[GitHub] marcoabreu commented on issue #9845: Flaky test_operator.test_reduce
marcoabreu commented on issue #9845: Flaky test_operator.test_reduce URL: https://github.com/apache/incubator-mxnet/issues/9845#issuecomment-379982790 == FAIL: test_operator.test_reduce -- Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest self.test(*self.arg) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/common.py", line 157, in test_new orig_test(*args, **kwargs) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1826, in test_reduce mx.symbol.max, test_none_axis=test_none) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1799, in test_reduce_inner assert equal_backward AssertionError: >> begin captured logging << common: INFO: Setting module np/mx/python random seeds, use MXNET_MODULE_SEED=754316987 to reproduce. common: INFO: Setting test np/mx/python random seeds, use MXNET_TEST_SEED=1117986172 to reproduce. - >> end captured logging << - -- Ran 1 test in 22.731s FAILED (failures=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] marcoabreu commented on issue #9845: Flaky test_operator.test_reduce
marcoabreu commented on issue #9845: Flaky test_operator.test_reduce URL: https://github.com/apache/incubator-mxnet/issues/9845#issuecomment-379982790 ``` == FAIL: test_operator.test_reduce -- Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest self.test(*self.arg) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/common.py", line 157, in test_new orig_test(*args, **kwargs) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1826, in test_reduce mx.symbol.max, test_none_axis=test_none) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1799, in test_reduce_inner assert equal_backward AssertionError: >> begin captured logging << common: INFO: Setting module np/mx/python random seeds, use MXNET_MODULE_SEED=754316987 to reproduce. common: INFO: Setting test np/mx/python random seeds, use MXNET_TEST_SEED=1117986172 to reproduce. - >> end captured logging << - -- Ran 1 test in 22.731s FAILED (failures=1) ``` from https://github.com/apache/incubator-mxnet/issues/10476 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 #10476: Flaky test for test_operator.test_reduce
marcoabreu commented on issue #10476: Flaky test for test_operator.test_reduce URL: https://github.com/apache/incubator-mxnet/issues/10476#issuecomment-379982768 Duplicate of https://github.com/apache/incubator-mxnet/issues/9845 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 closed issue #10476: Flaky test for test_operator.test_reduce
marcoabreu closed issue #10476: Flaky test for test_operator.test_reduce URL: https://github.com/apache/incubator-mxnet/issues/10476 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 closed pull request #10013: [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE
eric-haibin-lin closed pull request #10013: [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE URL: https://github.com/apache/incubator-mxnet/pull/10013 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/tutorials/index.md b/docs/tutorials/index.md index 917afe92837..00a15046ab9 100644 --- a/docs/tutorials/index.md +++ b/docs/tutorials/index.md @@ -2,10 +2,10 @@ MXNet has two primary high-level interfaces for its deep learning engine: the Gluon API and the Module API. Tutorials for each are provided below. -The difference between the two is an imperative versus symbolic programming style. Gluon makes it easy to prototype, build, and train deep learning models without sacrificing training speed by enabling both (1) intuitive imperative Python code development and (2) faster execution by automatically generating a symbolic execution graph using the hybridization feature. - `TL;DR:` If you are new to deep learning or MXNet, you should start with the Gluon tutorials. +The difference between the two is an imperative versus symbolic programming style. Gluon makes it easy to prototype, build, and train deep learning models without sacrificing training speed by enabling both (1) intuitive imperative Python code development and (2) faster execution by automatically generating a symbolic execution graph using the hybridization feature. + The Gluon and Module tutorials are in Python, but you can also find a variety of other MXNet tutorials, such as R, Scala, and C++ in the [Other Languages API Tutorials](#other-mxnet-api-tutorials) section below. [Example scripts and applications](#example-scripts-and-applications) as well as [contribution](#contributing-tutorials) info is below. @@ -246,7 +246,8 @@ The Gluon and Module tutorials are in Python, but you can also find a variety of - [Setup your MXNet with Scala on IntelliJ](/tutorials/scala/mxnet_scala_on_intellij.html) - [MNIST with the Scala API](/tutorials/scala/mnist.html) - [Use Scala to build a Long Short-Term Memory network that generates Barack Obama's speech patterns](/tutorials/scala/char_lstm.html) - + + diff --git a/docs/tutorials/scala/README.md b/docs/tutorials/scala/README.md new file mode 100644 index 000..0081f9c8216 --- /dev/null +++ b/docs/tutorials/scala/README.md @@ -0,0 +1,7 @@ +# MXNet-Scala Tutorials + +* [Setup Scala with MXNet and Create a MXNet-Scala Project with IntelliJ](mxnet_scala_on_intellij.md) +* [MNIST with MXNet-Scala](mnist.md) +* [Character-level Language Model with MXNet-Scala](char_lstm.md) + +See the [tutorials](http://mxnet.incubator.apache.org/tutorials/index.html#other-languages-api-tutorials) page on MXNet.io for more. diff --git a/docs/tutorials/scala/mxnet_scala_on_intellij.md b/docs/tutorials/scala/mxnet_scala_on_intellij.md index 037165fa0d7..8a14767d56b 100644 --- a/docs/tutorials/scala/mxnet_scala_on_intellij.md +++ b/docs/tutorials/scala/mxnet_scala_on_intellij.md @@ -1,62 +1,295 @@ # Run MXNet Scala Examples Using the IntelliJ IDE + + This tutorial guides you through setting up a Scala project in the IntelliJ IDE and shows how to use an MXNet package from your application. ## Prerequisites: -To use this tutorial, you need: +To use this tutorial you need the following items, however after this list, installation info for macOS is provided for your benefit: + +- [Java 8 JDK](http://www.oracle.com/technetwork/java/javase/downloads/index.html) +- [Maven](https://maven.apache.org/install.html) +- [Scala](https://www.scala-lang.org/download/) - comes with IntelliJ, so you don't need to install it separately +- [MXNet Shared Library and Scala Package](#build-the-mxnet-shared-library-and-scala-package) +- [IntelliJ IDE](https://www.jetbrains.com/idea/) + +## Mac Prerequisites Setup + +For other operating systems, visit each Prerequisite's website and follow their installations instructions. For macOS, you're in luck: + +1. Install brew: +``` +/usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)" +``` + +2. Install Java 8 JDK: +``` +brew tap caskroom/versions +brew cask install java8 +``` + +3. Install maven: +``` +brew update +brew install maven +``` + +## Build the MXNet Shared Library and Scala Package + +This depends on your operating system. Instructions for macOS, Ubuntu, and Windows are provided: + + +OS | Step 1 | Step 2 +---|---|--- +macOS | [Shared Library for macOS](http://mxnet.incubator.apache.org/install/osx_setup.html#build-the-shared-library) | [Scala Package for macOS](http://mxnet.incubator.apache.org/install/osx_setup.html#install-the-mxnet-package-for-scala) +Ubuntu | [Shared Library for
[incubator-mxnet] branch master updated: [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE (#10013)
This is an automated email from the ASF dual-hosted git repository. haibin pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 76f67b1 [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE (#10013) 76f67b1 is described below commit 76f67b1eba8261ae5e703453f15adf56fefe5b73 Author: Aaron MarkhamAuthorDate: Mon Apr 9 22:49:09 2018 -0700 [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE (#10013) * initial update on setting up scala ide with mxnet * moving images to web-data project * updated links to images; added readme for root folder * scala hello world feature added * workaround for make transitive error * fixed systempath * minor updates --- docs/tutorials/index.md | 7 +- docs/tutorials/scala/README.md | 7 + docs/tutorials/scala/mxnet_scala_on_intellij.md | 301 +--- 3 files changed, 278 insertions(+), 37 deletions(-) diff --git a/docs/tutorials/index.md b/docs/tutorials/index.md index 917afe9..00a1504 100644 --- a/docs/tutorials/index.md +++ b/docs/tutorials/index.md @@ -2,10 +2,10 @@ MXNet has two primary high-level interfaces for its deep learning engine: the Gluon API and the Module API. Tutorials for each are provided below. -The difference between the two is an imperative versus symbolic programming style. Gluon makes it easy to prototype, build, and train deep learning models without sacrificing training speed by enabling both (1) intuitive imperative Python code development and (2) faster execution by automatically generating a symbolic execution graph using the hybridization feature. - `TL;DR:` If you are new to deep learning or MXNet, you should start with the Gluon tutorials. +The difference between the two is an imperative versus symbolic programming style. Gluon makes it easy to prototype, build, and train deep learning models without sacrificing training speed by enabling both (1) intuitive imperative Python code development and (2) faster execution by automatically generating a symbolic execution graph using the hybridization feature. + The Gluon and Module tutorials are in Python, but you can also find a variety of other MXNet tutorials, such as R, Scala, and C++ in the [Other Languages API Tutorials](#other-mxnet-api-tutorials) section below. [Example scripts and applications](#example-scripts-and-applications) as well as [contribution](#contributing-tutorials) info is below. @@ -246,7 +246,8 @@ The Gluon and Module tutorials are in Python, but you can also find a variety of - [Setup your MXNet with Scala on IntelliJ](/tutorials/scala/mxnet_scala_on_intellij.html) - [MNIST with the Scala API](/tutorials/scala/mnist.html) - [Use Scala to build a Long Short-Term Memory network that generates Barack Obama's speech patterns](/tutorials/scala/char_lstm.html) - + + diff --git a/docs/tutorials/scala/README.md b/docs/tutorials/scala/README.md new file mode 100644 index 000..0081f9c --- /dev/null +++ b/docs/tutorials/scala/README.md @@ -0,0 +1,7 @@ +# MXNet-Scala Tutorials + +* [Setup Scala with MXNet and Create a MXNet-Scala Project with IntelliJ](mxnet_scala_on_intellij.md) +* [MNIST with MXNet-Scala](mnist.md) +* [Character-level Language Model with MXNet-Scala](char_lstm.md) + +See the [tutorials](http://mxnet.incubator.apache.org/tutorials/index.html#other-languages-api-tutorials) page on MXNet.io for more. diff --git a/docs/tutorials/scala/mxnet_scala_on_intellij.md b/docs/tutorials/scala/mxnet_scala_on_intellij.md index 037165f..8a14767 100644 --- a/docs/tutorials/scala/mxnet_scala_on_intellij.md +++ b/docs/tutorials/scala/mxnet_scala_on_intellij.md @@ -1,62 +1,295 @@ # Run MXNet Scala Examples Using the IntelliJ IDE + + This tutorial guides you through setting up a Scala project in the IntelliJ IDE and shows how to use an MXNet package from your application. ## Prerequisites: -To use this tutorial, you need: +To use this tutorial you need the following items, however after this list, installation info for macOS is provided for your benefit: + +- [Java 8 JDK](http://www.oracle.com/technetwork/java/javase/downloads/index.html) +- [Maven](https://maven.apache.org/install.html) +- [Scala](https://www.scala-lang.org/download/) - comes with IntelliJ, so you don't need to install it separately +- [MXNet Shared Library and Scala Package](#build-the-mxnet-shared-library-and-scala-package) +- [IntelliJ IDE](https://www.jetbrains.com/idea/) + +## Mac Prerequisites Setup + +For other operating systems, visit each Prerequisite's website and follow their installations instructions. For macOS, you're in luck: + +1. Install brew: +``` +/usr/bin/ruby -e "$(curl -fsSL
[GitHub] eric-haibin-lin closed issue #9973: update Scala tutorial - using IntelliJ IDE
eric-haibin-lin closed issue #9973: update Scala tutorial - using IntelliJ IDE URL: https://github.com/apache/incubator-mxnet/issues/9973 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] haojin2 opened a new pull request #10479: [MXNET-301] Setting a fixed seed for module tutorial
haojin2 opened a new pull request #10479: [MXNET-301] Setting a fixed seed for module tutorial URL: https://github.com/apache/incubator-mxnet/pull/10479 ## Description ## Setting a fix seed in the tutorial for module to ensure consistent behavior ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [x] The PR title starts with [MXNET-301] - [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 ### - [x] Set a fixed seed in tutorial for module API ## Comments ## With this fixed seed, the Jenkins tutorial passed. 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] anirudh2290 commented on issue #10013: [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE
anirudh2290 commented on issue #10013: [MXNET-48] update on setting up Scala with MXNet and the IntelliJ IDE URL: https://github.com/apache/incubator-mxnet/pull/10013#issuecomment-379977145 @eric-haibin-lin @nswamy is this good to 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] anirudh2290 commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn
anirudh2290 commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-379977059 @piiswrong is this good to 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] eric-haibin-lin commented on a change in pull request #10374: [MXNET-93] Sparse support for Custom Op
eric-haibin-lin commented on a change in pull request #10374: [MXNET-93] Sparse support for Custom Op URL: https://github.com/apache/incubator-mxnet/pull/10374#discussion_r180299395 ## File path: python/mxnet/operator.py ## @@ -522,6 +526,69 @@ def infer_type(self, in_type): return in_type, [in_type[0]]*len(self.list_outputs()), \ [in_type[0]]*len(self.list_auxiliary_states()) +def infer_storage_type(self, in_stype): +"""infer_storage_type interface. Used to infer storage type of +inputs and outputs in the forward pass. + +Parameters +-- +in_stype : list of stypes, Valid stypes are default, row_sparse and +csr + +Returns +--- +in_stype : list +list of argument stypes. +out_stype : list +list of output types calculated from in_stype, +in the same order as declared in list_outputs. +aux_type : Optional, list +list of aux types calculated from in_stype, +in the same order as declared in list_auxiliary_states. +""" +return in_stype, [in_stype[0]]*len(self.list_outputs()), \ +[in_stype[0]]*len(self.list_auxiliary_states()) + +def infer_storage_type_backward(self, ograd_stype, in_stype, out_stype, igrad_stype, aux_stype): +"""infer_storage_type_backward interface. Used to infer storage +type of inputs and outputs in the backward pass. + +Will raise an error if undefined storage type is returned. +Returned lists have to be the same size as the input lists to infer_storage_type_backward, +otherwise an exception will be thrown. When this interface is not implemented, +all stypes will fallback to default. Review comment: I will be careful about using the word "fallback" here since it has specific meaning for sparse ops. I'm a little bit confused about the default behavior of forward stype inference vs. backward stype inference. In forward you replicated the stype of in_stype to outputs. In backward you replicated the "default" stype to all 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] eric-haibin-lin commented on a change in pull request #10374: [MXNET-93] Sparse support for Custom Op
eric-haibin-lin commented on a change in pull request #10374: [MXNET-93] Sparse support for Custom Op URL: https://github.com/apache/incubator-mxnet/pull/10374#discussion_r180298634 ## File path: python/mxnet/operator.py ## @@ -522,6 +526,69 @@ def infer_type(self, in_type): return in_type, [in_type[0]]*len(self.list_outputs()), \ [in_type[0]]*len(self.list_auxiliary_states()) +def infer_storage_type(self, in_stype): +"""infer_storage_type interface. Used to infer storage type of +inputs and outputs in the forward pass. + +Parameters +-- +in_stype : list of stypes, Valid stypes are default, row_sparse and Review comment: Valid -> valid 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] anirudh2290 commented on issue #10415: [MXNET-284] (WIP) Update release version text to 1.2.0
anirudh2290 commented on issue #10415: [MXNET-284] (WIP) Update release version text to 1.2.0 URL: https://github.com/apache/incubator-mxnet/pull/10415#issuecomment-379976395 is this still wip ? 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] zhreshold opened a new pull request #10478: update imagenet training script links
zhreshold opened a new pull request #10478: update imagenet training script links URL: https://github.com/apache/incubator-mxnet/pull/10478 ## Description ## Update gluon model_zoo links to training scrips. The old ones are pointing to forked repo. ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) created (except PRs with tiny changes) - [ ] Changes are complete (i.e. I finished coding on this PR) - [ ] All changes have test coverage: - Unit tests are added for small changes to verify correctness (e.g. adding a new operator) - Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore) - Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL) - [ ] Code is well-documented: - For user-facing API changes, API doc string has been updated. - For new C++ functions in header files, their functionalities and arguments are documented. - For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable - Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html - [ ] 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 ## - If this change is a backward incompatible change, why must this change be made. - Interesting edge cases to note 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] eric-haibin-lin commented on issue #10183: [MXNET-120] Float16 support for distributed training
eric-haibin-lin commented on issue #10183: [MXNET-120] Float16 support for distributed training URL: https://github.com/apache/incubator-mxnet/pull/10183#issuecomment-379973769 BTW build is failing 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 a change in pull request #10183: [MXNET-120] Float16 support for distributed training
eric-haibin-lin commented on a change in pull request #10183: [MXNET-120] Float16 support for distributed training URL: https://github.com/apache/incubator-mxnet/pull/10183#discussion_r180296527 ## File path: amalgamation/Makefile ## @@ -23,6 +23,11 @@ ifndef OPENBLAS_ROOT export OPENBLAS_ROOT=/usr/local/opt/openblas endif +# use F16C if the architecture supports it, turned off by default +ifndef USE_F16C Review comment: what does 16C stand for? 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] anirudh2290 commented on issue #10449: [MXNET-293] Navigation updates
anirudh2290 commented on issue #10449: [MXNET-293] Navigation updates URL: https://github.com/apache/incubator-mxnet/pull/10449#issuecomment-379973552 @piiswrong can we merge 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] eric-haibin-lin commented on a change in pull request #10390: Initial PR for RAT License Check
eric-haibin-lin commented on a change in pull request #10390: Initial PR for RAT License Check URL: https://github.com/apache/incubator-mxnet/pull/10390#discussion_r180295347 ## File path: tests/nightly/apache_rat_license_check/.rat-excludes ## @@ -0,0 +1,54 @@ +.*xml +\..* +.*css +\\.* +.*ipynb +.*html +.*json +.*js +.*txt +.*md +3rdparty/* +R-package/* +src/operator/mkl/* +trunk/* +docker/* +docker_multiarch/* +.*\\.m +.*\\.mk +.*\\.R +contrib/* +Dockerfile* +.*svg +.*cfg +.*config +docs/* +__init__.py +build/* +.*\\.t +perl-package/* Review comment: @sergeykolychev any comment about adding apache license to these files? 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 a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU
eric-haibin-lin commented on a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU URL: https://github.com/apache/incubator-mxnet/pull/10371#discussion_r180295176 ## File path: src/operator/tensor/dot-inl.h ## @@ -216,32 +223,60 @@ inline bool DotForwardInferStorageType(const nnvm::NodeAttrs& attrs, bool only_lhs_transpose = param.transpose_a && !param.transpose_b; bool rhs_rsp_or_dns = rhs_stype == kRowSparseStorage || rhs_stype == kDefaultStorage; + NDArrayStorageType target_stype; if (!dispatched && lhs_stype == kDefaultStorage && rhs_stype == kDefaultStorage) { // dns, dns -> dns -dispatched = storage_type_assign(_stype, kDefaultStorage, dispatch_mode, - DispatchMode::kFCompute); +target_stype = (param.forward_stype_hint.has_value())? Review comment: nit: This line is quite long. Maybe cache the value of `param.forward_stype_hint.has_value()`? 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 a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU
eric-haibin-lin commented on a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU URL: https://github.com/apache/incubator-mxnet/pull/10371#discussion_r180294411 ## File path: src/operator/tensor/dot.cc ## @@ -51,12 +150,16 @@ NNVM_REGISTER_OP(dot) dot(x,y)[0,0,1,1] = 0 sum(x[0,0,:]*y[:,1,1]) = 0 -The storage type of ``dot`` output depends on storage types of inputs and transpose options: +The storage type of ``dot`` output depends on storage types of inputs, transpose options and given +hint for output storage type: +Implemented sprase operations include: - dot(csr, default) = default - dot(csr.T, default) = row_sparse - dot(csr, row_sparse) = default -- dot(default, csr) = csr +- dot(default, csr) = csr on CPU only +- dot(default, csr) = dense on GPU only Review comment: Could you show the output storage with specific values of `forward_stype_hint`? e.g. dot(csr.T, dense) = row_sparse dot(csr.T, dense, forward_stype_hint='default') = default dot(default, csr, forward_stype_hint='default') = default (GPU only) What happens if someone uses dot(dense, dense, forward_stype_hint='csr')? 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 a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU
eric-haibin-lin commented on a change in pull request #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU URL: https://github.com/apache/incubator-mxnet/pull/10371#discussion_r180294853 ## File path: src/operator/tensor/dot-inl.h ## @@ -45,13 +45,20 @@ namespace op { struct DotParam : public dmlc::Parameter { bool transpose_a; bool transpose_b; + dmlc::optional forward_stype_hint; DMLC_DECLARE_PARAMETER(DotParam) { DMLC_DECLARE_FIELD(transpose_a) .describe("If true then transpose the first input before dot.") .set_default(false); DMLC_DECLARE_FIELD(transpose_b) .describe("If true then transpose the second input before dot.") .set_default(false); +DMLC_DECLARE_FIELD(forward_stype_hint) + .describe("Desired storage type of the forward output.") Review comment: Need to be more detailed than this. Explicitly state that if no such combination is implemented, dense op is used. 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] spidyDev commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix
spidyDev commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix URL: https://github.com/apache/incubator-mxnet/pull/10425#discussion_r180290804 ## File path: python/mxnet/io.py ## @@ -645,8 +647,8 @@ def __init__(self, data, label=None, batch_size=1, shuffle=False, label_name='softmax_label'): super(NDArrayIter, self).__init__(batch_size) -self.data = _init_data(data, allow_empty=False, default_name=data_name) -self.label = _init_data(label, allow_empty=True, default_name=label_name) +self.data, self.renamedData = _init_data(data, allow_empty=False, default_name=data_name) Review comment: done ! 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] spidyDev commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix
spidyDev commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix URL: https://github.com/apache/incubator-mxnet/pull/10425#discussion_r180290853 ## File path: python/mxnet/module/base_module.py ## @@ -363,6 +363,19 @@ def predict(self, eval_data, num_batch=None, merge_batches=True, reset=True, eval_data.reset() output_list = [] +# If data provided is a dict. This makes sure that only relevant data items Review comment: Yes, thanks for pointing this out. Added to score() 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] eric-haibin-lin commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU
eric-haibin-lin commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns and dot(dns, csr.T) = dns on GPU URL: https://github.com/apache/incubator-mxnet/pull/10371#issuecomment-379965625 Please fix test ``` test_operator_gpu.test_sparse_dot ... /work/runtime_functions.sh: line 389: 7 Segmentation fault (core dumped) nosetests-2.7 --verbose tests/python/gpu ``` 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] ashokei commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn
ashokei commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-379965221 @anirudh2290 done...thanks. 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] asitstands commented on issue #10468: [MXNET-300] Fix ndarray assignment bug with index -1
asitstands commented on issue #10468: [MXNET-300] Fix ndarray assignment bug with index -1 URL: https://github.com/apache/incubator-mxnet/pull/10468#issuecomment-379964853 @reminisce I added some cases covering this PR (negative index in basic indexing) and removed two duplicated cases. 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] ZhichengHuang commented on issue #5812: Guided BackPropagation for visualization
ZhichengHuang commented on issue #5812: Guided BackPropagation for visualization URL: https://github.com/apache/incubator-mxnet/issues/5812#issuecomment-379960515 @bhokaal2k can you share your implementation of Guided BackPropagation, I have met the same problem, and can't find a proper solution. 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] kpmurali opened a new pull request #10477: [MXNET-299] Adding the Download page and linking it to the Install page
kpmurali opened a new pull request #10477: [MXNET-299] Adding the Download page and linking it to the Install page URL: https://github.com/apache/incubator-mxnet/pull/10477 ## Description ## Adding a download page with Apache mirror source links. ### Changes ### - [ ] Adding a download page on the mxnet website. - [ ] Download point to our mirror system, not gihtub managed links. ß--- ASF mirror http://54.210.6.225/install/download.html 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 issue #10445: Make the numpy version compatible with Official Mac OS system python-2.7.10's numpy
bhavinthaker commented on issue #10445: Make the numpy version compatible with Official Mac OS system python-2.7.10's numpy URL: https://github.com/apache/incubator-mxnet/pull/10445#issuecomment-379953639 Hi Marco: Can we keep range of Python version support the same in both CentOS and Ubuntu so that it is consistent across both OS versions and also in production and CI? 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] Sherry-Zhang commented on issue #10104: [WIP][MXNET-107] Fused RNN implementation for CPU
Sherry-Zhang commented on issue #10104: [WIP][MXNET-107] Fused RNN implementation for CPU URL: https://github.com/apache/incubator-mxnet/pull/10104#issuecomment-379951240 @sxjscience , not support yet, will do in the future(in other 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] anirudh2290 commented on issue #10452: [MXNET-294] Fix element wise multiply for csr ndarrays
anirudh2290 commented on issue #10452: [MXNET-294] Fix element wise multiply for csr ndarrays URL: https://github.com/apache/incubator-mxnet/pull/10452#issuecomment-379948348 Seems like a flaky test, opened an issue for the same: https://github.com/apache/incubator-mxnet/issues/10476. Can you please retrigger. 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] anirudh2290 opened a new issue #10476: Flaky test for test_operator.test_reduce
anirudh2290 opened a new issue #10476: Flaky test for test_operator.test_reduce URL: https://github.com/apache/incubator-mxnet/issues/10476 Try to use MXNET_TEST_SEED=1117986172 ``` == FAIL: test_operator.test_reduce -- Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest self.test(*self.arg) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/common.py", line 157, in test_new orig_test(*args, **kwargs) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1826, in test_reduce mx.symbol.max, test_none_axis=test_none) File "/home/ubuntu/experimentals/spidyDev_mxnet/tests/python/unittest/test_operator.py", line 1799, in test_reduce_inner assert equal_backward AssertionError: >> begin captured logging << common: INFO: Setting module np/mx/python random seeds, use MXNET_MODULE_SEED=754316987 to reproduce. common: INFO: Setting test np/mx/python random seeds, use MXNET_TEST_SEED=1117986172 to reproduce. - >> end captured logging << - -- Ran 1 test in 22.731s FAILED (failures=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] anirudh2290 commented on issue #10435: [MXNET-289] Allow specifying number of batches to run in an epoch to fix hang in dist sync for last epoch
anirudh2290 commented on issue #10435: [MXNET-289] Allow specifying number of batches to run in an epoch to fix hang in dist sync for last epoch URL: https://github.com/apache/incubator-mxnet/pull/10435#issuecomment-379947516 @piiswrong can this be merged ? 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] anirudh2290 commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix
anirudh2290 commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix URL: https://github.com/apache/incubator-mxnet/pull/10425#discussion_r180276893 ## File path: python/mxnet/module/base_module.py ## @@ -363,6 +363,19 @@ def predict(self, eval_data, num_batch=None, merge_batches=True, reset=True, eval_data.reset() output_list = [] +# If data provided is a dict. This makes sure that only relevant data items Review comment: Do we need to make this change for score 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] anirudh2290 commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix
anirudh2290 commented on a change in pull request #10425: [MXNET-175] NdArrayiterator dict fix URL: https://github.com/apache/incubator-mxnet/pull/10425#discussion_r180277044 ## File path: python/mxnet/io.py ## @@ -645,8 +647,8 @@ def __init__(self, data, label=None, batch_size=1, shuffle=False, label_name='softmax_label'): super(NDArrayIter, self).__init__(batch_size) -self.data = _init_data(data, allow_empty=False, default_name=data_name) -self.label = _init_data(label, allow_empty=True, default_name=label_name) +self.data, self.renamedData = _init_data(data, allow_empty=False, default_name=data_name) Review comment: snake case for renamedData? 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: Change the cuda version for docs test script (#10475)
This is an automated email from the ASF dual-hosted git repository. haibin pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 7d513db Change the cuda version for docs test script (#10475) 7d513db is described below commit 7d513dbde48c57ef336812454f3ae76f08fc00c9 Author: mbaijal <30911248+mbai...@users.noreply.github.com> AuthorDate: Mon Apr 9 18:23:42 2018 -0700 Change the cuda version for docs test script (#10475) --- tests/jenkins/run_test_installation_docs.sh | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tests/jenkins/run_test_installation_docs.sh b/tests/jenkins/run_test_installation_docs.sh index 90eecb9..4b3e449 100755 --- a/tests/jenkins/run_test_installation_docs.sh +++ b/tests/jenkins/run_test_installation_docs.sh @@ -302,13 +302,13 @@ echo echo "### Testing Virtualenv ###" echo "${virtualenv_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${virtualenv_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${virtualenv_commands}" echo echo "### Testing Pip ###" echo "${pip_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${pip_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${pip_commands}" echo echo "### Testing Docker ###" @@ -320,4 +320,4 @@ echo echo "### Testing Build From Source ###" echo "${buildfromsource_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${buildfromsource_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${buildfromsource_commands}" -- To stop receiving notification emails like this one, please contact hai...@apache.org.
[GitHub] eric-haibin-lin closed pull request #10475: Change the cuda version for docs test script
eric-haibin-lin closed pull request #10475: Change the cuda version for docs test script URL: https://github.com/apache/incubator-mxnet/pull/10475 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/tests/jenkins/run_test_installation_docs.sh b/tests/jenkins/run_test_installation_docs.sh index 90eecb95a32..4b3e4490296 100755 --- a/tests/jenkins/run_test_installation_docs.sh +++ b/tests/jenkins/run_test_installation_docs.sh @@ -302,13 +302,13 @@ echo echo "### Testing Virtualenv ###" echo "${virtualenv_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${virtualenv_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${virtualenv_commands}" echo echo "### Testing Pip ###" echo "${pip_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${pip_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${pip_commands}" echo echo "### Testing Docker ###" @@ -320,4 +320,4 @@ echo echo "### Testing Build From Source ###" echo "${buildfromsource_commands}" echo -nvidia-docker run --rm nvidia/cuda:7.5-cudnn5-devel bash -c "${buildfromsource_commands}" +nvidia-docker run --rm nvidia/cuda:9.0-cudnn7-devel bash -c "${buildfromsource_commands}" 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] zheng-da commented on issue #8532: mxnet-mkl (v0.12.0) crash when using (conda-installed) numpy with MKL
zheng-da commented on issue #8532: mxnet-mkl (v0.12.0) crash when using (conda-installed) numpy with MKL URL: https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-379943151 no, i don't have a workaround so far. 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 issue #10475: Change the cuda version for docs test script
mbaijal commented on issue #10475: Change the cuda version for docs test script URL: https://github.com/apache/incubator-mxnet/pull/10475#issuecomment-379940972 @eric-haibin-lin Can you please review and 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] b0noI commented on issue #10475: Change the cuda version for docs test script
b0noI commented on issue #10475: Change the cuda version for docs test script URL: https://github.com/apache/incubator-mxnet/pull/10475#issuecomment-379940778 LGTM 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] ashokei commented on issue #8532: mxnet-mkl (v0.12.0) crash when using (conda-installed) numpy with MKL
ashokei commented on issue #8532: mxnet-mkl (v0.12.0) crash when using (conda-installed) numpy with MKL URL: https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-379936533 @zheng-da is your openmp linking issue related to this in any way; Does your solution/workaround fix 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] mbaijal commented on a change in pull request #10390: Initial PR for RAT License Check
mbaijal commented on a change in pull request #10390: Initial PR for RAT License Check URL: https://github.com/apache/incubator-mxnet/pull/10390#discussion_r180267135 ## File path: tests/nightly/apache_rat_license_check/.rat-excludes ## @@ -0,0 +1,54 @@ +.*xml +\..* +.*css +\\.* +.*ipynb +.*html +.*json +.*js +.*txt +.*md +3rdparty/* +R-package/* +src/operator/mkl/* +trunk/* +docker/* +docker_multiarch/* +.*\\.m +.*\\.mk +.*\\.R +contrib/* +Dockerfile* +.*svg +.*cfg +.*config +docs/* +__init__.py +build/* +.*\\.t +perl-package/* Review comment: The following 13 files in the perl-package do not have a valid license - /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNet/Changes /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNet/MANIFEST /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNet/META.yml /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNetCAPI/Changes /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNetCAPI/MANIFEST /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNetCAPI/META.yml /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNetCAPI/mxnet.i /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-MXNetCAPI/mxnet_typemaps.i /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-NNVMCAPI/Changes /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-NNVMCAPI/MANIFEST /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-NNVMCAPI/META.yml /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-NNVMCAPI/nnvm.i /home/ubuntu/workspace/RAT_LicenseCheck/perl-package/AI-NNVMCAPI/nnvm_typemaps.i Do you know who can resolve 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] eric-haibin-lin commented on a change in pull request #10390: Initial PR for RAT License Check
eric-haibin-lin commented on a change in pull request #10390: Initial PR for RAT License Check URL: https://github.com/apache/incubator-mxnet/pull/10390#discussion_r180262066 ## File path: tests/nightly/apache_rat_license_check/.rat-excludes ## @@ -0,0 +1,54 @@ +.*xml +\..* +.*css +\\.* +.*ipynb +.*html +.*json +.*js +.*txt +.*md +3rdparty/* +R-package/* +src/operator/mkl/* +trunk/* +docker/* +docker_multiarch/* +.*\\.m +.*\\.mk +.*\\.R +contrib/* Review comment: How does other apache projects handle contrib folder? I thought licenses are required 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 #10474: [MXNET-298] Scala Infer API docs landing page
aaronmarkham commented on issue #10474: [MXNET-298] Scala Infer API docs landing page URL: https://github.com/apache/incubator-mxnet/pull/10474#issuecomment-379926690 @nswamy @Roshrini @lanking520 - please let me know your feedback. 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 #10424: Improved error message
ankkhedia commented on issue #10424: Improved error message URL: https://github.com/apache/incubator-mxnet/pull/10424#issuecomment-379926490 @piiswrong I have made changes as suggested. Could you please take a look? 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 opened a new pull request #10475: Change the cuda version for docs test script
mbaijal opened a new pull request #10475: Change the cuda version for docs test script URL: https://github.com/apache/incubator-mxnet/pull/10475 ## Description ## Installation Guide Nightly test is failing due to old cuda version. Update the CUDA version from 7.5 to 9.0. @aaronmarkham @b0noI Can you please review soon! @anirudh2290 This needs to make it into 1.2 release NOTE: I have not updated the cuda version to 9.1 since the slaves on the old(nightly) CI do not have the correct nvidia drivers. These tests are being migrated to the official CI in the coming weeks. ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) created (except PRs with tiny changes) - [ ] Changes are complete (i.e. I finished coding on this PR) - [ ] All changes have test coverage: - Unit tests are added for small changes to verify correctness (e.g. adding a new operator) - Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore) - Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL) - [ ] Code is well-documented: - For user-facing API changes, API doc string has been updated. - For new C++ functions in header files, their functionalities and arguments are documented. - For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable - Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html - [ ] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change ### Changes ### - [ ] Updated cuda from 7.5 to 9.0 - [ ] Updated Cudnn from 5 to 7 ## Comments ## - We will know whether this change fixes the nightly test as expected only after the PR is merged and I run the nightly test. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] aaronmarkham opened a new pull request #10474: Scala Infer API docs landing page
aaronmarkham opened a new pull request #10474: Scala Infer API docs landing page URL: https://github.com/apache/incubator-mxnet/pull/10474 ## Description ## Website docs updates for the Infer API. ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [x] The PR title starts with [MXNET-$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] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change ### Changes ### - [x] Added infer landing page. - [x] Added infer overview to scala api landing. - [x] Added readme for scala tutorials folder. - [x] Fixed broken links on scala landing page (due to namespace 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] lanking520 commented on issue #8344: Broken link for "MXNet R Reference Manual"?
lanking520 commented on issue #8344: Broken link for "MXNet R Reference Manual"? URL: https://github.com/apache/incubator-mxnet/issues/8344#issuecomment-379920422 You are right. Once the PR merged, the link will be fixed. 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] dslate commented on issue #8344: Broken link for "MXNet R Reference Manual"?
dslate commented on issue #8344: Broken link for "MXNet R Reference Manual"? URL: https://github.com/apache/incubator-mxnet/issues/8344#issuecomment-379919933 @lanking520 Thanks, I was able to download the new R mxnet manual at https://s3.amazonaws.com/mxnet-prod/docs/R/mxnet-r-reference-manual.pdf. However, the page I originally referenced, http://mxnet.incubator.apache.org/api/r/index.html, still contains a link titled 'MXNet R Reference Manual' that points to the wrong document (mxnet-test.pdf). 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] cjolivier01 commented on issue #8866: src/operator/./bilinear_sampler-inl.h:105: Have not implemented the data req combinations! gdata_req=0 ggrid_req=1
cjolivier01 commented on issue #8866: src/operator/./bilinear_sampler-inl.h:105: Have not implemented the data req combinations! gdata_req=0 ggrid_req=1 URL: https://github.com/apache/incubator-mxnet/issues/8866#issuecomment-379918394 I don't understand the problem description and am unfamiliar with this code. seems like @sxjscience knows what the fix should be, so reassigning. 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] cjolivier01 commented on issue #6644: incompatibility between backward mirror and batch normalization
cjolivier01 commented on issue #6644: incompatibility between backward mirror and batch normalization URL: https://github.com/apache/incubator-mxnet/issues/6644#issuecomment-379916034 no longer an issue 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] cjolivier01 closed issue #6644: incompatibility between backward mirror and batch normalization
cjolivier01 closed issue #6644: incompatibility between backward mirror and batch normalization URL: https://github.com/apache/incubator-mxnet/issues/6644 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 #10286: [MXNET-250] CSS and .htaccess fix
aaronmarkham commented on issue #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#issuecomment-379421679 We need to have the package links in there, but this will be handled with a subsequent 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] spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model
spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model URL: https://github.com/apache/incubator-mxnet/pull/10472#discussion_r180247510 ## File path: tests/python-pytest/onnx/onnx_test.py ## @@ -214,8 +211,8 @@ def test_bvlc_reference_caffenet(): # run test for each test file for input_data, output_data in zip(inputs, outputs): # create module -mod = mx.mod.Module(symbol=sym, data_names=['input_0'], context=mx.cpu(), label_names=None) -mod.bind(for_training=False, data_shapes=[('input_0', input_data.shape)], label_shapes=None) +mod = mx.mod.Module(symbol=sym, data_names=['data_0'], context=mx.cpu(), label_names=None) Review comment: same as 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
[GitHub] aaronmarkham commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix
aaronmarkham commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#discussion_r180208385 ## File path: docs/build_version_doc/update_all_version.sh ## @@ -68,31 +73,44 @@ fi # Write all version numbers into $tag_file for AddVersion.py to use later # Master is added by that script by default for tag in $tag_list; do -if [ $tag != 'master' ] -then -echo "$tag" >> "$tag_file" -fi +echo "$tag" >> "$tag_file" done +function update_mxnet_css { + tag=$1 + echo "Begin update fixes.." + # All fixes are done on the master branch of mxnet-incubator repository + # During a nightly build, these fixes will be patched to all the versions in the asf-site repository including the master folder under versions directory. + # copy + + echo "Copying mxnet.css from master branch to all versions" + cp "$MASTER_SOURCE_DIR/$STATIC_FILES_DIR/mxnet.css" "$built/versions/$tag/_static" + + echo "Update fixes complete.." +} + # Update the specified tags with the Versions dropdown for tag in $tag_list; do # This Python script is expecting the tag_list.txt and it will use that as the entries to populate + python AddVersion.py --root_url "$root_url" --file_path "$built/versions/$tag" --current_version "$tag" || exit 1 -if [ $tag != 'master' ] -then -python AddPackageLink.py --file_path "$built/versions/master/install/index.html" \ Review comment: Why are these removed entirely? 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] spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model
spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model URL: https://github.com/apache/incubator-mxnet/pull/10472#discussion_r180247544 ## File path: tests/python-pytest/onnx/onnx_test.py ## @@ -214,8 +211,8 @@ def test_bvlc_reference_caffenet(): # run test for each test file for input_data, output_data in zip(inputs, outputs): # create module -mod = mx.mod.Module(symbol=sym, data_names=['input_0'], context=mx.cpu(), label_names=None) -mod.bind(for_training=False, data_shapes=[('input_0', input_data.shape)], label_shapes=None) +mod = mx.mod.Module(symbol=sym, data_names=['data_0'], context=mx.cpu(), label_names=None) +mod.bind(for_training=False, data_shapes=[('data_0', input_data.shape)], label_shapes=None) Review comment: same as 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
[GitHub] spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model
spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model URL: https://github.com/apache/incubator-mxnet/pull/10472#discussion_r180247473 ## File path: tests/python-pytest/onnx/onnx_test.py ## @@ -192,8 +189,8 @@ def test_bvlc_googlenet(): # run test for each test file for input_data, output_data in zip(inputs, outputs): # create module -mod = mx.mod.Module(symbol=sym, data_names=['input_0'], context=mx.cpu(), label_names=None) -mod.bind(for_training=False, data_shapes=[('input_0', input_data.shape)], label_shapes=None) +mod = mx.mod.Module(symbol=sym, data_names=['data_0'], context=mx.cpu(), label_names=None) Review comment: Can we fetch the data_names from sym ? 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] spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model
spidyDev commented on a change in pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model URL: https://github.com/apache/incubator-mxnet/pull/10472#discussion_r180247574 ## File path: tests/python-pytest/onnx/onnx_test.py ## @@ -236,8 +233,8 @@ def test_bvlc_rcnn_ilsvrc13(): # run test for each test file for input_data, output_data in zip(inputs, outputs): # create module -mod = mx.mod.Module(symbol=sym, data_names=['input_0'], context=mx.cpu(), label_names=None) -mod.bind(for_training=False, data_shapes=[('input_0', input_data.shape)], label_shapes=None) +mod = mx.mod.Module(symbol=sym, data_names=['data_0'], context=mx.cpu(), label_names=None) Review comment: same as 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
[GitHub] Jerryzcn commented on issue #10453: Bug of CuDNN RNN with variable sequence length
Jerryzcn commented on issue #10453: Bug of CuDNN RNN with variable sequence length URL: https://github.com/apache/incubator-mxnet/issues/10453#issuecomment-379910659 I have similar issue when training speech model. even after `export MXNET_GPU_MEM_POOL_RESERVE=7` will try larger RESERVE 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] anirudh2290 commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn
anirudh2290 commented on issue #10021: [MXNET-33] SSD example not working with mkl-dnn URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-379909946 @ashokei can you please check if the CI failure is related to your change and retrigger it if not ? 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 #10414: There is no API Reference Doc for R bindings.
lanking520 commented on issue #10414: There is no API Reference Doc for R bindings. URL: https://github.com/apache/incubator-mxnet/issues/10414#issuecomment-379908725 Update in this PR: https://github.com/apache/incubator-mxnet/pull/10473 @anirudhacharya 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 a change in pull request #10449: [MXNET-293] Navigation updates
aaronmarkham commented on a change in pull request #10449: [MXNET-293] Navigation updates URL: https://github.com/apache/incubator-mxnet/pull/10449#discussion_r180245363 ## File path: docs/_static/mxnet-theme/navbar.html ## @@ -7,47 +7,39 @@ Install - - Gluon - -About -http://gluon.mxnet.io;>Tutorials Review comment: You can get to the Gluon book from the tutorials page. Plus @thomelane will be moving that link on the tutorials page above the fold, so it will be easier to see. 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 #8344: Broken link for "MXNet R Reference Manual"?
lanking520 commented on issue #8344: Broken link for "MXNet R Reference Manual"? URL: https://github.com/apache/incubator-mxnet/issues/8344#issuecomment-379908057 @dslate Hi, we have added the old R documentation in this PR. https://github.com/apache/incubator-mxnet/pull/10473 Hope this helps. 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 #10473: [MXNET-297] updated R docs
lanking520 commented on issue #10473: [MXNET-297] updated R docs URL: https://github.com/apache/incubator-mxnet/pull/10473#issuecomment-379907684 Looks good! we finally have a R documentation there 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] anirudh2290 commented on issue #10317: [MXNET-264] Improve performance of MKLDNN in small batch sizes.
anirudh2290 commented on issue #10317: [MXNET-264] Improve performance of MKLDNN in small batch sizes. URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-379907529 @piiswrong can this be merged ? 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] anirudh2290 commented on issue #10367: [MXNET-262] Implement mx.random.seed_context to seed random number generators of a specific device context
anirudh2290 commented on issue #10367: [MXNET-262] Implement mx.random.seed_context to seed random number generators of a specific device context URL: https://github.com/apache/incubator-mxnet/pull/10367#issuecomment-379907142 @piiswrong can this be merged. will add this to api changes in release notes. 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 opened a new pull request #10473: updated R docs
aaronmarkham opened a new pull request #10473: updated R docs URL: https://github.com/apache/incubator-mxnet/pull/10473 ## Description ## R docs have been regenerated and a new pdf is available. ## 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] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change ### Changes ### - [x] This updates the link to the reference docs. 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] piiswrong closed pull request #10443: bug fix
piiswrong closed pull request #10443: bug fix URL: https://github.com/apache/incubator-mxnet/pull/10443 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: bug fix (#10443)
This is an automated email from the ASF dual-hosted git repository. jxie pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 81a3827 bug fix (#10443) 81a3827 is described below commit 81a38274f399f8e24b7415188ce6f1745062e269 Author: aluesAuthorDate: Tue Apr 10 05:45:28 2018 +0800 bug fix (#10443) --- python/mxnet/__init__.py | 1 + 1 file changed, 1 insertion(+) diff --git a/python/mxnet/__init__.py b/python/mxnet/__init__.py index 4e2c4f0..58b8bd8 100644 --- a/python/mxnet/__init__.py +++ b/python/mxnet/__init__.py @@ -41,6 +41,7 @@ from . import random as rnd from . import random from . import optimizer from . import model +from . import metric from . import notebook from . import initializer # use mx.init as short for mx.initializer -- To stop receiving notification emails like this one, please contact j...@apache.org.
[GitHub] piiswrong closed pull request #10470: CUDNN not training on backward pass similar to pytorch
piiswrong closed pull request #10470: CUDNN not training on backward pass similar to pytorch URL: https://github.com/apache/incubator-mxnet/pull/10470 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/src/operator/nn/cudnn/cudnn_batch_norm-inl.h b/src/operator/nn/cudnn/cudnn_batch_norm-inl.h index e3d5dd9204b..d4b9f84ed2f 100644 --- a/src/operator/nn/cudnn/cudnn_batch_norm-inl.h +++ b/src/operator/nn/cudnn/cudnn_batch_norm-inl.h @@ -165,8 +165,6 @@ class CuDNNBatchNormOp { using namespace mshadow::expr; CHECK_EQ(inputs.size(), 8U); CHECK_EQ(outputs.size(), 3U); -CHECK(ctx.is_train && !param_.use_global_stats) -<< "use global statistics is not yet supported in CuDNNBatchNorm"; // Rename the inputs and outputs. const TBlob _grad = inputs[0]; @@ -183,6 +181,8 @@ class CuDNNBatchNormOp { in_grad[cudnnbatchnorm::kData].get_with_shape(shape_, s); Tensor dy = out_grad.get_with_shape (shape_, s); +const bool global_stats = !ctx.is_train || param_.use_global_stats; + #if CUDNN_VERSION >= 4007 #if CUDNN_VERSION >= 7002 auto mode = CUDNN_BATCHNORM_SPATIAL_PERSISTENT; @@ -226,8 +226,8 @@ class CuDNNBatchNormOp { dgamma.dptr_, dbeta.dptr_, param_.eps, -save_mean.dptr_, -save_inv_var.dptr_)); +global_stats ? nullptr : save_mean.dptr_, +global_stats ? nullptr : save_inv_var.dptr_)); if (param_.fix_gamma) dgamma = 0.f; }) #else // CUDNN_VERSION < 4007 @@ -264,8 +264,8 @@ class CuDNNBatchNormOp { dgamma.dptr_, dbeta.dptr_, param_.eps, - save_mean.dptr_, - save_inv_var.dptr_)); + global_stats ? nullptr : save_mean.dptr_, + global_stats ? nullptr : save_inv_var.dptr_)); if (param_.fix_gamma) dgamma = 0.f; }) #endif diff --git a/tests/python/gpu/test_operator_gpu.py b/tests/python/gpu/test_operator_gpu.py index 2dd66ee2d10..36af524737d 100644 --- a/tests/python/gpu/test_operator_gpu.py +++ b/tests/python/gpu/test_operator_gpu.py @@ -27,6 +27,7 @@ from nose.tools import assert_raises from mxnet.test_utils import check_consistency, set_default_context, assert_almost_equal from mxnet.base import MXNetError +from mxnet import autograd from numpy.testing import assert_allclose curr_path = os.path.dirname(os.path.abspath(os.path.expanduser(__file__))) @@ -1786,7 +1787,24 @@ def test_incorrect_gpu(): # Try setting dev_id to a really big number assert_raises(MXNetError, mx.nd.ones, (2,2), ctx=mx.gpu(11)) +@with_seed() +def test_batchnorm_backwards_notrain(): +for ctx in [mx.cpu(0), mx.gpu(0)]: +for cudnn_o in [False, True]: +B,C,H,W = 4,3,2,2 +x = mx.nd.random.poisson(1,shape=(B,C,H,W)).as_in_context(ctx) +gamma = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +beta = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +mean = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +std = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +x.attach_grad() + +with autograd.record(False): +y = mx.ndarray.BatchNorm(x, gamma, beta, mean, std.square(), + fix_gamma=False, cudnn_off=cudnn_o) +loss=y.square().sum() +loss.backward(train_mode=False) + if __name__ == '__main__': import nose nose.runmodule() - 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: CUDNN not training on backward pass similar to pytorch (#10470)
This is an automated email from the ASF dual-hosted git repository. jxie pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 649a0a1 CUDNN not training on backward pass similar to pytorch (#10470) 649a0a1 is described below commit 649a0a16b91e4c81ec911ecc3be0439d9ebc2a52 Author: Chris OlivierAuthorDate: Mon Apr 9 14:43:53 2018 -0700 CUDNN not training on backward pass similar to pytorch (#10470) * CUDNN not training on backward pass similar to pytorch https://github.com/pytorch/pytorch/issues/4284 * add test * add seed decorator * Trigger --- src/operator/nn/cudnn/cudnn_batch_norm-inl.h | 12 ++-- tests/python/gpu/test_operator_gpu.py| 20 +++- 2 files changed, 25 insertions(+), 7 deletions(-) diff --git a/src/operator/nn/cudnn/cudnn_batch_norm-inl.h b/src/operator/nn/cudnn/cudnn_batch_norm-inl.h index e3d5dd9..d4b9f84 100644 --- a/src/operator/nn/cudnn/cudnn_batch_norm-inl.h +++ b/src/operator/nn/cudnn/cudnn_batch_norm-inl.h @@ -165,8 +165,6 @@ class CuDNNBatchNormOp { using namespace mshadow::expr; CHECK_EQ(inputs.size(), 8U); CHECK_EQ(outputs.size(), 3U); -CHECK(ctx.is_train && !param_.use_global_stats) -<< "use global statistics is not yet supported in CuDNNBatchNorm"; // Rename the inputs and outputs. const TBlob _grad = inputs[0]; @@ -183,6 +181,8 @@ class CuDNNBatchNormOp { in_grad[cudnnbatchnorm::kData].get_with_shape (shape_, s); Tensor dy = out_grad.get_with_shape (shape_, s); +const bool global_stats = !ctx.is_train || param_.use_global_stats; + #if CUDNN_VERSION >= 4007 #if CUDNN_VERSION >= 7002 auto mode = CUDNN_BATCHNORM_SPATIAL_PERSISTENT; @@ -226,8 +226,8 @@ class CuDNNBatchNormOp { dgamma.dptr_, dbeta.dptr_, param_.eps, -save_mean.dptr_, -save_inv_var.dptr_)); +global_stats ? nullptr : save_mean.dptr_, +global_stats ? nullptr : save_inv_var.dptr_)); if (param_.fix_gamma) dgamma = 0.f; }) #else // CUDNN_VERSION < 4007 @@ -264,8 +264,8 @@ class CuDNNBatchNormOp { dgamma.dptr_, dbeta.dptr_, param_.eps, - save_mean.dptr_, - save_inv_var.dptr_)); + global_stats ? nullptr : save_mean.dptr_, + global_stats ? nullptr : save_inv_var.dptr_)); if (param_.fix_gamma) dgamma = 0.f; }) #endif diff --git a/tests/python/gpu/test_operator_gpu.py b/tests/python/gpu/test_operator_gpu.py index 85b3e26..08c749e 100644 --- a/tests/python/gpu/test_operator_gpu.py +++ b/tests/python/gpu/test_operator_gpu.py @@ -27,6 +27,7 @@ import unittest from nose.tools import assert_raises from mxnet.test_utils import check_consistency, set_default_context, assert_almost_equal from mxnet.base import MXNetError +from mxnet import autograd from numpy.testing import assert_allclose curr_path = os.path.dirname(os.path.abspath(os.path.expanduser(__file__))) @@ -1815,7 +1816,24 @@ def test_incorrect_gpu(): # Try setting dev_id to a really big number assert_raises(MXNetError, mx.nd.ones, (2,2), ctx=mx.gpu(11)) +@with_seed() +def test_batchnorm_backwards_notrain(): +for ctx in [mx.cpu(0), mx.gpu(0)]: +for cudnn_o in [False, True]: +B,C,H,W = 4,3,2,2 +x = mx.nd.random.poisson(1,shape=(B,C,H,W)).as_in_context(ctx) +gamma = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +beta = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +mean = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +std = mx.nd.random.normal(shape=(C)).as_in_context(ctx) +x.attach_grad() + +with autograd.record(False): +y = mx.ndarray.BatchNorm(x, gamma, beta, mean, std.square(), + fix_gamma=False, cudnn_off=cudnn_o) +loss=y.square().sum() +loss.backward(train_mode=False) + if __name__ == '__main__': import nose nose.runmodule() - -- To stop receiving notification emails like this one, please contact j...@apache.org.
[GitHub] rahul003 commented on a change in pull request #10452: [MXNET-294] Fix element wise multiply for csr ndarrays
rahul003 commented on a change in pull request #10452: [MXNET-294] Fix element wise multiply for csr ndarrays URL: https://github.com/apache/incubator-mxnet/pull/10452#discussion_r180239182 ## File path: src/operator/tensor/elemwise_binary_op-inl.h ## @@ -292,9 +292,7 @@ void ElemwiseBinaryOp::CsrCsrOp(mshadow::Stream *s, OpBase::FillDense(s, next.shape_.Size(), IType(-1), req, next.dptr_); OpBase::FillDense(s, lhs_row.shape_.Size(), DType(0), req, lhs_row.dptr_); - if (!same_lhs_rhs) { -OpBase::FillDense(s, rhs_row.shape_.Size(), DType(0), req, rhs_row.dptr_); - } + OpBase::FillDense(s, rhs_row.shape_.Size(), DType(0), req, rhs_row.dptr_); Review comment: Thanks for the suggestion, changed 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 opened a new pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model
anirudhacharya opened a new pull request #10472: [MXNET-106] [ONNX_MXNet] Change parameter names in imported model URL: https://github.com/apache/incubator-mxnet/pull/10472 ## Description This PR fixes the names of inputs parameters in the imported model. Previously the imported model had input parameter names like - input_0, param_0, param_1 etc.. This PR changes it to use the same names as in the ONNX model file. While the current format works fine, this change was done to remain consistent with the ONNX tool chain. So that users moving between frameworks will not have to encounter different input and parameter names in their models. ## 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) - [ ] All changes have test coverage: - [ ] 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] bhavinthaker commented on issue #10185: [Documentation][Installation][macOS] Building From Source Has Incomplete Instructions For macOS 10.13.3
bhavinthaker commented on issue #10185: [Documentation][Installation][macOS] Building From Source Has Incomplete Instructions For macOS 10.13.3 URL: https://github.com/apache/incubator-mxnet/issues/10185#issuecomment-379898234 @conradwt : Could you please check if this script runs fine in your environment? curl -O https://raw.githubusercontent.com/dmlc/mxnet/master/setup-utils/install-mxnet-osx-python.sh 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 closed issue #10249: Custom interpolation layer. Gluon.
szha closed issue #10249: Custom interpolation layer. Gluon. URL: https://github.com/apache/incubator-mxnet/issues/10249 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 #10249: Custom interpolation layer. Gluon.
zhanghang1989 commented on issue #10249: Custom interpolation layer. Gluon. URL: https://github.com/apache/incubator-mxnet/issues/10249#issuecomment-379896540 ```python class BilinearInterp(Block): def __init__(self, factor, target_height, target_width, **kwargs): super(BilinearInterp, self).__init__(**kwargs) self.height = target_height self.width = target_height def forward(self, x): return mx.nd.contrib.BilinearResize2D(x, self.height, self.width) ``` 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 #10249: Custom interpolation layer. Gluon.
zhanghang1989 commented on issue #10249: Custom interpolation layer. Gluon. URL: https://github.com/apache/incubator-mxnet/issues/10249#issuecomment-379896540 ```python class BilinearInterp(Block): def __init__(self, factor, target_height, target_width, **kwargs): super(BilinearInterp, self).__init__(**kwargs) self.height = target_height self.width = target_height def forward(self, x): return mx.nd. BilinearResize2D(x, self.height, self.width) ``` 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 closed issue #9138: I cannot make mxnet.ndarray.UpSampling work with bilinear interpolation
szha closed issue #9138: I cannot make mxnet.ndarray.UpSampling work with bilinear interpolation URL: https://github.com/apache/incubator-mxnet/issues/9138 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 closed issue #8560: Upsampling: lack of arguments crashes the kernel
szha closed issue #8560: Upsampling: lack of arguments crashes the kernel URL: https://github.com/apache/incubator-mxnet/issues/8560 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 closed issue #4134: how to use mx.sym.UpSampling for bilinear upsampling
szha closed issue #4134: how to use mx.sym.UpSampling for bilinear upsampling URL: https://github.com/apache/incubator-mxnet/issues/4134 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 issue #10390: Initial PR for RAT License Check
mbaijal commented on issue #10390: Initial PR for RAT License Check URL: https://github.com/apache/incubator-mxnet/pull/10390#issuecomment-379892387 I think this PR can be merged as is. The list of directories included in the .rat-excludes file has been discussed in the previous release and reasons for most decisions are documented here - https://cwiki.apache.org/confluence/display/MXNET/MXNet+Source+Licenses In most cases, either the file should have no license or I am unsure about the correct license. As we move forward, where ever possible the community can add the correct licenses and clean this list further. 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] Roshrini commented on issue #10446: [MXNET-173]fix acc metric shape miss match
Roshrini commented on issue #10446: [MXNET-173]fix acc metric shape miss match URL: https://github.com/apache/incubator-mxnet/pull/10446#issuecomment-379888132 @reminisce @sandeep-krishnamurthy Can you 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] anirudh2290 commented on issue #10390: Initial PR for RAT License Check
anirudh2290 commented on issue #10390: Initial PR for RAT License Check URL: https://github.com/apache/incubator-mxnet/pull/10390#issuecomment-379886418 @mbaijal any updates ? 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] gautamkmr commented on issue #10445: Make the numpy version compatible with Official Mac OS system python-2.7.10's numpy
gautamkmr commented on issue #10445: Make the numpy version compatible with Official Mac OS system python-2.7.10's numpy URL: https://github.com/apache/incubator-mxnet/pull/10445#issuecomment-379883343 So looks like 1.7 also didnt work. Do you want me to add installation for this package "python-dev|python-devel" ? I still thinks we don't need to have setup.py and CI with same pinned version given setup.py can't be pinned. The reason is our customers are based on different version of MAC OS X. WDYT? 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] anirudh2290 commented on issue #10417: [MXNET-283] Error handling for non-positive reps of tile op
anirudh2290 commented on issue #10417: [MXNET-283] Error handling for non-positive reps of tile op URL: https://github.com/apache/incubator-mxnet/pull/10417#issuecomment-379882685 @piiswrong is this good to 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] anirudh2290 commented on issue #10424: Improved error message
anirudh2290 commented on issue #10424: Improved error message URL: https://github.com/apache/incubator-mxnet/pull/10424#issuecomment-379882169 ping @ankkhedia 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: Revert "[MXNET-265] Update optimizer doc to clarify wd behaviors (#10388)" (#10471)
This is an automated email from the ASF dual-hosted git repository. haibin pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new f9a34f8 Revert "[MXNET-265] Update optimizer doc to clarify wd behaviors (#10388)" (#10471) f9a34f8 is described below commit f9a34f874cc80d6199fb55cd49c6bca1df1dd486 Author: Haibin LinAuthorDate: Mon Apr 9 12:54:56 2018 -0700 Revert "[MXNET-265] Update optimizer doc to clarify wd behaviors (#10388)" (#10471) This reverts commit c24765f6088fd66d072dfc3402998680de659217. --- python/mxnet/optimizer.py | 40 ++- tests/python/unittest/test_optimizer.py | 89 ++--- 2 files changed, 7 insertions(+), 122 deletions(-) diff --git a/python/mxnet/optimizer.py b/python/mxnet/optimizer.py index 18bd5c6..6589e77 100644 --- a/python/mxnet/optimizer.py +++ b/python/mxnet/optimizer.py @@ -538,7 +538,7 @@ class SGD(Optimizer): class Signum(Optimizer): """The Signum optimizer that takes the sign of gradient or momentum. -The optimizer updates the weight by:: +The optimizer updates the weight by: rescaled_grad = rescale_grad * clip(grad, clip_gradient) + wd * weight state = momentum * state + (1-momentum)*rescaled_grad @@ -604,14 +604,6 @@ class FTML(Optimizer): *FTML - Follow the Moving Leader in Deep Learning*, available at http://proceedings.mlr.press/v70/zheng17a/zheng17a.pdf. -Denote time step by t. The optimizer updates the weight by:: - -rescaled_grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -v = beta2 * v + (1 - beta2) * square(rescaled_grad) -d_t = (1 - power(beta1, t)) / lr * square_root(v / (1 - power(beta2, t))) + epsilon) -z = beta1 * z + (1 - beta1) * rescaled_grad - (d_t - beta1 * d_(t-1)) * weight -weight = - z / d_t - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1076,13 +1068,6 @@ class AdaGrad(Optimizer): Methods for Online Learning and Stochastic Optimization*, and available at http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf. -This optimizer updates each weight by:: - -grad = clip(grad * rescale_grad + weight * wd, clip_gradient) -history += square(grad) -div = grad / sqrt(history + float_stable_eps) -weight += div * -lr - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1120,12 +1105,12 @@ class AdaGrad(Optimizer): kwargs['clip_gradient'] = self.clip_gradient sparse.adagrad_update(weight, grad, history, out=weight, lr=lr, wd=wd, **kwargs) else: -grad = grad * self.rescale_grad + weight * wd +grad = grad * self.rescale_grad if self.clip_gradient is not None: grad = clip(grad, -self.clip_gradient, self.clip_gradient) history[:] += square(grad) div = grad / sqrt(history + self.float_stable_eps) -weight[:] += div * -lr +weight[:] += (div + weight * wd) * -lr @register class RMSProp(Optimizer): @@ -1210,15 +1195,6 @@ class AdaDelta(Optimizer): This class implements AdaDelta, an optimizer described in *ADADELTA: An adaptive learning rate method*, available at https://arxiv.org/abs/1212.5701. -This optimizer updates each weight by:: - -grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -acc_grad = rho * acc_grad + (1. - rho) * grad * grad -delta = sqrt(acc_delta + epsilon) / sqrt(acc_grad + epsilon) * grad -acc_delta = rho * acc_delta + (1. - rho) * delta * delta -weight -= delta - - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1246,7 +1222,6 @@ class AdaDelta(Optimizer): # preprocess grad grad *= self.rescale_grad -grad += wd * weight if self.clip_gradient is not None: grad = clip(grad, -self.clip_gradient, self.clip_gradient) @@ -1259,7 +1234,7 @@ class AdaDelta(Optimizer): acc_delta[:] = self.rho * acc_delta + (1. - self.rho) * current_delta * current_delta # update weight -weight[:] -= current_delta +weight[:] -= current_delta + wd * weight #pylint: disable=invalid-name #pylint: disable=line-too-long @@ -1346,13 +1321,6 @@ class Adamax(Optimizer): It is a variant of Adam based on the infinity norm available at http://arxiv.org/abs/1412.6980 Section 7. -The optimizer updates the weight by:: - -grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -m = beta1 * m_t + (1 - beta1) * grad -u = maximum(beta2 * u,
[GitHub] eric-haibin-lin closed pull request #10471: Revert "[MXNET-265] Update optimizer doc to clarify wd behaviors (#10388)
eric-haibin-lin closed pull request #10471: Revert "[MXNET-265] Update optimizer doc to clarify wd behaviors (#10388) URL: https://github.com/apache/incubator-mxnet/pull/10471 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/python/mxnet/optimizer.py b/python/mxnet/optimizer.py index 18bd5c6e4fe..6589e77e453 100644 --- a/python/mxnet/optimizer.py +++ b/python/mxnet/optimizer.py @@ -538,7 +538,7 @@ def update_multi_precision(self, index, weight, grad, state): class Signum(Optimizer): """The Signum optimizer that takes the sign of gradient or momentum. -The optimizer updates the weight by:: +The optimizer updates the weight by: rescaled_grad = rescale_grad * clip(grad, clip_gradient) + wd * weight state = momentum * state + (1-momentum)*rescaled_grad @@ -604,14 +604,6 @@ class FTML(Optimizer): *FTML - Follow the Moving Leader in Deep Learning*, available at http://proceedings.mlr.press/v70/zheng17a/zheng17a.pdf. -Denote time step by t. The optimizer updates the weight by:: - -rescaled_grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -v = beta2 * v + (1 - beta2) * square(rescaled_grad) -d_t = (1 - power(beta1, t)) / lr * square_root(v / (1 - power(beta2, t))) + epsilon) -z = beta1 * z + (1 - beta1) * rescaled_grad - (d_t - beta1 * d_(t-1)) * weight -weight = - z / d_t - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1076,13 +1068,6 @@ class AdaGrad(Optimizer): Methods for Online Learning and Stochastic Optimization*, and available at http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf. -This optimizer updates each weight by:: - -grad = clip(grad * rescale_grad + weight * wd, clip_gradient) -history += square(grad) -div = grad / sqrt(history + float_stable_eps) -weight += div * -lr - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1120,12 +1105,12 @@ def update(self, index, weight, grad, state): kwargs['clip_gradient'] = self.clip_gradient sparse.adagrad_update(weight, grad, history, out=weight, lr=lr, wd=wd, **kwargs) else: -grad = grad * self.rescale_grad + weight * wd +grad = grad * self.rescale_grad if self.clip_gradient is not None: grad = clip(grad, -self.clip_gradient, self.clip_gradient) history[:] += square(grad) div = grad / sqrt(history + self.float_stable_eps) -weight[:] += div * -lr +weight[:] += (div + weight * wd) * -lr @register class RMSProp(Optimizer): @@ -1210,15 +1195,6 @@ class AdaDelta(Optimizer): This class implements AdaDelta, an optimizer described in *ADADELTA: An adaptive learning rate method*, available at https://arxiv.org/abs/1212.5701. -This optimizer updates each weight by:: - -grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -acc_grad = rho * acc_grad + (1. - rho) * grad * grad -delta = sqrt(acc_delta + epsilon) / sqrt(acc_grad + epsilon) * grad -acc_delta = rho * acc_delta + (1. - rho) * delta * delta -weight -= delta - - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. @@ -1246,7 +1222,6 @@ def update(self, index, weight, grad, state): # preprocess grad grad *= self.rescale_grad -grad += wd * weight if self.clip_gradient is not None: grad = clip(grad, -self.clip_gradient, self.clip_gradient) @@ -1259,7 +1234,7 @@ def update(self, index, weight, grad, state): acc_delta[:] = self.rho * acc_delta + (1. - self.rho) * current_delta * current_delta # update weight -weight[:] -= current_delta +weight[:] -= current_delta + wd * weight #pylint: disable=invalid-name #pylint: disable=line-too-long @@ -1346,13 +1321,6 @@ class Adamax(Optimizer): It is a variant of Adam based on the infinity norm available at http://arxiv.org/abs/1412.6980 Section 7. -The optimizer updates the weight by:: - -grad = clip(grad * rescale_grad + wd * weight, clip_gradient) -m = beta1 * m_t + (1 - beta1) * grad -u = maximum(beta2 * u, abs(grad)) -weight -= lr / (1 - beta1**t) * m / u - This optimizer accepts the following parameters in addition to those accepted by :class:`.Optimizer`. diff --git a/tests/python/unittest/test_optimizer.py b/tests/python/unittest/test_optimizer.py index 07b306702f3..bbd7845f66f 100644 ---
[GitHub] reminisce commented on a change in pull request #10429: [MXNET-210]give warning for variables with same name in graph visualization
reminisce commented on a change in pull request #10429: [MXNET-210]give warning for variables with same name in graph visualization URL: https://github.com/apache/incubator-mxnet/pull/10429#discussion_r180208742 ## File path: python/mxnet/visualization.py ## @@ -252,6 +252,10 @@ def plot_network(symbol, title="plot", save_format='pdf', shape=None, node_attrs shape_dict = dict(zip(interals.list_outputs(), out_shapes)) conf = json.loads(symbol.tojson()) nodes = conf["nodes"] +# check if multiple nodes have the same name +if len(nodes) != len(set([node["name"] for node in nodes])): +logging.warning("There are multiple variables with the same name in your graph, " Review comment: Second 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] aaronmarkham commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix
aaronmarkham commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#discussion_r180208385 ## File path: docs/build_version_doc/update_all_version.sh ## @@ -68,31 +73,44 @@ fi # Write all version numbers into $tag_file for AddVersion.py to use later # Master is added by that script by default for tag in $tag_list; do -if [ $tag != 'master' ] -then -echo "$tag" >> "$tag_file" -fi +echo "$tag" >> "$tag_file" done +function update_mxnet_css { + tag=$1 + echo "Begin update fixes.." + # All fixes are done on the master branch of mxnet-incubator repository + # During a nightly build, these fixes will be patched to all the versions in the asf-site repository including the master folder under versions directory. + # copy + + echo "Copying mxnet.css from master branch to all versions" + cp "$MASTER_SOURCE_DIR/$STATIC_FILES_DIR/mxnet.css" "$built/versions/$tag/_static" + + echo "Update fixes complete.." +} + # Update the specified tags with the Versions dropdown for tag in $tag_list; do # This Python script is expecting the tag_list.txt and it will use that as the entries to populate + python AddVersion.py --root_url "$root_url" --file_path "$built/versions/$tag" --current_version "$tag" || exit 1 -if [ $tag != 'master' ] -then -python AddPackageLink.py --file_path "$built/versions/master/install/index.html" \ Review comment: Why are these removed entirely? 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] thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix
thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#discussion_r180206044 ## File path: docs/build_version_doc/update_all_version.sh ## @@ -68,31 +73,51 @@ fi # Write all version numbers into $tag_file for AddVersion.py to use later # Master is added by that script by default for tag in $tag_list; do -if [ $tag != 'master' ] -then -echo "$tag" >> "$tag_file" -fi +echo "$tag" >> "$tag_file" done +function update_mxnet_css { + tag=$1 + echo "Begin update fixes.." + # All fixes are done on the master branch of mxnet-incubator repository + # During a nightly build, these fixes will be patched to all the versions in the asf-site repository including the master folder under versions directory. + # copy + + #SIM: https://sim.amazon.com/issues/AWSDBUX-19727 + echo "Copying mxnet.css from master branch to all versions" + cp "$MASTER_SOURCE_DIR/$STATIC_FILES_DIR/mxnet.css" "$built/versions/$tag/_static" + + echo "Update fixes complete.." +} + # Update the specified tags with the Versions dropdown for tag in $tag_list; do # This Python script is expecting the tag_list.txt and it will use that as the entries to populate + python AddVersion.py --root_url "$root_url" --file_path "$built/versions/$tag" --current_version "$tag" || exit 1 -if [ $tag != 'master' ] -then -python AddPackageLink.py --file_path "$built/versions/master/install/index.html" \ - --current_version "$tag" || exit 1 -fi +#if [ $tag != 'master' ] Review comment: Done 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] thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix
thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#discussion_r180206032 ## File path: docs/build_version_doc/update_all_version.sh ## @@ -68,31 +73,51 @@ fi # Write all version numbers into $tag_file for AddVersion.py to use later # Master is added by that script by default for tag in $tag_list; do -if [ $tag != 'master' ] -then -echo "$tag" >> "$tag_file" -fi +echo "$tag" >> "$tag_file" done +function update_mxnet_css { + tag=$1 + echo "Begin update fixes.." + # All fixes are done on the master branch of mxnet-incubator repository + # During a nightly build, these fixes will be patched to all the versions in the asf-site repository including the master folder under versions directory. + # copy + + #SIM: https://sim.amazon.com/issues/AWSDBUX-19727 Review comment: Done 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] thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix
thinksanky commented on a change in pull request #10286: [MXNET-250] CSS and .htaccess fix URL: https://github.com/apache/incubator-mxnet/pull/10286#discussion_r180203366 ## File path: docs/build_version_doc/build_all_version.sh ## @@ -68,6 +68,7 @@ for tag in $tag_list; do git checkout master -- docs/mxdoc.py fi git submodule update || exit 1 +git submodule update --init --recursive Review comment: I have tried updating it together, but its unstable. I have found this to be better and working consistently This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] thinksanky commented on issue #10241: Css and install fix
thinksanky commented on issue #10241: Css and install fix URL: https://github.com/apache/incubator-mxnet/pull/10241#issuecomment-379864208 closing this out. as a dup for another PR This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] thinksanky closed pull request #10241: Css and install fix
thinksanky closed pull request #10241: Css and install fix URL: https://github.com/apache/incubator-mxnet/pull/10241 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake of provenance: As this is a foreign pull request (from a fork), the diff is supplied below (as it won't show otherwise due to GitHub magic): diff --git a/docs/.htaccess b/docs/.htaccess new file mode 100644 index 000..76d0893a528 --- /dev/null +++ b/docs/.htaccess @@ -0,0 +1,5 @@ +RewriteEngine on +RewriteRule ^get_started/why_mxnet.html$ /faq/why_mxnet.html [R=301,L] +RewriteRule ^get_started.*$ /install/ [R=301,L] +RewriteRule ^how_to.*$ /faq/ [R=301,L] +RewriteRule ^api/python/symbol.html$ /api/python/symbol/symbol.html [R=301,L] diff --git a/docs/_static/mxnet.css b/docs/_static/mxnet.css index 6e116dbb745..a62ceb59890 100644 --- a/docs/_static/mxnet.css +++ b/docs/_static/mxnet.css @@ -131,9 +131,7 @@ img { .navbar .container { position: relative; -left: 20px; width: 100%; -align-content: center; } @media screen and (max-width: 510px) { @@ -393,7 +391,7 @@ div .burgerIcon a { } #plusMenu .dropdown-submenu>.dropdown-menu { -left: 138px; +left: 127px; top: 0; background: #fff; } diff --git a/docs/build_version_doc/AddVersion.py b/docs/build_version_doc/AddVersion.py index c4d088a4b0f..f4b8304006f 100755 --- a/docs/build_version_doc/AddVersion.py +++ b/docs/build_version_doc/AddVersion.py @@ -36,7 +36,6 @@ with open('tag_list.txt', 'r') as tag_file: for line in tag_file: tag_list.append(line.lstrip().rstrip()) -tag_list.append('master') version_str = '' \ @@ -57,9 +56,6 @@ for name in files: if not name.endswith('.html'): continue -if 'install' in path: -print("Skipping this path: {}".format(path)) -continue with open(os.path.join(path, name), 'r') as html_file: content = bs(html_file, 'html.parser') navbar = content.find(id="main-nav") @@ -86,25 +82,26 @@ outstr = outstr.replace('http://mxnet.io', 'https://mxnet.incubator.apache.org/' 'versions/%s' % (args.current_version)) -# Fix git clone and pip installation to specific tag -pip_pattern = ['', '-cu80', '-cu75', '-cu80mkl', '-cu75mkl', '-mkl'] -if args.current_version == 'master': -outstr = outstr.replace('git clone --recursive https://github.com/dmlc/mxnet', -'git clone --recursive https://github.com/apache/incubator-mxnet.git mxnet') -for trail in pip_pattern: -outstr = outstr.replace('pip install mxnet%s<' % (trail), -'pip install mxnet%s --pre<' % (trail)) -outstr = outstr.replace('pip install mxnet%s\n<' % (trail), -'pip install mxnet%s --pre\n<' % (trail)) -else: -outstr = outstr.replace('git clone --recursive https://github.com/dmlc/mxnet', -'git clone --recursive https://github.com/apache/incubator-mxnet.git mxnet ' -'--branch %s' % (args.current_version)) -for trail in pip_pattern: -outstr = outstr.replace('pip install mxnet%s<' % (trail), -'pip install mxnet%s==%s<' % (trail, args.current_version)) -outstr = outstr.replace('pip install mxnet%s\n<' % (trail), -'pip install mxnet%s==%s\n<' % (trail, args.current_version)) +if 'install' in path: +# Fix git clone and pip installation to specific tag +pip_pattern = ['', '-cu80', '-cu75', '-cu80mkl', '-cu75mkl', '-mkl'] +if args.current_version == 'master': +outstr = outstr.replace('git clone --recursive https://github.com/dmlc/mxnet', +'git clone --recursive https://github.com/apache/incubator-mxnet.git mxnet') +for trail in pip_pattern: +outstr = outstr.replace('pip install mxnet%s<' % (trail), +'pip install mxnet%s --pre<' % (trail)) +outstr = outstr.replace('pip install mxnet%s\n<' % (trail), +'pip install mxnet%s --pre\n<' % (trail)) +else: +outstr = outstr.replace('git clone --recursive https://github.com/dmlc/mxnet', +'git clone --recursive
[GitHub] cjolivier01 commented on a change in pull request #10470: CUDNN not training on backward pass similar to pytorch
cjolivier01 commented on a change in pull request #10470: CUDNN not training on backward pass similar to pytorch URL: https://github.com/apache/incubator-mxnet/pull/10470#discussion_r180201171 ## File path: tests/python/gpu/test_operator_gpu.py ## @@ -1786,7 +1787,23 @@ def test_incorrect_gpu(): # Try setting dev_id to a really big number assert_raises(MXNetError, mx.nd.ones, (2,2), ctx=mx.gpu(11)) +def test_batchnorm_backwards_notrain(): Review comment: done 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 a change in pull request #10429: [MXNET-210]give warning for variables with same name in graph visualization
sandeep-krishnamurthy commented on a change in pull request #10429: [MXNET-210]give warning for variables with same name in graph visualization URL: https://github.com/apache/incubator-mxnet/pull/10429#discussion_r180199629 ## File path: python/mxnet/visualization.py ## @@ -252,6 +252,10 @@ def plot_network(symbol, title="plot", save_format='pdf', shape=None, node_attrs shape_dict = dict(zip(interals.list_outputs(), out_shapes)) conf = json.loads(symbol.tojson()) nodes = conf["nodes"] +# check if multiple nodes have the same name +if len(nodes) != len(set([node["name"] for node in nodes])): +logging.warning("There are multiple variables with the same name in your graph, " Review comment: It would good to print the duplicate name in your warning message. 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 #10470: CUDNN not training on backward pass similar to pytorch
marcoabreu commented on a change in pull request #10470: CUDNN not training on backward pass similar to pytorch URL: https://github.com/apache/incubator-mxnet/pull/10470#discussion_r180197804 ## File path: tests/python/gpu/test_operator_gpu.py ## @@ -1786,7 +1787,23 @@ def test_incorrect_gpu(): # Try setting dev_id to a really big number assert_raises(MXNetError, mx.nd.ones, (2,2), ctx=mx.gpu(11)) +def test_batchnorm_backwards_notrain(): Review comment: Please add with seed decorator 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 #8560: Upsampling: lack of arguments crashes the kernel
zhanghang1989 commented on issue #8560: Upsampling: lack of arguments crashes the kernel URL: https://github.com/apache/incubator-mxnet/issues/8560#issuecomment-379859176 @szha This issue can be closed in favor of https://github.com/apache/incubator-mxnet/pull/9688 ```python import mxnet as mx x1 = mx.nd.ones(shape=(2,3,4,4)) y1 = mx.nd.contrib.BilinearResize2D(x1, out_height=5, out_width=5) ``` 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 #4134: how to use mx.sym.UpSampling for bilinear upsampling
zhanghang1989 commented on issue #4134: how to use mx.sym.UpSampling for bilinear upsampling URL: https://github.com/apache/incubator-mxnet/issues/4134#issuecomment-379859043 @szha This issue can be closed in favor of https://github.com/apache/incubator-mxnet/pull/9688 ```python import mxnet as mx x1 = mx.nd.ones(shape=(2,3,4,4)) y1 = mx.nd.contrib.BilinearResize2D(x1, out_height=5, out_width=5) ``` 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