu) CC=$(brew --prefix llvm)/bin/clang++ CXX=$(brew --prefix llvm)/bin/clang++ USE_OPENCV=0 USE_OPENMP=1 USE_MKLDNN=1 USE_BLAS=apple USE_PROFILER=1 Review comment: USE_OPENCV=1 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] xinyu-intel commented on a change in pull request #12724: simplify mac mkldnn build
xinyu-intel commented on a change in pull request #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#discussion_r224648857 ## File path: MKLDNN_README.md ## @@ -70,24 +69,10 @@ git clone --recursive https://github.com/apache/incubator-mxnet.git cd incubator-mxnet ``` -### Enable OpenMP for MacOS - -If you want to enable OpenMP for better performance, you should modify the Makefile in MXNet root dictionary: - -Add CFLAGS '-fopenmp' for Darwin. - -``` -ifeq ($(USE_OPENMP), 1) -# ifneq ($(UNAME_S), Darwin) -CFLAGS += -fopenmp -# endif -endif -``` - ### Build MXNet with MKL-DNN ``` -make -j $(sysctl -n hw.ncpu) CC=gcc-4.9 CXX=g++-4.9 USE_OPENCV=0 USE_OPENMP=1 USE_MKLDNN=1 USE_BLAS=apple USE_PROFILER=1 +LIBRARY_PATH=$(brew --prefix llvm)/lib/ make -j $(sysctl -n hw.ncpu) CC=$(brew --prefix llvm)/bin/clang++ CXX=$(brew --prefix llvm)/bin/clang++ USE_OPENCV=0 USE_OPENMP=1 USE_MKLDNN=1 USE_BLAS=apple USE_PROFILER=1 ``` *Note: Temporarily disable OPENCV.* Review comment: Please remove this line since llvm support opencv well now. This is an automated message from the Apache Git Service. To respond to the message, please log on 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 49fc4f0 Bump the publish timestamp. 49fc4f0 is described below commit 49fc4f08e20c993790cbeab05e3edcbd39e87bf0 Author: mxnet-ci AuthorDate: Fri Oct 12 00:57:28 2018 + 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..1f6911d --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Fri Oct 12 00:57:27 UTC 2018
[GitHub] safrooze commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...`
safrooze commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...` URL: https://github.com/apache/incubator-mxnet/issues/12795#issuecomment-429151434 @lostella while the issue is being root caused, one work around in this case would be to use different blocks with shared parameters: ``` class MyBlock(mx.gluon.HybridBlock): def __init__(self): super().__init__() with self.name_scope(): self.model0 = mx.gluon.nn.Dense(units=5) self.model1 = mx.gluon.nn.Dense(units=5, params=self.model0.collect_params()) def hybrid_forward(self, F, x, y): return self.model0(x) + self.model1(y) ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] vandanavk commented on issue #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
vandanavk commented on issue #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#issuecomment-429142991 @anirudhacharya Addressed the comments above. Working on the test for square export. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on a change in pull request #11778: [MXNET-483] C++ tests for mkldnn convolution/deconvolution operator
azai91 commented on a change in pull request #11778: [MXNET-483] C++ tests for mkldnn convolution/deconvolution operator URL: https://github.com/apache/incubator-mxnet/pull/11778#discussion_r224624098 ## File path: tests/cpp/operator/mkldnn.cc ## @@ -515,14 +516,81 @@ OpAttrs GetConcatBackwardsOp(int num_args, int dim) { std::string CreateShapeString(int value, int dim) { std::stringstream ss; ss << "("; - for (int i = 0; i < dim; i++) { + for (size_t i = 0; i < dim; ++i) { ss << value; if (i != dim - 1) ss << ","; } ss << ")"; return ss.str(); } +OpAttrs GetConvOp(int kernel, int num_filters, int dim, int stride, int pad) { + OpAttrs attrs; + attrs.attrs.op = Op::Get("Convolution"); + attrs.num_inputs = 3; + attrs.num_outputs = 1; + attrs.attrs.dict.insert({"kernel" , CreateShapeString(kernel, dim)}); + attrs.attrs.dict.insert({"num_filter" , std::to_string(num_filters)}); + attrs.attrs.dict.insert({"stride" , CreateShapeString(stride, dim)}); + attrs.attrs.dict.insert({"pad" , CreateShapeString(pad, dim)}); + attrs.attrs.op->attr_parser(); + attrs.input_types = ArrayTypes::Normal | + ArrayTypes::MKLDNN | + ArrayTypes::NormalReshaped | + ArrayTypes::MKLDNNReshaped | + ArrayTypes::NormalReused | + ArrayTypes::MKLDNNReused | + ArrayTypes::NormalReshapedReused; + attrs.output_types = ArrayTypes::Normal | + ArrayTypes::MKLDNN | + ArrayTypes::NormalReshaped | + ArrayTypes::MKLDNNReshaped | + ArrayTypes::NormalReused | + ArrayTypes::MKLDNNReused | + ArrayTypes::NormalReshapedReused | + ArrayTypes::NormalReusedDiffDtype; + return attrs; +} + +OpAttrs GetConvBackwardOp(int kernel, int num_filters, int dim, int stride, int pad) { + OpAttrs attrs; + attrs.attrs.op = Op::Get("_backward_Convolution"); + attrs.num_inputs = 4; + attrs.num_outputs = 3; + attrs.attrs.dict.insert({"kernel" , CreateShapeString(kernel, dim)}); + attrs.attrs.dict.insert({"num_filter" , std::to_string(num_filters)}); + attrs.attrs.dict.insert({"stride" , CreateShapeString(stride, dim)}); + attrs.attrs.dict.insert({"pad" , CreateShapeString(pad, dim)}); + attrs.attrs.op->attr_parser(); Review comment: input and output are just specified in the forward operator. the output of the forward is the input of the backwards. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sandeep-krishnamurthy edited a comment on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...`
sandeep-krishnamurthy edited a comment on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...` URL: https://github.com/apache/incubator-mxnet/issues/12795#issuecomment-429130537 https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/gluon/block.py#L1050 this is the new addition after 1.2.1 which calls output symbol.get_internals() and then finds duplicate names and fails, exposing the duplicate names issue from 1.3 onwards. Duplicate output names when we have 2 inputs passing through the same block was always the case, I am not very sure why this has not created issues for our users. Trying to root cause and solution with the help of @safrooze and @zhreshold @szha This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ShownX closed issue #12334: Load pretrained model with new load_parameters() API
ShownX closed issue #12334: Load pretrained model with new load_parameters() API URL: https://github.com/apache/incubator-mxnet/issues/12334 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so
lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429135619 @sameer-kapoor @milandesai We are trying to reproduce the issue you may facing by the steps came from the CI: ``` bash ci/docker/install/centos7_core.sh yum install -y maven yum install -y java-1.8.0-openjdk ``` And build the MXNet backend with the following command: https://github.com/apache/incubator-mxnet/blob/master/ci/docker/runtime_functions.sh#L239-L253 We tested with python install and it succeeds with: ``` pip install -U -e . ``` However, Scala contains linking issues: ``` undefined symbol: _ZN2ps4Meta6kEmptyE (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: __gcov_merge_add(./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8Customer11WaitRequestEi (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10PostofficeC1Ev (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10Postoffice18GetServerKeyRangesEv (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: __gcov_init(./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8CustomerC1EiiRKSt8functionIFvRKNS_7MessageEEE (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8Customer10NewRequestEi (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10Postoffice12GetDeadNodesEi (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8Customer11NumResponseEi (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8Customer11AddResponseEii (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10Postoffice8FinalizeEib (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10Postoffice5StartEiPKcb (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps8CustomerD1Ev (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps10Postoffice7BarrierEii (./libmxnet-init-scala-linux-x86_64.so) undefined symbol: _ZN2ps3Van4SendERKNS_7MessageE (./libmxnet-init-scala-linux-x86_64.so) ``` From the community feedback from @szha , `ps-lite` and `test-coverage` dependencies are missing. We are diagnosing the issue and see if we can get it work. In the meantime, can you check with ``` ldd -r libmxnet.so ``` To see if there is any linking problem? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] Vikas89 closed issue #12800: getting segfault while running train_cifar10.py program in example directory
Vikas89 closed issue #12800: getting segfault while running train_cifar10.py program in example directory URL: https://github.com/apache/incubator-mxnet/issues/12800 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] Vikas89 commented on issue #12800: getting segfault while running train_cifar10.py program in example directory
Vikas89 commented on issue #12800: getting segfault while running train_cifar10.py program in example directory URL: https://github.com/apache/incubator-mxnet/issues/12800#issuecomment-429133125 I am not seeing this issue now, closing this This is an automated message from the Apache Git Service. To respond to the message, please log on 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] milandesai commented on issue #12802: Crash while loading libmxnet-scala.so
milandesai commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429132263 Hello, to build the C++ library, we simply run 'make' on a CentOS 7 machine. We build the Scala package using a custom SBT script so that we can create artifacts for Scala 2.12. We will run some tests on a sample project using Scala 2.11 to see if the same issue arises when we run `make scalapkg`. That way we can rule out whether there's an issue with our Scala build process. As to @piyushghai 's last comment,`opencv-devel` and `atlas-devel` are already installed on the machine. Additionally, modification of the mshadow linker flags is no longer needed. I submitted a change a while back that sets those flags automatically for CentOS machines, and I can see during the build process that the flags are indeed set. Without those packages and linker flags, there would be a linker failure. Currently, the entire compilation and linkage completes successfully. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12591: USE_MKLDNN=1 is default in make build (mkldnn must be explicitly turned off)
azai91 commented on issue #12591: USE_MKLDNN=1 is default in make build (mkldnn must be explicitly turned off) URL: https://github.com/apache/incubator-mxnet/pull/12591#issuecomment-429131687 @juliusshufan can you provide benchmarks comparing mkldnn vs non? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sandeep-krishnamurthy commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...`
sandeep-krishnamurthy commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...` URL: https://github.com/apache/incubator-mxnet/issues/12795#issuecomment-429130537 https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/gluon/block.py#L1050 this is the new addition after 1.2.1 which calls output symbol.get_internals() and then finds duplicate names and fails. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] lostella commented on issue #12783: Name conflict when serializing LSTMCell
lostella commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429128509 @szha yes, thanks. That's basically #12794. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] mbrookhart edited a comment on issue #12502: [NGRAPH] MXNet - nGraph initial integration
mbrookhart edited a comment on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-429127369 Hi @eric-haibin-lin, the test failure is related to a change in the subgraph API here: https://github.com/apache/incubator-mxnet/pull/12530 that broke this test's use of duplicated input parameter names. We're trying to figure out how to get around it while supporting the MKLDNN use case, but don't have a full solution yet. We could skip this particular test as an edge case, but we'd prefer to fix the subgraph API to support it. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] mbrookhart commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration
mbrookhart commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-429127369 Hi @eric-haibin-lin, the test failure is related to a change in the subgraph API here: https://github.com/apache/incubator-mxnet/pull/12530 that broke with this tests use of duplicated input parameter names. We're trying to figure out how to get around it while supporting the MKLDNN use case, but don't have a full solution yet. We could skip this particular test as an edge case, but we'd prefer to fix the subgraph API to support it. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] eric-haibin-lin commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration
eric-haibin-lin commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-429122386 Is the test failure relevant? This is an automated message from the Apache Git Service. To respond to the message, please log on 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 (213ab09 -> 06bf600)
This is an automated email from the ASF dual-hosted git repository. haibin pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. from 213ab09 Updated tvm submodule head (#12764) add 06bf600 Adagrad optimizer with row-wise learning rate (#12365) No new revisions were added by this update. Summary of changes: docs/api/python/index.md | 1 + docs/api/python/optimization/contrib.md| 52 + .../opencv => python/mxnet/optimizer}/__init__.py | 9 +- python/mxnet/optimizer/contrib.py | 100 + python/mxnet/{ => optimizer}/optimizer.py | 20 +- python/mxnet/test_utils.py | 41 src/operator/contrib/optimizer_op-inl.h| 247 + src/operator/contrib/optimizer_op.cc | 89 .../contrib/{transformer.cu => optimizer_op.cu}| 14 +- tests/python/unittest/test_contrib_optimizer.py| 100 + tests/python/unittest/test_optimizer.py| 37 --- 11 files changed, 655 insertions(+), 55 deletions(-) create mode 100644 docs/api/python/optimization/contrib.md copy {plugin/opencv => python/mxnet/optimizer}/__init__.py (83%) create mode 100644 python/mxnet/optimizer/contrib.py rename python/mxnet/{ => optimizer}/optimizer.py (98%) create mode 100644 src/operator/contrib/optimizer_op-inl.h create mode 100644 src/operator/contrib/optimizer_op.cc copy src/operator/contrib/{transformer.cu => optimizer_op.cu} (78%) create mode 100644 tests/python/unittest/test_contrib_optimizer.py
[GitHub] eric-haibin-lin closed pull request #12365: Adagrad optimizer with row-wise learning rate
eric-haibin-lin closed pull request #12365: Adagrad optimizer with row-wise learning rate URL: https://github.com/apache/incubator-mxnet/pull/12365 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake of provenance: As this is a foreign pull request (from a fork), the diff is supplied below (as it won't show otherwise due to GitHub magic): diff --git a/docs/api/python/index.md b/docs/api/python/index.md index 8f60bcd0f13..de86aedff69 100644 --- a/docs/api/python/index.md +++ b/docs/api/python/index.md @@ -136,6 +136,7 @@ Code examples are placed throughout the API documentation and these can be run a :maxdepth: 1 optimization/optimization.md + optimization/contrib.md ``` ## Profiler API diff --git a/docs/api/python/optimization/contrib.md b/docs/api/python/optimization/contrib.md new file mode 100644 index 000..8fc261f4f05 --- /dev/null +++ b/docs/api/python/optimization/contrib.md @@ -0,0 +1,52 @@ +# Contrib Optimization API + +```eval_rst +.. currentmodule:: mxnet.optimizer.contrib +``` + +## Overview + +This document summaries the contrib APIs used to initialize and update the model +weights during training + +```eval_rst +.. autosummary:: +:nosignatures: + +mxnet.optimizer.contrib +``` + +The `Contrib Optimization` API, defined in the `optimizer.contrib` package, provides +many useful experimental APIs for new features. +This is a place for the community to try out the new features, +so that feature contributors can receive feedback. + +```eval_rst +.. warning:: This package contains experimental APIs and may change in the near future. +``` + +In the rest of this document, we list routines provided by the `optimizer.contrib` package. + +## Contrib + +```eval_rst +.. currentmodule:: mxnet.optimizer.contrib + +.. autosummary:: +:nosignatures: + + GroupAdaGrad +``` + +## API Reference + + + +```eval_rst + +.. automodule:: mxnet.optimizer.contrib +:members: + +``` + +auto_index("api-reference"); diff --git a/python/mxnet/optimizer/__init__.py b/python/mxnet/optimizer/__init__.py new file mode 100644 index 000..72eb5a74152 --- /dev/null +++ b/python/mxnet/optimizer/__init__.py @@ -0,0 +1,24 @@ +# 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. +"""Optimizer API of MXNet.""" + +from . import optimizer, contrib +# pylint: disable=wildcard-import +from .optimizer import * +# pylint: enable=wildcard-import + +__all__ = optimizer.__all__ + ['contrib'] diff --git a/python/mxnet/optimizer/contrib.py b/python/mxnet/optimizer/contrib.py new file mode 100644 index 000..d269aa1bd06 --- /dev/null +++ b/python/mxnet/optimizer/contrib.py @@ -0,0 +1,100 @@ +# coding: utf-8 +# 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. + +# pylint: disable=too-many-lines +"""Contrib optimizers.""" +from ..ndarray import (NDArray, clip, contrib, mean, sqrt, square, zeros) +from .optimizer import Optimizer + +# convenience wrapper for Optimizer.Register +register = Optimizer.register # pylint: disable=invalid-name + +__all__ = ['GroupAdaGrad'] + + +@register +class GroupAdaGrad(Optimizer): +"""Adagrad optimizer with row-wise learning rates. + +This class implements the AdaGrad optimizer described in *Adaptive +Subgradient Methods for Online Learning and Stochastic Optimization*, and +available at http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf but +uses only a single learning rate
[GitHub] ChaiBapchya opened a new pull request #12749: [MXNET-1029] Feature request: randint operator
ChaiBapchya opened a new pull request #12749: [MXNET-1029] Feature request: randint operator URL: https://github.com/apache/incubator-mxnet/pull/12749 ## Description ## Incorporates the feature request to support randint operator. Fixes #11218 ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [X] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) created (except PRs with tiny changes) - [X] Changes are complete (i.e. I finished coding on this PR) - [X] All changes have test coverage: - Unit tests are added for small changes to verify correctness (e.g. adding a new operator) - Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore) - Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL) - [X] Code is well-documented: - For user-facing API changes, API doc string has been updated. - For new C++ functions in header files, their functionalities and arguments are documented. - For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable - Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html - [X] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12724: simplify mac mkldnn build
azai91 commented on issue #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#issuecomment-429114619 @mseth10 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224595446 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -2077,3 +2112,79 @@ def convert_sqrt(node, **kwargs): name=name, ) return [node] + +@mx_op.register("square") +def convert_square(node, **kwargs): +"""Map MXNet's square operator attributes to onnx's Pow operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] + +input_node_a_id = kwargs["index_lookup"][inputs[0][0]] +input_node_a = proc_nodes[input_node_a_id].name + +initializer = kwargs["initializer"] +np_arr = np.array([2]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] +dims = np.shape(np_arr) Review comment: can this be removed? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224597228 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -214,17 +222,44 @@ def convert_fully_connected(node, **kwargs): onnx = import_onnx_modules() name = node["name"] inputs = node["inputs"] +attrs = node["attrs"] +initializer = kwargs["initializer"] + +no_bias = get_boolean_attribute_value(attrs, "no_bias") + input_node_id = kwargs["index_lookup"][inputs[0][0]] weight_node_id = kwargs["index_lookup"][inputs[1][0]] -bias_node_id = kwargs["index_lookup"][inputs[2][0]] + proc_nodes = kwargs["proc_nodes"] -input_node = proc_nodes[input_node_id] -weights_node = proc_nodes[weight_node_id] -bias_node = proc_nodes[bias_node_id] +input_node = proc_nodes[input_node_id] input_name = input_node.name + +weights_node = proc_nodes[weight_node_id] weights_name = weights_node.name -bias_name = bias_node.name + +fcnode = [] + +if no_bias == 0: +bias_node_id = kwargs["index_lookup"][inputs[2][0]] +bias_node = proc_nodes[bias_node_id] +bias_name = bias_node.name +else: +np_arr = np.array([0]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] Review comment: please use `np.int64` here and remove the previous line This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224595328 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -2077,3 +2112,79 @@ def convert_sqrt(node, **kwargs): name=name, ) return [node] + +@mx_op.register("square") +def convert_square(node, **kwargs): +"""Map MXNet's square operator attributes to onnx's Pow operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] + +input_node_a_id = kwargs["index_lookup"][inputs[0][0]] +input_node_a = proc_nodes[input_node_a_id].name + +initializer = kwargs["initializer"] +np_arr = np.array([2]) Review comment: can you please remove this line This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224596258 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -2077,3 +2112,79 @@ def convert_sqrt(node, **kwargs): name=name, ) return [node] + +@mx_op.register("square") +def convert_square(node, **kwargs): +"""Map MXNet's square operator attributes to onnx's Pow operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] + +input_node_a_id = kwargs["index_lookup"][inputs[0][0]] +input_node_a = proc_nodes[input_node_a_id].name + +initializer = kwargs["initializer"] +np_arr = np.array([2]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] +dims = np.shape(np_arr) + +power2_name = "square_tensor" + str(kwargs["idx"]) +tensor_node = onnx.helper.make_tensor_value_info(power2_name, data_type, dims) +initializer.append( +onnx.helper.make_tensor( +name=power2_name, +data_type=data_type, +dims=dims, +vals=[2], +raw=False, +) +) + +node = onnx.helper.make_node( +"Pow", +[input_node_a, power2_name], +[name], +name=name +) +return [tensor_node, node] + +@mx_op.register("sum") +def convert_sum(node, **kwargs): +"""Map MXNet's sum operator attributes to onnx's ReduceSum operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] +attrs = node["attrs"] + +mx_axis = attrs.get("axis", None) +axes = convert_string_to_list(str(mx_axis)) if mx_axis is not None else None + +keepdims = get_boolean_attribute_value(attrs, "keepdims") + +input_node_id = kwargs["index_lookup"][inputs[0][0]] +input_node = proc_nodes[input_node_id].name + +if axes is not None: Review comment: can this be `if axes` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12746: Add env flag to disable MKLDNN cache (MXNET_MKLDNN_CACHE_ENABLED)
azai91 commented on issue #12746: Add env flag to disable MKLDNN cache (MXNET_MKLDNN_CACHE_ENABLED) URL: https://github.com/apache/incubator-mxnet/pull/12746#issuecomment-429114430 @mseth10 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224595541 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -2077,3 +2112,79 @@ def convert_sqrt(node, **kwargs): name=name, ) return [node] + +@mx_op.register("square") +def convert_square(node, **kwargs): +"""Map MXNet's square operator attributes to onnx's Pow operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] + +input_node_a_id = kwargs["index_lookup"][inputs[0][0]] +input_node_a = proc_nodes[input_node_a_id].name + +initializer = kwargs["initializer"] +np_arr = np.array([2]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] +dims = np.shape(np_arr) + +power2_name = "square_tensor" + str(kwargs["idx"]) +tensor_node = onnx.helper.make_tensor_value_info(power2_name, data_type, dims) +initializer.append( +onnx.helper.make_tensor( +name=power2_name, +data_type=data_type, +dims=dims, Review comment: use tuple `(1,)` here This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224595370 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -2077,3 +2112,79 @@ def convert_sqrt(node, **kwargs): name=name, ) return [node] + +@mx_op.register("square") +def convert_square(node, **kwargs): +"""Map MXNet's square operator attributes to onnx's Pow operator +and return the created node. +""" +onnx = import_onnx_modules() +name = node["name"] +proc_nodes = kwargs["proc_nodes"] +inputs = node["inputs"] + +input_node_a_id = kwargs["index_lookup"][inputs[0][0]] +input_node_a = proc_nodes[input_node_a_id].name + +initializer = kwargs["initializer"] +np_arr = np.array([2]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] Review comment: use np.int64 here This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square
anirudhacharya commented on a change in pull request #12646: ONNX export: Fully connected operator w/o bias, ReduceSum, Square URL: https://github.com/apache/incubator-mxnet/pull/12646#discussion_r224597492 ## File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py ## @@ -214,17 +222,44 @@ def convert_fully_connected(node, **kwargs): onnx = import_onnx_modules() name = node["name"] inputs = node["inputs"] +attrs = node["attrs"] +initializer = kwargs["initializer"] + +no_bias = get_boolean_attribute_value(attrs, "no_bias") + input_node_id = kwargs["index_lookup"][inputs[0][0]] weight_node_id = kwargs["index_lookup"][inputs[1][0]] -bias_node_id = kwargs["index_lookup"][inputs[2][0]] + proc_nodes = kwargs["proc_nodes"] -input_node = proc_nodes[input_node_id] -weights_node = proc_nodes[weight_node_id] -bias_node = proc_nodes[bias_node_id] +input_node = proc_nodes[input_node_id] input_name = input_node.name + +weights_node = proc_nodes[weight_node_id] weights_name = weights_node.name -bias_name = bias_node.name + +fcnode = [] + +if no_bias == 0: +bias_node_id = kwargs["index_lookup"][inputs[2][0]] +bias_node = proc_nodes[bias_node_id] +bias_name = bias_node.name +else: +np_arr = np.array([0]) +data_type = onnx.mapping.NP_TYPE_TO_TENSOR_TYPE[np_arr.dtype] +dims = np.shape(np_arr) +bias_name = "bias" + str(kwargs["idx"]) +tensor_node = onnx.helper.make_tensor_value_info(bias_name, data_type, dims) +initializer.append( +onnx.helper.make_tensor( +name=bias_name, +data_type=data_type, +dims=dims, Review comment: you can directly use the tuple value `(1,)` and remove the dims variable. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai edited a comment on issue #12802: Crash while loading libmxnet-scala.so
piyushghai edited a comment on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429110352 @sameer-kapoor Can you try modifying the MSHADOW Flags as mentioned here before building MXNet from source? https://philipskokoh.github.io/blog/mxnet-on-centos As a pre-requisite, can you also install ```opencv-devel``` and ```atlas-devel``` before running the make scalapkg command ? Let mw know if this works for you. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] marcoabreu commented on issue #12803: Fix Extremely Unbalanced Issue in split_and_load
marcoabreu commented on issue #12803: Fix Extremely Unbalanced Issue in split_and_load URL: https://github.com/apache/incubator-mxnet/pull/12803#issuecomment-429110890 Can we double check that this is properly covered by tests? I'm afraid this might be a behaviour change, so we might run into compatibility issues This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so
piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429110352 @sameer-kapoor Can you try modifying the MSHADOW Flags as mentioned here before building MXNet from source? https://philipskokoh.github.io/blog/mxnet-on-centos This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sandeep-krishnamurthy commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...`
sandeep-krishnamurthy commented on issue #12795: Deserialization problem with gluon `ValueError: There are multiple outputs with name ...` URL: https://github.com/apache/incubator-mxnet/issues/12795#issuecomment-429109606 I observed that the issue occurs when 2 inputs pass through the same block. Trying to understand the root cause for this as it works fine with mxnet 1.2.1 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so
piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429108766 Here's a reference of the Scala Package building steps : https://github.com/apache/incubator-mxnet/tree/master/scala-package#build This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12724: simplify mac mkldnn build
azai91 commented on issue #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#issuecomment-429108625 updated the docs to support using llvm. @pengzhao-intel can you review this and if it looks good we can merge. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so
lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429108402 @sameer-kapoor Could you please bring the steps you have taken to build from source? Note: We haven't tested the Scala package build on Cent OS. We do a test on Ubuntu 16.04 and above. This is an automated message from the Apache Git Service. To respond to the message, please log on 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 4daefd3 Bump the publish timestamp. 4daefd3 is described below commit 4daefd34cdfeb21b271392132415a38d08e4d7b3 Author: mxnet-ci AuthorDate: Thu Oct 11 20:28:41 2018 + 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..dd16cfb --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Thu Oct 11 20:28:41 UTC 2018
[GitHub] sameer-kapoor commented on issue #12802: Crash while loading libmxnet-scala.so
sameer-kapoor commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429106659 @lanking520 We build from source using SBT and on the same OS versions as where the tests are run. We haven't tried the Maven jar package since we use Scala 2.12. As far as the crash goes there is nothing else available to us other than what I have posted. If you'd like like I can post the thread dump from the hs_err_pid.log. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] larroy commented on issue #12094: [MXNET-793] ★ Virtualized testing in CI with QEMU ★
larroy commented on issue #12094: [MXNET-793] ★ Virtualized testing in CI with QEMU ★ URL: https://github.com/apache/incubator-mxnet/pull/12094#issuecomment-429106341 @mxnet-label-bot: [pr-awaiting-review] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so
piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429102671 @mxnet-label-bot [Pending Requester Info] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ChaiBapchya closed pull request #12749: [MXNET-1029] Feature request: randint operator
ChaiBapchya closed pull request #12749: [MXNET-1029] Feature request: randint operator URL: https://github.com/apache/incubator-mxnet/pull/12749 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake of provenance: As this is a foreign pull request (from a fork), the diff is supplied below (as it won't show otherwise due to GitHub magic): This is an automated message from the Apache Git Service. To respond to the message, please log on 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] lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so
lanking520 commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429099117 @sameer-kapoor Thanks for your question. Please provide some detailed information on the crash message. Have you tried to build from source or use the jar package on Maven? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ChaiBapchya commented on issue #12801: Inconsistent function signature of NDarray.randn
ChaiBapchya commented on issue #12801: Inconsistent function signature of NDarray.randn URL: https://github.com/apache/incubator-mxnet/issues/12801#issuecomment-429099168 Yup, I'm working on another PR that will have fix for this too. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] Vikas89 commented on issue #12800: getting segfault while running train_cifar10.py program in example directory
Vikas89 commented on issue #12800: getting segfault while running train_cifar10.py program in example directory URL: https://github.com/apache/incubator-mxnet/issues/12800#issuecomment-429097826 I think there was a fix made in relevant area of code: https://github.com/dmlc/dmlc-core/commit/e3377de933d3e069635150fb330ce6cc33b57950#diff-855ba648d1f4003608aa37ba3d060043 I will retry this on latest code and check if I can reproduce this again. I will keep this thread updated. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] zhanghang1989 commented on issue #12803: Fix Extremely Unbalanced Issue in split_and_load
zhanghang1989 commented on issue #12803: Fix Extremely Unbalanced Issue in split_and_load URL: https://github.com/apache/incubator-mxnet/pull/12803#issuecomment-429097682 FYI @eric-haibin-lin This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12724: simplify mac mkldnn build
azai91 commented on issue #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#issuecomment-429097572 Benchmarking with mobilenet is better since the architecture is smaller: with MKLDNN ``` INFO:root:batch size 1, dtype float32, images/sec: 107.649766 INFO:root:batch size 2, dtype float32, images/sec: 67.138190 INFO:root:batch size 4, dtype float32, images/sec: 115.700230 INFO:root:batch size 8, dtype float32, images/sec: 115.093671 INFO:root:batch size 16, dtype float32, images/sec: 116.750479 INFO:root:batch size 32, dtype float32, images/sec: 107.679933 ``` w/o ``` INFO:root:batch size 1, dtype float32, images/sec: 39.162575 INFO:root:batch size 2, dtype float32, images/sec: 41.462039 INFO:root:batch size 4, dtype float32, images/sec: 42.517774 INFO:root:batch size 8, dtype float32, images/sec: 42.316606 INFO:root:batch size 16, dtype float32, images/sec: 40.210146 INFO:root:batch size 32, dtype float32, images/sec: 38.261067 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12801: Inconsistent function signature of NDarray.randn
piyushghai commented on issue #12801: Inconsistent function signature of NDarray.randn URL: https://github.com/apache/incubator-mxnet/issues/12801#issuecomment-429097435 Thank you for reporting this issue @ChaiBapchya @mxnet-label-bot [NDArray] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] zhanghang1989 opened a new pull request #12803: Fix Extremely Unbalanced Issue in split_and_load
zhanghang1989 opened a new pull request #12803: Fix Extremely Unbalanced Issue in split_and_load URL: https://github.com/apache/incubator-mxnet/pull/12803 ## Description ## Existing implementation: ```python import mxnet as mx x = mx.nd.ones((38, 3, 4, 4)) ctx = [mx.gpu(i) for i in range(8)] xs = mx.gluon.utils.split_and_load(x, ctx, even_split=False) for xi in xs: print(xi.shape[0]) ``` The output is: ```bash 4 4 4 4 4 4 4 10 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so
piyushghai commented on issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802#issuecomment-429097144 Thank you for reporting this issue. We will look into it. @mxnet-label-bot [Scala, Breaking] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] azai91 commented on issue #12724: simplify mac mkldnn build
azai91 commented on issue #12724: simplify mac mkldnn build URL: https://github.com/apache/incubator-mxnet/pull/12724#issuecomment-429097045 @xinyu-intel I get the following when running on resnet-50 with MKLDNN ``` INFO:root:batch size 1, dtype float32, images/sec: 16.791932 INFO:root:batch size 2, dtype float32, images/sec: 16.436919 INFO:root:batch size 4, dtype float32, images/sec: 15.737164 INFO:root:batch size 8, dtype float32, images/sec: 15.317152 INFO:root:batch size 16, dtype float32, images/sec: 11.681419 INFO:root:batch size 32, dtype float32, images/sec: 14.112597 ``` vs without ``` INFO:root:batch size 1, dtype float32, images/sec: 12.369148 INFO:root:batch size 2, dtype float32, images/sec: 15.216185 INFO:root:batch size 4, dtype float32, images/sec: 14.709387 INFO:root:batch size 8, dtype float32, images/sec: 14.508584 INFO:root:batch size 16, dtype float32, images/sec: 13.949826 INFO:root:batch size 32, dtype float32, images/sec: 13.240351 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] vandanavk commented on issue #12783: Name conflict when serializing LSTMCell
vandanavk commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429097068 Similar error occurred in https://github.com/apache/incubator-mxnet/issues/11542 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #11542: Import to gluon not working for ONNX model
piyushghai commented on issue #11542: Import to gluon not working for ONNX model URL: https://github.com/apache/incubator-mxnet/issues/11542#issuecomment-429096687 Similar issues perhaps : #12795 , #12794 This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sameer-kapoor opened a new issue #12802: Crash while loading libmxnet-scala.so
sameer-kapoor opened a new issue #12802: Crash while loading libmxnet-scala.so URL: https://github.com/apache/incubator-mxnet/issues/12802 ## Description Random crashes while loading libmxnet-scala.so with both Mxnet 1.2.0 and 1.3.0 ## Environment info (Required) ``` Mxnet 1.3.0 CentOS Linux release 7.4.1708 JRE version: Java(TM) SE Runtime Environment (8.0_112-b15) (build 1.8.0_112-b15) ``` Package used: Scala For Scala user, please provide: 1. Java version: 1.8.0_112-b15 2. SBT version: 0.13 3. Scala runtime: 2.12 ## Build info (Required if built from source) Compiler (gcc/clang/mingw/visual studio): MXNet commit hash: b3be92f4a48bce62a5a8424271871c2f81c8f7f1 Build config: default config.mk ## Error Message: [2018-10-10 20:46:29,040] INFO - pool-1-thread-1 - NativeLibraryLoader - - Loading libmxnet-scala.so from /lib/native/ copying to mxnet-scala A fatal error has been detected by the Java Runtime Environment: SIGILL (0x4) at pc=0x7fe87774dff4, pid=4035, tid=0x7fe93039d700 JRE version: Java(TM) SE Runtime Environment (8.0_112-b15) (build 1.8.0_112-b15) Java VM: Java HotSpot(TM) 64-Bit Server VM (25.112-b15 mixed mode linux-amd64 compressed oops) Problematic frame: C [mxnet-scala+0x341aff4] mxnet::op::OperatorTune::Initialize()+0x2c4 Core dump written. Default location: /xyz/core or core.4035 An error report file with more information is saved as: /xyz/hs_err_pid4035.log ## Minimum reproducible example Random crash while loading libmxnet-scala.so This crash occurs while running Scala tests. During this time there are several times that libmxnet-scala.so is loaded and unloaded. ## Steps to reproduce (Paste the commands you ran that produced the error.) 1. 2. ## What have you tried to solve it? 1. Saw this in Mxnet 1.2.0 so tried upgrading to 1.3.0 but saw the same issue 2. Downgraded back to 1.1.0 and the issue disappears. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12800: getting segfault while running train_cifar10.py program in example directory
piyushghai commented on issue #12800: getting segfault while running train_cifar10.py program in example directory URL: https://github.com/apache/incubator-mxnet/issues/12800#issuecomment-429095935 Thanks for reporting this issue @Vikas89 @mxnet-label-bot [Example, Bug] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ankkhedia commented on issue #6783: mx.set.seed() does not work using GPU
ankkhedia commented on issue #6783: mx.set.seed() does not work using GPU URL: https://github.com/apache/incubator-mxnet/issues/6783#issuecomment-429095180 @nswamy @sandeep-krishnamurthy Could you please close the issue as the confusion has been resolved. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] szha commented on issue #12783: Name conflict when serializing LSTMCell
szha commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429092674 This problem exists in RNN and GRU as well, so all three needs to be patched. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] szha commented on issue #12783: Name conflict when serializing LSTMCell
szha commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429092412 For the problem with LSTM alone, the problem is in not naming some of the elementwise operations. In LSTM there are three names that are repeating: ``` "name": "myblock0_lstm0__plus0", "name": "myblock0_lstm0__mul0", "name": "myblock0_lstm0__mul1", ``` They come from the [plus0 here](https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/gluon/rnn/rnn_cell.py#L514), and the [mul0 and mul1 here](https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/gluon/rnn/rnn_cell.py#L524) The fix should be to replace these operations with F.elemwise_X, with the proper prefix just like other operators. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] lostella commented on issue #12783: Name conflict when serializing LSTMCell
lostella commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429091366 Unfortunately, that does not seem to solve the issue. See also the simpler example in my previous comment, which does not involve HybridSequentialRNNCell. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] danithaca commented on issue #6783: mx.set.seed() does not work using GPU
danithaca commented on issue #6783: mx.set.seed() does not work using GPU URL: https://github.com/apache/incubator-mxnet/issues/6783#issuecomment-429089214 Read the code again, and the results and explanation makes sense. Thanks @ankkhedia. @nswamy -- thanks for reopening the issue. I agree that the problem seems to be resolved and can be closed. Thank you both. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] szha edited a comment on issue #12783: Name conflict when serializing LSTMCell
szha edited a comment on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429086305 For the first piece of code, the problem is not using the container's name_scope. Since HybridSequentialRNNCell is a container block, you need to use its name_scope if you intend to properly export it as a symbol. ``` class MyBlock(mx.gluon.HybridBlock): def __init__(self): super().__init__() with self.name_scope(): self.lstm = mx.gluon.rnn.HybridSequentialRNNCell() with self.lstm.name_scope(): for layer in range(3): self.lstm.add(mx.gluon.rnn.LSTMCell(hidden_size=20)) def hybrid_forward(self, F, seq): outputs, state = self.lstm.unroll(inputs=seq, length=10, layout="NTC", merge_outputs=True) return outputs ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] szha commented on issue #12783: Name conflict when serializing LSTMCell
szha commented on issue #12783: Name conflict when serializing LSTMCell URL: https://github.com/apache/incubator-mxnet/issues/12783#issuecomment-429086305 The problem is in the sample code. Since HybridSequentialRNNCell is a container block, you need to use its name_scope if you intend to properly export it as a symbol. ``` class MyBlock(mx.gluon.HybridBlock): def __init__(self): super().__init__() with self.name_scope(): self.lstm = mx.gluon.rnn.HybridSequentialRNNCell() with self.lstm.name_scope(): for layer in range(3): self.lstm.add(mx.gluon.rnn.LSTMCell(hidden_size=20)) def hybrid_forward(self, F, seq): outputs, state = self.lstm.unroll(inputs=seq, length=10, layout="NTC", merge_outputs=True) return outputs ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] szha commented on issue #11585: Adds support for finding cudnn inside cuda path
szha commented on issue #11585: Adds support for finding cudnn inside cuda path URL: https://github.com/apache/incubator-mxnet/pull/11585#issuecomment-429084947 @lebeg is this something you can pick up? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on issue #12801: Inconsistent function signature of NDarray.randn
anirudhacharya commented on issue #12801: Inconsistent function signature of NDarray.randn URL: https://github.com/apache/incubator-mxnet/issues/12801#issuecomment-429083958 can you please add links and references in the above issue description This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ChaiBapchya opened a new issue #12801: Inconsistent function signature of NDarray.randn
ChaiBapchya opened a new issue #12801: Inconsistent function signature of NDarray.randn URL: https://github.com/apache/incubator-mxnet/issues/12801 ## Description NDArray.randn function signature `def rand (*shape, **kwargs):` is incorrect and inconsistent with rest of the signatures in Random Number Generator NDArray API a. Rest of the parameters - loc, scale, shape, dtype, ctx, out need to be included b. *shape has to be replaced by shape c. Why is another level of abstraction added by putting everything in **kwargs and then popping it up. If that's required, why not for others? This is an automated message from the Apache Git Service. To respond to the message, please log on 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 9c8d603 Bump the publish timestamp. 9c8d603 is described below commit 9c8d603f42a78420094a25aad0c618d4730401a1 Author: mxnet-ci AuthorDate: Thu Oct 11 18:57:36 2018 + 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..0dccade --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Thu Oct 11 18:57:36 UTC 2018
[GitHub] vandanavk commented on issue #12785: [MXNET-906] ONNX export/import: Selu
vandanavk commented on issue #12785: [MXNET-906] ONNX export/import: Selu URL: https://github.com/apache/incubator-mxnet/pull/12785#issuecomment-429074177 Adding @Roshrini @zhreshold for review This is an automated message from the Apache Git Service. To respond to the message, please log on 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] mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release
mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release URL: https://github.com/apache/incubator-mxnet/pull/12791#discussion_r224558542 ## File path: docker/docker-python/Dockerfile.mxnet.python3.cpu ## @@ -0,0 +1,28 @@ +# -*- mode: dockerfile -*- +# 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. +# +# Dockerfile to build MXNet for CPU + +FROM ubuntu:16.04 + +RUN apt-get update +RUN apt-get install -y wget python3 gcc +RUN wget https://bootstrap.pypa.io/get-pip.py +RUN python3 get-pip.py Review comment: @KellenSunderland @bhavinthaker Thank you, I agree. I will update the PR. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release
mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release URL: https://github.com/apache/incubator-mxnet/pull/12791#discussion_r224558239 ## File path: docker/docker-python/README.md ## @@ -39,3 +47,10 @@ Credentials can be provided in the following ways: * **Set Environment Variables:** Set the following environment variables which the script will pick up to login to dockerhub at runtime - * $MXNET_DOCKERHUB_PASSWORD * $MXNET_DOCKERHUB_USERNAME + + +### Using the Docker Images +* The MXNet Python Docker images can be found here: https://hub.docker.com/r/mxnet/python/ + +* Docker Pull Command: `docker pull mxnet/python:` Review comment: What details are you looking for? The tags can be looked up in the dockerhub link I have provided above - https://hub.docker.com/r/mxnet/python/tags/ Do you want me to add any other details to the README? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release
mbaijal commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release URL: https://github.com/apache/incubator-mxnet/pull/12791#discussion_r224557167 ## File path: docker/docker-python/Dockerfile.mxnet.python3.cpu ## @@ -0,0 +1,28 @@ +# -*- mode: dockerfile -*- +# 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. +# +# Dockerfile to build MXNet for CPU + +FROM ubuntu:16.04 + +RUN apt-get update +RUN apt-get install -y wget python3 gcc +RUN wget https://bootstrap.pypa.io/get-pip.py +RUN python3 get-pip.py + +RUN pip3 install mxnet Review comment: Do you mean parameterize the MXNet version in the dockerfiles? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] bhavinthaker commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release
bhavinthaker commented on a change in pull request #12791: [MXNET-1093] Add python3 Docker images for each MXNet release URL: https://github.com/apache/incubator-mxnet/pull/12791#discussion_r224554242 ## File path: docker/docker-python/Dockerfile.mxnet.python3.cpu ## @@ -0,0 +1,28 @@ +# -*- mode: dockerfile -*- +# 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. +# +# Dockerfile to build MXNet for CPU + +FROM ubuntu:16.04 + +RUN apt-get update +RUN apt-get install -y wget python3 gcc +RUN wget https://bootstrap.pypa.io/get-pip.py +RUN python3 get-pip.py Review comment: +1 to Kellen's ask: I think it is NOT a nit-pick but an important requirement because each RUN corresponds to a filesystem layer for Docker. See this: https://stackoverflow.com/questions/31222377/what-are-docker-image-layers This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ankkhedia commented on issue #12798: Fix Flaky Topk
ankkhedia commented on issue #12798: Fix Flaky Topk URL: https://github.com/apache/incubator-mxnet/pull/12798#issuecomment-429063399 Hi @sxjscience, request you to please look into the failing CI for the PR. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] marcoabreu commented on a change in pull request #12094: [MXNET-793] ★ Virtualized testing in CI with QEMU ★
marcoabreu commented on a change in pull request #12094: [MXNET-793] ★ Virtualized testing in CI with QEMU ★ URL: https://github.com/apache/incubator-mxnet/pull/12094#discussion_r224546252 ## File path: tests/nightly/apache_rat_license_check/rat-excludes ## @@ -57,3 +57,4 @@ apache-rat-tasks/* moderngpu/* deformable_im2col.cuh deformable_im2col.h +ci/qemu/.* Review comment: Yes, everything we publish in the repository needs a licence This is an automated message from the Apache Git Service. To respond to the message, please log on 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] aaronmarkham commented on issue #12689: Search update
aaronmarkham commented on issue #12689: Search update URL: https://github.com/apache/incubator-mxnet/pull/12689#issuecomment-429055202 @sandeep-krishnamurthy @safrooze I've re-enabled the preview. You should clear your cache to make sure you're getting the latest .htaccess file. This was pretty challenging to resolve and I hope I have the different angles covered with: https://github.com/apache/incubator-mxnet/pull/12689/files#diff-8752dbf83f90b837af0035f3c1291ecdR31 That should redirect traffic to the root to master. Requests to other versions go through. I added an env var to htaccess to hold the default version, which is currently 'master'. I kind of wish I wasn't combining the search fix with this kind of structure fix, but anyway, here we are. Please test it out. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] Vikas89 opened a new issue #12800: getting segfault while running train_cifar10.py program in example directory
Vikas89 opened a new issue #12800: getting segfault while running train_cifar10.py program in example directory URL: https://github.com/apache/incubator-mxnet/issues/12800 I am trying to run this command: python example/image-classification/train_cifar10.py And getting the segfault. It is very consistent. ``` INFO:root:Epoch[0] Batch [20]Speed: 18.95 samples/sec accuracy=0.164807 INFO:root:Epoch[0] Batch [40]Speed: 20.34 samples/sec accuracy=0.246875 INFO:root:Epoch[0] Batch [60]Speed: 21.66 samples/sec accuracy=0.308984 INFO:root:Epoch[0] Batch [80]Speed: 22.15 samples/sec accuracy=0.328125 INFO:root:Epoch[0] Batch [100] Speed: 22.38 samples/sec accuracy=0.362500 INFO:root:Epoch[0] Batch [120] Speed: 21.52 samples/sec accuracy=0.378125 INFO:root:Epoch[0] Batch [140] Speed: 22.96 samples/sec accuracy=0.417969 INFO:root:Epoch[0] Batch [160] Speed: 20.04 samples/sec accuracy=0.426563 INFO:root:Epoch[0] Batch [180] Speed: 18.58 samples/sec accuracy=0.430078 INFO:root:Epoch[0] Batch [200] Speed: 21.39 samples/sec accuracy=0.443750 INFO:root:Epoch[0] Batch [220] Speed: 17.69 samples/sec accuracy=0.469531 INFO:root:Epoch[0] Batch [240] Speed: 18.59 samples/sec accuracy=0.469141 INFO:root:Epoch[0] Batch [260] Speed: 21.68 samples/sec accuracy=0.470313 INFO:root:Epoch[0] Batch [280] Speed: 21.05 samples/sec accuracy=0.487891 INFO:root:Epoch[0] Batch [300] Speed: 22.30 samples/sec accuracy=0.503125 INFO:root:Epoch[0] Batch [320] Speed: 20.98 samples/sec accuracy=0.534766 INFO:root:Epoch[0] Batch [340] Speed: 15.53 samples/sec accuracy=0.526172 INFO:root:Epoch[0] Batch [360] Speed: 13.88 samples/sec accuracy=0.528516 Traceback (most recent call last): File "example/image-classification/train_cifar10.py", line 79, in fit.fit(args, sym, data.get_rec_iter) File "/Users/vikumar/incubator-mxnet/example/image-classification/common/fit.py", line 333, in fit monitor=monitor) File "/Users/vikumar/incubator-mxnet/python/mxnet/module/base_module.py", line 563, in fit next_data_batch = next(data_iter) File "/Users/vikumar/incubator-mxnet/python/mxnet/io/io.py", line 228, in __next__ return self.next() File "/Users/vikumar/incubator-mxnet/python/mxnet/io/io.py", line 840, in next check_call(_LIB.MXDataIterNext(self.handle, ctypes.byref(next_res))) File "/Users/vikumar/incubator-mxnet/python/mxnet/base.py", line 252, in check_call raise MXNetError(py_str(_LIB.MXGetLastError())) mxnet.base.MXNetError: [09:58:31] src/io/recordio_split.cc:29: Check failed: (reinterpret_cast(end) & 3UL) == 0U (1 vs. 0) Stack trace returned 10 entries: [bt] (0) 0 libmxnet.so 0x00010ae2cf30 dmlc::StackTrace() + 272 [bt] (1) 1 libmxnet.so 0x00010ae2ccdf dmlc::LogMessageFatal::~LogMessageFatal() + 47 [bt] (2) 2 libmxnet.so 0x00010c68a01c dmlc::io::RecordIOSplitter::FindLastRecordBegin(char const*, char const*) + 444 [bt] (3) 3 libmxnet.so 0x00010c68d10c dmlc::io::InputSplitBase::ReadChunk(void*, unsigned long*) + 284 [bt] (4) 4 libmxnet.so 0x00010c68d230 dmlc::io::InputSplitBase::Chunk::Load(dmlc::io::InputSplitBase*, unsigned long) + 144 [bt] (5) 5 libmxnet.so 0x00010c6a0b8f dmlc::ThreadedIter::Init(std::__1::function, std::__1::function)::'lambda'()::operator()() const + 895 [bt] (6) 6 libmxnet.so 0x00010c6a071d void* std::__1::__thread_proxy >, dmlc::ThreadedIter::Init(std::__1::function, std::__1::function)::'lambda'()> >(void*) + 45 [bt] (7) 7 libsystem_pthread.dylib 0x7fff903c293b _pthread_body + 180 [bt] (8) 8 libsystem_pthread.dylib 0x7fff903c2887 _pthread_body + 0 [bt] (9) 9 libsystem_pthread.dylib 0x7fff903c208d thread_start + 13 ``` ## Environment info build command - make -j8 USE_DIST_KVSTORE=1 os: mac ``` What to do: 1. o/p of diagnose script: 88e9fe53272d:incubator-mxnet vikumar$ python /tmp/lk.py --Python Info-- Version : 3.6.5 Compiler : GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final) Build: ('default', 'Apr 26 2018 08:42:37') Arch : ('64bit', '') Pip Info--- Version : 10.0.1 Directory: /Users/vikumar/anaconda3/lib/python3.6/site-packages/pip --MXNet Info--- /Users/vikumar/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning:
[GitHub] marcoabreu commented on issue #12796: Add embedding to print_summary
marcoabreu commented on issue #12796: Add embedding to print_summary URL: https://github.com/apache/incubator-mxnet/pull/12796#issuecomment-429051934 Am I catching it wrong or is test_print_summary() actually not doing any validation? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] soeque1 edited a comment on issue #12796: Add embedding to print_summary
soeque1 edited a comment on issue #12796: Add embedding to print_summary URL: https://github.com/apache/incubator-mxnet/pull/12796#issuecomment-429038277 The Post (#12778) does not have enough information to run the code. So I added something to the sample code. ``` from mxnet import gluon from mxnet.gluon import HybridBlock, nn import mxnet as mx vocab_size ``` ``` from mxnet import gluon from mxnet.gluon import HybridBlock, nn import mxnet as mx class SentClassificationModel(gluon.HybridBlock): def __init__(self, vocab_size, num_embed, **kwargs): super(SentClassificationModel, self).__init__(**kwargs) with self.name_scope(): self.embed = nn.Embedding(input_dim=vocab_size, output_dim=num_embed) self.drop = nn.Dropout(0.3) self.fc = nn.Dense(100, activation='relu') self.out = nn.Dense(2) def hybrid_forward(self, F ,inputs): em_out = self.drop(self.embed(inputs)) fc_out = self.fc(em_out) return(self.out(fc_out)) ctx = mx.gpu() model = SentClassificationModel(vocab_size=20, num_embed=50) model.initialize(mx.init.Xavier(),ctx=ctx) model.hybridize() mx.viz.print_summary( model(mx.sym.var('data')), shape={'data':(1,30)}, #set your shape here ) ``` The output is the below ``` Layer (type)Output Shape Param # Previous Layer data(null) 30 0 sentclassificationmodel2_embedding0_fwd(Embedding) 30x50 1000data sentclassificationmodel2_dropout0_fwd(Dropout) 30x50 0 sentclassificationmodel2_embeddi sentclassificationmodel2_dense0_fwd(FullyConnected) 100 3100sentclassificationmodel2_dropout sentclassificationmodel2_dense0_relu_fwd(Activation)100 0 sentclassificationmodel2_dense0_ sentclassificationmodel2_dense1_fwd(FullyConnected) 2 202 sentclassificationmodel2_dense0_ Total params: 4302 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] ArmageddonKnight commented on a change in pull request #11364: [MXNET-490] Added OpenLSTMRNN together with benchmarks and Tensorboard callback routines.
ArmageddonKnight commented on a change in pull request #11364: [MXNET-490] Added OpenLSTMRNN together with benchmarks and Tensorboard callback routines. URL: https://github.com/apache/incubator-mxnet/pull/11364#discussion_r224534555 ## File path: src/operator/contrib/cu_open_lstm_rnn-inl.cuh ## @@ -0,0 +1,907 @@ +/* + * 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. + */ +/*! + * Copyright (c) 2018 by Contributors + * \file open_lstm_rnn-inl.cuh + * \brief LSTM RNN Open-Source CUDA Implementation + * \author Bojian (Jack) Zheng, Gennady Pekhimenko, Jeremy Appleyard + */ +#ifndef MXNET_OPERATOR_CONTRIB_CU_OPEN_LSTM_RNN_INL_CUH_ Review comment: @apeforest The reason why I start this file with `cu` is because `cu` is considered as the abbreviation for **CUDA** implementation (e.g., cuDNN, cuBLAS, cuSparse). This is an automated message from the Apache Git Service. To respond to the message, please log on 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] vandanavk commented on a change in pull request #12785: [MXNET-906] ONNX export/import: Selu
vandanavk commented on a change in pull request #12785: [MXNET-906] ONNX export/import: Selu URL: https://github.com/apache/incubator-mxnet/pull/12785#discussion_r224532125 ## File path: tests/python-pytest/onnx/import/test_cases.py ## @@ -98,6 +98,7 @@ 'test_MaxPool', 'test_PReLU', 'test_ReLU', +'test_selu_default', Review comment: update based on offline discussions - MXNet's implementation fixed these values based on the recommendation in the paper, so alpha and lambda will continue to be the default values and not configurable parameters to LeakyRelu. following this, only test_selu_default will be added as a test for ONNX import/export. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] soeque1 commented on issue #12796: Add embedding to print_summary
soeque1 commented on issue #12796: Add embedding to print_summary URL: https://github.com/apache/incubator-mxnet/pull/12796#issuecomment-429038277 The Post (#12778) does not have enough information to run the code. So I added something to the sample code. ``` from mxnet import gluon from mxnet.gluon import HybridBlock, nn import mxnet as mx vocab_size ``` ``` from mxnet import gluon from mxnet.gluon import HybridBlock, nn import mxnet as mx class SentClassificationModel(gluon.HybridBlock): def __init__(self, vocab_size, num_embed, **kwargs): super(SentClassificationModel, self).__init__(**kwargs) with self.name_scope(): self.embed = nn.Embedding(input_dim=vocab_size, output_dim=num_embed) self.drop = nn.Dropout(0.3) self.fc = nn.Dense(100, activation='relu') self.out = nn.Dense(2) def hybrid_forward(self, F ,inputs): em_out = self.drop(self.embed(inputs)) fc_out = self.fc(em_out) return(self.out(fc_out)) ctx = mx.gpu() model = SentClassificationModel(vocab_size = 20, num_embed=50) model.initialize(mx.init.Xavier(),ctx=ctx) model.hybridize() mx.viz.print_summary( model(mx.sym.var('data')), shape={'data':(1,30)}, #set your shape here ) ``` The output is the below ``` Layer (type)Output Shape Param # Previous Layer data(null) 30 0 sentclassificationmodel2_embedding0_fwd(Embedding) 30x50 1000data sentclassificationmodel2_dropout0_fwd(Dropout) 30x50 0 sentclassificationmodel2_embeddi sentclassificationmodel2_dense0_fwd(FullyConnected) 100 3100sentclassificationmodel2_dropout sentclassificationmodel2_dense0_relu_fwd(Activation)100 0 sentclassificationmodel2_dense0_ sentclassificationmodel2_dense1_fwd(FullyConnected) 2 202 sentclassificationmodel2_dense0_ Total params: 4302 ``` This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12794: fixed symbols naming in RNNCell and LSTMCell
piyushghai commented on issue #12794: fixed symbols naming in RNNCell and LSTMCell URL: https://github.com/apache/incubator-mxnet/pull/12794#issuecomment-429036417 @lostella You are right. The issue is with the ```-``` and ```/``` as well. Just tried the Dense Layer example as well. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] larroy commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu)
larroy commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu) URL: https://github.com/apache/incubator-mxnet/issues/12799#issuecomment-429034160 @piyushghai is not flaky. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] larroy commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu)
larroy commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu) URL: https://github.com/apache/incubator-mxnet/issues/12799#issuecomment-429034472 @mxnet-label-bot, please add labels :[ARM, NDArray, Bug] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] piyushghai commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu)
piyushghai commented on issue #12799: ERROR: test_ndarray.test_order fails on ARMv7 (Qemu) URL: https://github.com/apache/incubator-mxnet/issues/12799#issuecomment-429033855 @mxnet-label-bot [Flaky] This is an automated message from the Apache Git Service. To respond to the message, please log on 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] anirudhacharya commented on a change in pull request #12748: [MXNET-1028] Adding CELU Activation function
anirudhacharya commented on a change in pull request #12748: [MXNET-1028] Adding CELU Activation function URL: https://github.com/apache/incubator-mxnet/pull/12748#discussion_r224525239 ## File path: python/mxnet/gluon/nn/activations.py ## @@ -181,6 +181,25 @@ def hybrid_forward(self, F, x): return F.LeakyReLU(x, act_type='selu', name='fwd') +class CELU(HybridBlock): +r""" +Continuous Exponential Linear Unit (CELU) +https://arxiv.org/abs/1704.07483 Review comment: i will do that. But I will do it more on these lines - http://mxnet.incubator.apache.org/api/python/gluon/nn.html#mxnet.gluon.nn.LeakyReLU as this a gluon block rather than an ndarray function. This is an automated message from the Apache Git Service. To respond to the message, please log on 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] marcoabreu commented on issue #12798: Fix Flaky Topk
marcoabreu commented on issue #12798: Fix Flaky Topk URL: https://github.com/apache/incubator-mxnet/pull/12798#issuecomment-429033493 Thanks for the fix! Could you submit a test case that would have failed previously? Also, do we have any performance benchmarks for this kernel change? This is an automated message from the Apache Git Service. To respond to the message, please log on 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] sandeep-krishnamurthy commented on issue #12794: fixed symbols naming in RNNCell and LSTMCell
sandeep-krishnamurthy commented on issue #12794: fixed symbols naming in RNNCell and LSTMCell URL: https://github.com/apache/incubator-mxnet/pull/12794#issuecomment-429033286 @piyushghai @szha This is an automated message from the Apache Git Service. To respond to the message, please log on 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