[GitHub] [incubator-mxnet] szha commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap

[GitBox]

szha commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap
URL: 
https://github.com/apache/incubator-mxnet/issues/16167#issuecomment-532066487
 
 
   @pengzhao-intel a tentative target date is by end of Q1 2020.
   
   @zachgk we will create a branch for 2.0. Initially we will keep master to be 
1.x and have 2.0 in a new branch. After 1.6 release we will revisit how to make 
the 2.0 branch the master.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16131: Fix for duplicate subgraph inputs/outputs

[GitBox]

ZhennanQin commented on a change in pull request #16131: Fix for duplicate 
subgraph inputs/outputs
URL: https://github.com/apache/incubator-mxnet/pull/16131#discussion_r324979243
 
 

 ##
 File path: src/operator/subgraph/build_subgraph.cc
 ##
 @@ -450,17 +450,18 @@ void FindInputEntries(const nnvm::Graph& g,
  * \param simple_nods vector of simple nodes in top sorted order
  * \param subgraph_nodes vector of pointers of simples of a subgraph.
  * \param entry_top_order_map mapping entry pointer to its top sorted position
- * \param output_entries output entries of the subgraph
+ * \param output_map output entries of the subgraph
  */
 void FindOutputEntries(nnvm::Graph* g,
const std::vector& simple_nodes,
const std::vector& subgraph_nodes,
const std::unordered_map&
  entry_top_order_map,
-   std::vector* output_entries) {
+   std::map > 
_map) {
 
 Review comment:
   Use `std::vector>` instead. This is fast(O(1) 
for search and push_back), and can avoid element missing(for std::map, we need 
additional check to ensure that for each i < output_map.size(), output_map[i] 
exists).


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532056594
 
 
   thanks this has been very helpful.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532055192
 
 
   for bias, you can try the first solution first(NDArray save), if it doesn't 
work, then access with pointer will always work.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532054969
 
 
   and for bias also we can do the same thing?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532054263
 
 
   
https://github.com/ZhennanQin/incubator-mxnet/commit/5276bfac3962fb9511482ac7232a0981f97d9b8e
 is added to show how to print s8 convolution weights.
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

shoubhik commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532051643
 
 
   @ZhennanQin in your previous comment for s8 it should be?
   
   `int8_t* ptr = 
static_cast(cached_weight_.GetMKLDNNData()->get_data_handle())`
   
   and for u8 case in your code snippet you have written
   
   ```
   MKLDNNStream::Get()->Submit();
   
   data[0] = cached_weight_.Reorder2Default();
   ```
   
   for s8 should it be
   ```
   MKLDNNStream::Get()->Submit();
   
   `int8_t* ptr = 
static_cast(cached_weight_.GetMKLDNNData()->get_data_handle())
   ```


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16088: remove 'foo' and other print msg from test

[GitBox]

Vikas-kum commented on issue #16088: remove 'foo' and other print msg from test
URL: https://github.com/apache/incubator-mxnet/pull/16088#issuecomment-532048956
 
 
   @apeforest This PR is causing tests to fail . 
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Fwindows-cpu/detail/master/1030/pipeline/
 
   
   ==
   
   FAIL: test_profiler.test_profile_task
   
   --
   
   Traceback (most recent call last):
   
 File "C:\Python37\lib\site-packages\nose\case.py", line 198, in runTest
   
   self.test(*self.arg)
   
 File 
"C:\jenkins_slave\workspace\ut-python-cpu@2\tests\python\unittest\test_profiler.py",
 line 115, in test_profile_task
   
   assert stop > start
   
   AssertionError
   
   
   ==
   
   FAIL: test_profiler.test_profile_tune_pause_resume
   
   --
   
   Traceback (most recent call last):
   
 File "C:\Python37\lib\site-packages\nose\case.py", line 198, in runTest
   
   self.test(*self.arg)
   
 File 
"C:\jenkins_slave\workspace\ut-python-cpu@2\tests\python\unittest\test_profiler.py",
 line 178, in test_profile_tune_pause_resume
   
   test_profile_task()
   
 File 
"C:\jenkins_slave\workspace\ut-python-cpu@2\tests\python\unittest\test_profiler.py",
 line 115, in test_profile_task
   
   assert stop > start
   
   AssertionError
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #15768: Fix gradient tensor mutate in `{adam/ftrl/rmprop/rmspropalex}_update`.

[GitBox]

Vikas-kum commented on issue #15768: Fix gradient tensor mutate in 
`{adam/ftrl/rmprop/rmspropalex}_update`.
URL: https://github.com/apache/incubator-mxnet/pull/15768#issuecomment-532046408
 
 
   @kshitij12345 @apeforest Test is failing for this change - 
   export MXNET_TEST_SEED=412298777 
   nosetests -v tests/python/unittest/test_ndarray.py:test_update_ops_mutation
   
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Funix-cpu/detail/master/1039/pipeline


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] shdyn commented on issue #16174: How to build new operators?

[GitBox]

shdyn commented on issue #16174: How to build new operators?
URL: 
https://github.com/apache/incubator-mxnet/issues/16174#issuecomment-532038005
 
 
   Thanks for your reply. There was no error in compiling. I also refered
   
https://github.com/reminisce/mxnet/blob/900a6997ded8f7124fc6323aa66408cb70e8253f/docs/how_to/add_op_in_backend.md
   
   On Tue, Sep 17, 2019 at 7:10 AM Zach Kimberg 
   wrote:
   
   > Did you receive any kind of error when you tried to compile mxnet? Also,
   > the new operator FAQ might be able to help you
   > https://mxnet.incubator.apache.org/versions/master/faq/new_op.html?
   >
   > —
   > You are receiving this because you authored the thread.
   > Reply to this email directly, view it on GitHub
   > 
,
   > or mute the thread
   > 

   > .
   >
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

ZhennanQin commented on issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532027621
 
 
   Hi @shoubhik,
   Here's a demo code on how to dump mkldnn u8 convolution weights.
   
https://github.com/ZhennanQin/incubator-mxnet/commit/c775dfc4f444a013302420112cc9c214fc59cd00
   
   For s8 convolution, mkldnn doesn't support to convert its s8 weight back to 
default format, thus you can't directly save them with aobve code. But you can 
still access its s8 weight in nhwc format:
   ```
   uint8_t* ptr = 
static_cast(cached_weight_.GetMKLDNNData()->get_data_handle()).
   ```
   You can write a loop to visit each element of it and save into a file.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16131: Fix for duplicate subgraph inputs/outputs

[GitBox]

ZhennanQin commented on a change in pull request #16131: Fix for duplicate 
subgraph inputs/outputs
URL: https://github.com/apache/incubator-mxnet/pull/16131#discussion_r324950644
 
 

 ##
 File path: src/operator/subgraph/subgraph_property.h
 ##
 @@ -296,8 +296,20 @@ class SubgraphProperty {
*/
   virtual void ConnectSubgraphOutputs(const nnvm::NodePtr subgraph_node,
   std::vector* 
output_entries) const {
+// Collapse output_entries pointing to same NodeEntry
 
 Review comment:
   This solution is more clear! I like this change. Please go ahead to 
implement it. Thanks.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.

[anirudh2290]

This is an automated email from the ASF dual-hosted git repository.

anirudh2290 pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git


The following commit(s) were added to refs/heads/asf-site by this push:
 new 0f17b51  Bump the publish timestamp.
0f17b51 is described below

commit 0f17b5125df794046a61d36602eb3f8defc80a8c
Author: mxnet-ci 
AuthorDate: Tue Sep 17 01:34:02 2019 +

Bump the publish timestamp.
---
 date.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/date.txt b/date.txt
new file mode 100644
index 000..6c16a29
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Tue Sep 17 01:34:02 UTC 2019

[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #16131: Fix for duplicate subgraph inputs/outputs

[GitBox]

samskalicky commented on a change in pull request #16131: Fix for duplicate 
subgraph inputs/outputs
URL: https://github.com/apache/incubator-mxnet/pull/16131#discussion_r324948919
 
 

 ##
 File path: src/operator/subgraph/subgraph_property.h
 ##
 @@ -296,8 +296,20 @@ class SubgraphProperty {
*/
   virtual void ConnectSubgraphOutputs(const nnvm::NodePtr subgraph_node,
   std::vector* 
output_entries) const {
+// Collapse output_entries pointing to same NodeEntry
 
 Review comment:
   @ZhennanQin Would you approve changing the connectSubgraphOutputs API to 
pass in a map instead for output_entries? Then it will simplify the 
implementation of connectSubgraphOutputs and help backend developers avoid 
making bugs by having to dedupe in this weird way. 
   
   @HahTK and I are willing to make this change. 
   
   Heres what we propose:
   
   Change FindOutputEntries to populate a map > instead 
of the vector. The integer refers to the i'th output of the 
subgraph, and nodeentry pointer is the same as it is in output_entries. The job 
of connectSubgraphOutputs is to connect all nodeentries in the vector to that 
indexed output of the subgraph.
   
   Then in CreateSubgraphNode (in build_subgraph.cc) the code will look like:
   
   ```
   nnvm::Symbol sym;
 size_t idx = 0;
 sym.outputs.resize(output_map.size());
 for (size_t i = 0; i < output_map.size(); ++i) {
   sym.outputs[i] = *output_map[i][0];
 }
   ```
   
   And ConnectSubgraphOutputs will take the output_map and do:
   
   ```
   for (size_t i = 0; i < output_map->size(); ++i) {
  for(auto e : output_map[i]) {
*e = nnvm::NodeEntry{subgraph_node, static_cast(i), 0};
   }
   }
   ```
   
   This vastly simplifies/clarifies the code backend devs need to write. And 
helps avoid confusion and prevents bugs. 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wuxun-zhang opened a new pull request #16184: Add large tensor nightly tests for MKL-DNN operators

[GitBox]

wuxun-zhang opened a new pull request #16184: Add large tensor nightly tests 
for MKL-DNN operators
URL: https://github.com/apache/incubator-mxnet/pull/16184
 
 
   To track the correctness of MKL-DNN operators when switching to use int64 
tensor size, we added nightly tests script here. In this script, we specify the 
data type of input as float32 since mkl-dnn only supported float32 input data 
currently, which is different from existed large tensor test script (default is 
int64 input).
   
   @pengzhao-intel @TaoLv @apeforest @ChaiBapchya 
   
   
   ## 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 to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] R1ncy commented on issue #15320: Weird C++ Error / Bug when calling asnumpy() or exporting the weight of darknet53 while training

[GitBox]

R1ncy commented on issue #15320: Weird C++ Error / Bug when calling asnumpy() 
or exporting the weight of darknet53 while training 
URL: 
https://github.com/apache/incubator-mxnet/issues/15320#issuecomment-532016111
 
 
   the label of my class index is from 0 to 999, so the class number=1000. i 
changed the class number to 1001 and the problem was solved. @nacorti


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add MKL-DNN Convolution

[GitBox]

ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add 
MKL-DNN Convolution
URL: https://github.com/apache/incubator-mxnet/pull/16141#discussion_r324945129
 
 

 ##
 File path: src/operator/nn/mkldnn/mkldnn_convolution-inl.h
 ##
 @@ -79,54 +79,63 @@ struct MKLDNNConvFullParam {
   MKLDNNPostEltwiseParam postsum_act_param;
 };
 
-mkldnn::convolution_forward::primitive_desc GetConvFwdImpl(const 
MKLDNNConvFullParam ,
-   const bool is_train,
-   const NDArray ,
-   const NDArray 
,
-   const NDArray *bias,
-   const NDArray 
);
+std::shared_ptr GetConvFwdImpl(
+const ConvolutionParam , const bool is_train, const NDArray , 
const NDArray ,
+const NDArray *bias, const NDArray );
 
 class MKLDNNConvForward {
  public:
-  mkldnn::convolution_forward::primitive_desc fwd_pd;
-
   MKLDNNConvForward(const MKLDNNConvFullParam , const bool is_train, 
const NDArray ,
-const NDArray , const NDArray *bias, const NDArray 
);
+const NDArray , const NDArray *bias, const NDArray 
);
 
-  void SetNewMem(const mkldnn::memory , const mkldnn::memory ,
- const mkldnn::memory *bias, const mkldnn::memory );
+  const mkldnn::convolution_forward () const { return *fwd_; }
 
-  void SetNewMem(const mkldnn::memory , const mkldnn::memory ) {
-this->data_->set_data_handle(data.get_data_handle());
-this->out_->set_data_handle(output.get_data_handle());
-  }
-
-  const mkldnn::convolution_forward () const {
-return *fwd_;
-  }
+  const mkldnn::convolution_forward::primitive_desc () const { return 
*pd_; }
 
  private:
   std::shared_ptr fwd_;
-  std::shared_ptr data_;
-  std::shared_ptr weight_;
-  std::shared_ptr bias_;
-  std::shared_ptr out_;
+  std::shared_ptr pd_;
 };
 
 typedef ParamOpSign MKLDNNConvSignature;
 
-MKLDNNConvForward (const ConvolutionParam ,
-  const bool is_train, const NDArray ,
-  const NDArray , const NDArray *bias,
-  const NDArray );
-
 void MKLDNNConvolutionForwardFullFeature(const MKLDNNConvFullParam ,
  const OpContext ,
  MKLDNNConvForward *fwd,
  const std::vector _data,
  const std::vector ,
  const std::vector _data);
 
+void MKLDNNConvolutionForward(const nnvm::NodeAttrs ,
+  const OpContext ,
+  const std::vector _data,
+  const std::vector ,
+  const std::vector _data);
+
+class MKLDNNConvBackward {
+ public:
+  MKLDNNConvBackward(const MKLDNNConvFullParam , const NDArray , 
const NDArray ,
+ const NDArray *bias, const NDArray );
+
+  const mkldnn::convolution_backward_data () const { return 
*bwd_data_; }
+
+  const mkldnn::convolution_backward_weights () const { return 
*bwd_weight_; }
+
+  const mkldnn::convolution_backward_data::primitive_desc () const {
+return *bwd_data_pd_;
+  }
+
+  const mkldnn::convolution_backward_weights::primitive_desc () 
const {
+return *bwd_weights_pd_;
+  }
+
+ private:
+  std::shared_ptr 
bwd_data_pd_;
+  std::shared_ptr 
bwd_weights_pd_;
+  std::shared_ptr bwd_data_;
+  std::shared_ptr bwd_weight_;
+};
+
 
 Review comment:
   This is a difference between mkldnn and mxnet. mxnet prefer to use `weight` 
while mkldnn prefer `weights`. So I don't know which should be the boundary of 
mkldnn and mxnet. @rongzha1 You can make the decision.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add MKL-DNN Convolution

[GitBox]

ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add 
MKL-DNN Convolution
URL: https://github.com/apache/incubator-mxnet/pull/16141#discussion_r324945190
 
 

 ##
 File path: src/operator/nn/mkldnn/mkldnn_base-inl.h
 ##
 @@ -277,6 +277,19 @@ inline static mkldnn::memory::desc GetWeightDesc(const 
NDArray ,
   }
 }
 
+inline static const std::vector GetMKLDNNInputArray(const 
std::vector ) {
+  std::vector ret;
+  ret.reserve(inputs.size());
+  for (const auto  : inputs) {
+if (in.IsView() && in.IsMKLDNNData()) {
+  ret.push_back(in.Reorder2Default());
+} else {
+  ret.push_back(in);
 
 Review comment:
   Yes, we can do that.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add MKL-DNN Convolution

[GitBox]

ZhennanQin commented on a change in pull request #16141: [mkldnn-v1.0] Add 
MKL-DNN Convolution
URL: https://github.com/apache/incubator-mxnet/pull/16141#discussion_r324944841
 
 

 ##
 File path: src/operator/operator_common.h
 ##
 @@ -526,17 +526,46 @@ class OpSignature {
* and the layout to sign the op.
*/
 
-#if MXNET_USE_MKLDNN == 1
+#if MXNET_USE_MKLDNN == 100
   void AddSign(const mkldnn::memory ) {
-auto desc = mem.get_primitive_desc().desc();
-hash = hash * 2 + desc.data.format;
-eles.push_back(desc.data.format);
+auto desc = mem.get_desc();
+hash = hash * 2 + desc.data.format_kind;
+eles.push_back(desc.data.format_kind);
 hash = hash * 2 + desc.data.data_type;
 eles.push_back(desc.data.data_type);
 for (int i = 0; i < desc.data.ndims; i++) {
   hash = hash * 2 + desc.data.dims[i];
   eles.push_back(desc.data.dims[i]);
 }
+switch (desc.data.format_kind) {
 
 Review comment:
   OK. Ideally, we can save the hash result into MKLDNNMemory, only computing 
it when hash result doesn't exist.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] YouhuiBai commented on issue #15674: Straggler in latest mxnet when training with distributed parameter server

[GitBox]

YouhuiBai commented on issue #15674: Straggler in latest mxnet when training 
with distributed parameter server
URL: 
https://github.com/apache/incubator-mxnet/issues/15674#issuecomment-532010392
 
 
   @apeforest Very sorry, I forgot to check the gmail because it is uncommonly 
used. I used branch 1.5.0 ("current newest mxnet" as I said), and checked from 
version 1.0.0 to `master` branch, there is straggler. What can we do is just 
allocate the pinned memory for `comm_buf` rather than reuse the `merged`, which 
is pageable memory. Thanks a lot!


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16131: Fix for duplicate subgraph inputs/outputs

[GitBox]

ZhennanQin commented on a change in pull request #16131: Fix for duplicate 
subgraph inputs/outputs
URL: https://github.com/apache/incubator-mxnet/pull/16131#discussion_r324939057
 
 

 ##
 File path: src/operator/subgraph/subgraph_property.h
 ##
 @@ -296,8 +296,20 @@ class SubgraphProperty {
*/
   virtual void ConnectSubgraphOutputs(const nnvm::NodePtr subgraph_node,
   std::vector* 
output_entries) const {
+// Collapse output_entries pointing to same NodeEntry
 
 Review comment:
   @samskalicky Before this PR, backend still has the chance to remove the 
duplicated sym.outputs in CreateSubgraphNode(). The existing APIs are enough to 
solve the problem you're trying to fix, so I think this PR is to make all of 
this happen in a automatic way. But the change in connectSubgraphOutputs() 
doesn't achieve that goal, but making the solution semi-automatic. 
   
   From a backend developer position, I rather prefer to have full control of 
my code(current APIs works in this way), or making most of the code 
automatically(the way I suggest to change). For the semi-automatic solution, 
developer have to spend extra effort to figure out which parts are handled 
automatically(sym.outputs), and which parts are not(connectSubgraphOutputs). 
And also increased the change that backend developer to make a bug when 
overriding connectSubgraphOutputs(I can't find any reason for connecting 
entries with same value to different sym.outputs, but by mistake).
   
   Anyway, I don't against to make the change in your way. At least maybe we 
should add some comments in connectSubgraphOutputs to highlight the difference 
between sym.outputs and output_entries, and let developer notice that they need 
to handle the duplicated entires carefully?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated (2a55cd7 -> 3dacabe)

[anirudh2290]

This is an automated email from the ASF dual-hosted git repository.

anirudh2290 pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


from 2a55cd7  fixing test for model compatibility checker (#16159)
 add 3dacabe  Tutorials nighly fix (#16179)

No new revisions were added by this update.

Summary of changes:
 docs/tutorials/gluon/custom_layer.md  |  2 +-
 docs/tutorials/python/profiler.md |  3 +++
 docs/tutorials/sparse/train_gluon.md  |  2 +-
 tests/nightly/JenkinsfileForBinaries  |  6 +++---
 tests/tutorials/test_tutorials.py | 17 ++---
 tests/utils/notebook_test/__init__.py |  2 +-
 6 files changed, 19 insertions(+), 13 deletions(-)

[GitHub] [incubator-mxnet] anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix

[GitBox]

anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#discussion_r324938174
 
 

 ##
 File path: tests/utils/notebook_test/__init__.py
 ##
 @@ -31,7 +31,7 @@
 
 
 IPYTHON_VERSION = 4  # Pin to ipython version 4.
-TIME_OUT = 10*60  # Maximum 10 mins/test. Reaching timeout causes test failure.
+TIME_OUT = 15*60  # Maximum 10 mins/test. Reaching timeout causes test failure.
 
 Review comment:
   i am going to merge this PR to unblock nightlies. @access2rohit since this 
change is small, would you be able to add this change to one of your WIP PRs?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] anirudh2290 merged pull request #16179: Tutorials nighly fix

[GitBox]

anirudh2290 merged pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

pengzhao-intel commented on issue #16177: How to dump quantized weights from 
MKLDNN as Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-532003397
 
 
   Thanks @zhennanqin to follow up :)


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #16175: [Dataset] add shard API

[GitBox]

wkcn commented on a change in pull request #16175: [Dataset] add shard API
URL: https://github.com/apache/incubator-mxnet/pull/16175#discussion_r324928963
 
 

 ##
 File path: python/mxnet/gluon/data/dataset.py
 ##
 @@ -64,6 +64,37 @@ def filter(self, fn):
 from . import FilterSampler
 return _SampledDataset(self, FilterSampler(fn, self))
 
+def shard(self, num_shards, index):
+"""Returns a new dataset includes only 1/num_shards of this dataset.
+
+For distributed training, be sure to shard before you randomize the 
dataset
+(such as shuffle), if you want each worker to reach a unique subset.
+
+Parameters
+--
+num_shards : int
+A integer representing the number of data shards.
+index : int
+A integer representing the index of the current shard.
+
+Returns
+---
+Dataset
+The result dataset.
+"""
+assert index < num_shards, 'Shard index of out bound: %d out of 
%d'%(index, num_shards)
+assert num_shards > 0, 'Number of shards must be greater than 0'
+assert index >= 0, 'Index must be non-negative'
+length = len(self)
+shard_len = length // num_shards
+rest = length % num_shards
+# Compute the start index for this partition
+start = shard_len * index + min(index, rest)
 
 Review comment:
   It will distribute the samples more uniformly.
   For example, when there are 6 samples and 4 partitions, the partitions will 
get 2, 2, 1, 1 samples respectively.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16182: contrib.cond operator does not parameterized block execution

[GitBox]

zachgk commented on issue #16182: contrib.cond operator does not parameterized 
block execution
URL: 
https://github.com/apache/incubator-mxnet/issues/16182#issuecomment-531995373
 
 
   @junrushao1994 Can you take a look?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

zachgk commented on issue #16177: How to dump quantized weights from MKLDNN as 
Ndarray
URL: 
https://github.com/apache/incubator-mxnet/issues/16177#issuecomment-531993249
 
 
   ping @pengzhao-intel


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16174: How to build new operators?

[GitBox]

zachgk commented on issue #16174: How to build new operators?
URL: 
https://github.com/apache/incubator-mxnet/issues/16174#issuecomment-531992399
 
 
   Did you receive any kind of error when you tried to compile mxnet? Also, the 
new operator FAQ might be able to help you 
https://mxnet.incubator.apache.org/versions/master/faq/new_op.html?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap

[GitBox]

zachgk commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap
URL: 
https://github.com/apache/incubator-mxnet/issues/16167#issuecomment-531989453
 
 
   Is there a plan to create a branch either for the 1.x version and have 
master reflect 2.0 or to create a branch for the 2.0 version and keep master on 
1.x for now?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16162: [CI] [Flaky] [Test] Unix Cpu test failed on Clojure cases

[GitBox]

zachgk commented on issue #16162: [CI] [Flaky] [Test] Unix Cpu test failed on 
Clojure cases
URL: 
https://github.com/apache/incubator-mxnet/issues/16162#issuecomment-531987600
 
 
   That error means the CI machine was unable to connect to Maven central to 
download the package. We see it sometimes and are tracking the issue under 
https://github.com/apache/incubator-mxnet/issues/15605. 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16183: Fix README Build Status

[GitBox]

zachgk commented on issue #16183: Fix README Build Status
URL: https://github.com/apache/incubator-mxnet/pull/16183#issuecomment-531986305
 
 
   @aaronmarkham Ok. I figured out how to label the badges so the current 
version is 
https://github.com/zachgk/incubator-mxnet/blob/de80f2a0e27fc7af7489d0cb8fd2d3f190854d89/README.md.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #16131: Fix for duplicate subgraph inputs/outputs

[GitBox]

samskalicky commented on a change in pull request #16131: Fix for duplicate 
subgraph inputs/outputs
URL: https://github.com/apache/incubator-mxnet/pull/16131#discussion_r324915391
 
 

 ##
 File path: src/operator/subgraph/subgraph_property.h
 ##
 @@ -296,8 +296,20 @@ class SubgraphProperty {
*/
   virtual void ConnectSubgraphOutputs(const nnvm::NodePtr subgraph_node,
   std::vector* 
output_entries) const {
+// Collapse output_entries pointing to same NodeEntry
 
 Review comment:
   Hey Guys,
   
   I think we're getting offtrack here. Lets just clarify what we're working 
toward. We want the subgraph to only have unique outputs, so this code is 
necessary in creating the sym.outputs correctly:
   
   
https://github.com/apache/incubator-mxnet/pull/16131/files#diff-962e4e83b570943c087e477058485e25R578-R587
   
   Then, the problem is we need to wire up the those unique outputs to each 
entry in output_entries. Since we pass the new symbol "sym" into 
connectSubgraphOutputs indirectly via the node "n", technically the set of 
unique outputs from the subgraph is already available in the 
connectSubgraphOutputs function. 
   
   To clarify, the ideal solution would be to pass in a map from output node to 
output entry and then let connectSubgraphOutputs just wire them up as it sees 
fit. 
   
   Do we agree that we do not want to change the API of connectSubgraphOutputs 
to enable this? Is there any reason not to just do this now? it shouldnt be too 
much work to go change, there is only a few subgraph properties that would need 
to change:
   
   
https://github.com/apache/incubator-mxnet/search?q=connectSubgraphOutputs_q=connectSubgraphOutputs
   
   And I think the creators of those wouldnt mind just fixing this once and for 
all now. This would go in the 1.6 release. 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] access2rohit commented on issue #16179: Tutorials nighly fix

[GitBox]

access2rohit commented on issue #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#issuecomment-531984632
 
 
   one nitpicky comment rest LGTM!


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #16179: Tutorials nighly fix

[GitBox]

access2rohit commented on a change in pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#discussion_r324913952
 
 

 ##
 File path: tests/utils/notebook_test/__init__.py
 ##
 @@ -31,7 +31,7 @@
 
 
 IPYTHON_VERSION = 4  # Pin to ipython version 4.
-TIME_OUT = 10*60  # Maximum 10 mins/test. Reaching timeout causes test failure.
+TIME_OUT = 15*60  # Maximum 10 mins/test. Reaching timeout causes test failure.
 
 Review comment:
   nit: can you update the comment to max 15 mins/ test


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ChaiBapchya commented on a change in pull request #16155: set fixed seed for profiler

[GitBox]

ChaiBapchya commented on a change in pull request #16155: set fixed seed for 
profiler
URL: https://github.com/apache/incubator-mxnet/pull/16155#discussion_r324912567
 
 

 ##
 File path: benchmark/opperf/README.md
 ##
 @@ -173,6 +173,22 @@ This utility queries MXNet operator registry to fetch all 
operators registered w
 However, fully automated tests are enabled only for simpler operators such as 
- broadcast operators, element_wise operators etc... For the purpose of 
readability and giving more control to the users, complex operators such as 
convolution (2D, 3D), Pooling, Recurrent are not fully automated but expressed 
as default rules.
 See `utils/op_registry_utils.py` for more details.
 
+## Use nativa python timer
 
 Review comment:
   nitpick: native (also acc to our naming convention we have 2 parameters
   `native` and `python` where (`native` resembles MXNet's CPP pased profiler) 
and `python` indicates the python's time function. So I hope people don't get 
confused between native and native python


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ChaiBapchya commented on a change in pull request #16155: set fixed seed for profiler

[GitBox]

ChaiBapchya commented on a change in pull request #16155: set fixed seed for 
profiler
URL: https://github.com/apache/incubator-mxnet/pull/16155#discussion_r324912567
 
 

 ##
 File path: benchmark/opperf/README.md
 ##
 @@ -173,6 +173,22 @@ This utility queries MXNet operator registry to fetch all 
operators registered w
 However, fully automated tests are enabled only for simpler operators such as 
- broadcast operators, element_wise operators etc... For the purpose of 
readability and giving more control to the users, complex operators such as 
convolution (2D, 3D), Pooling, Recurrent are not fully automated but expressed 
as default rules.
 See `utils/op_registry_utils.py` for more details.
 
+## Use nativa python timer
 
 Review comment:
   nitpick: native (also acc to our naming convention we have 2 parameters
   `native` and `python` where (`native` resembles MXNet's CPP pased profiler) 
and `python` indicates the python's time function


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk closed issue #16146: Website build on CI is down due to Julia error

[GitBox]

zachgk closed issue #16146: Website build on CI is down due to Julia error
URL: https://github.com/apache/incubator-mxnet/issues/16146
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16146: Website build on CI is down due to Julia error

[GitBox]

zachgk commented on issue #16146: Website build on CI is down due to Julia error
URL: 
https://github.com/apache/incubator-mxnet/issues/16146#issuecomment-531982522
 
 
   Since #16147 was raised that seems to address this issue and the website 
publishing has been successful for the last 15 runs, I will close this issue. 
Please reopen if I am closing in error.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk opened a new pull request #16183: Fix README Build Status

[GitBox]

zachgk opened a new pull request #16183: Fix README Build Status
URL: https://github.com/apache/incubator-mxnet/pull/16183
 
 
   ## Description ##
   We display a build status at the top of the README between the mxnet logo 
and the Portable, Efficient, Scalable graphics. It is currently broken due to 
splitting the CI validation into multiple steps. This PR displays a build 
status for each of the steps as I don't know if Jenkins has a combined view.
   
   Old Broken - 
https://github.com/apache/incubator-mxnet/blob/2a55cd7d1801e6ed801057cac9018b707fe3e27f/README.md
   New - 
https://github.com/zachgk/incubator-mxnet/blob/c62becbc73d74ae066fb1857a7aa3140a38bfb07/README.md
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [X] Changes are complete (i.e. I finished coding on this PR)
   - [X] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [X] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [X] To the my best knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16181: disabled 3 tutorials tests

[GitBox]

Vikas-kum commented on issue #16181: disabled 3 tutorials tests
URL: 
https://github.com/apache/incubator-mxnet/issues/16181#issuecomment-531979949
 
 
   @sad- Thanks. I tried that but doesn't look like profiler tests were 
passing- 
   
   Looks like there is more to fix here - New error that came here was - 
   ```
   MXNetError: [19:29:32] 
/work/mxnet/3rdparty/dmlc-core/include/dmlc/thread_group.h:227: Check failed: 
auto_remove_ == false (1 vs. 0) : 
   
   Stack trace:
   
 [bt] (0) 
/work/mxnet/python/mxnet/../../lib/libmxnet.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x32)
 [0x7effabb7aed2]
   
 [bt] (1) 
/work/mxnet/python/mxnet/../../lib/libmxnet.so(dmlc::ThreadGroup::Thread::joinable()
 const+0xf4) [0x7effae507734]
   
 [bt] (2) 
/work/mxnet/python/mxnet/../../lib/libmxnet.so(mxnet::profiler::Profiler::SetContinuousProfileDump(bool,
 float)+0x108) [0x7effae504d28]
   
 [bt] (3) 
/work/mxnet/python/mxnet/../../lib/libmxnet.so(mxnet::profiler::Profiler::SetConfig(int,
 std::__cxx11::basic_string, std::allocator 
>, bool, float, bool)+0x95) [0x7effae505b95]
   
 [bt] (4) 
/work/mxnet/python/mxnet/../../lib/libmxnet.so(MXSetProcessProfilerConfig+0x3d6)
 [0x7effaecb80a6]
   
 [bt] (5) 
/usr/lib/python3.5/lib-dynload/_ctypes.cpython-35m-x86_64-linux-gnu.so(ffi_call_unix64+0x4c)
 [0x7efffb8bae20]
   
 [bt] (6) 
/usr/lib/python3.5/lib-dynload/_ctypes.cpython-35m-x86_64-linux-gnu.so(ffi_call+0x2eb)
 [0x7efffb8ba88b]
   
 [bt] (7) 
/usr/lib/python3.5/lib-dynload/_ctypes.cpython-35m-x86_64-linux-gnu.so(_ctypes_callproc+0x49a)
 [0x7efffb8b501a]
   
 [bt] (8) 
/usr/lib/python3.5/lib-dynload/_ctypes.cpython-35m-x86_64-linux-gnu.so(+0x9fcb) 
[0x7efffb8a8fcb]
   
   ```
   
   Logs for reference - 
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/NightlyTestsForBinaries/detail/tutorials_nighly_fix/10/pipeline
   
   Can you please try to provide fixes for 3 tests and then we can enable the 
tests in nightly. MKLDNN is mostly using wrong binary. (Currently using GPU 
binary without mkldnn libraries.)


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on a change in pull request #16179: Tutorials nighly fix

[GitBox]

Vikas-kum commented on a change in pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#discussion_r324903596
 
 

 ##
 File path: tests/nightly/JenkinsfileForBinaries
 ##
 @@ -39,7 +39,7 @@ core_logic: {
   utils.pack_lib('gpu', mx_lib)
 }
   }
-},
+}/*,
 
 Review comment:
   https://github.com/apache/incubator-mxnet/pull/16164 -- This PR is for 
re-enabling large tensor test. @access2rohit will be enabling the test in his 
PR.  


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated (692f49f -> 2a55cd7)

[anirudh2290]

This is an automated email from the ASF dual-hosted git repository.

anirudh2290 pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


from 692f49f  Sequence last fix (#16156)
 add 2a55cd7  fixing test for model compatibility checker (#16159)

No new revisions were added by this update.

Summary of changes:
 tests/nightly/model_backwards_compatibility_check/JenkinsfileForMBCC | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

[GitHub] [incubator-mxnet] zachgk commented on issue #16121: Using custom gcc version

[GitBox]

zachgk commented on issue #16121: Using custom gcc version
URL: 
https://github.com/apache/incubator-mxnet/issues/16121#issuecomment-531972844
 
 
   Hi @mahmoodn. If you are using the MAKE base build instead of the CMake one, 
then you can pass in a new value for CXX through `make CXX=xxx ...`. You can 
also use bash right before the command such as `CXX=xxx someBashCommand...`. 
   
   Can you explain what your use case is, how you are building mxnet, and what 
prevents you from passing in the gcc version through the CC and CXX environment 
variables?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] anirudh2290 merged pull request #16159: fixing test for model compatibility checker

[GitBox]

anirudh2290 merged pull request #16159: fixing test for model compatibility 
checker
URL: https://github.com/apache/incubator-mxnet/pull/16159
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] igolan opened a new issue #16182: contrib.cond operator does not parameterized block execution

[GitBox]

igolan opened a new issue #16182: contrib.cond operator does not parameterized 
block execution
URL: https://github.com/apache/incubator-mxnet/issues/16182
 
 
   ## Description
   ``contrib.cond`` operator does not support parameterized block execution 
(cannot infer shape).
   
   ## Environment info (Required)
   
   ```
   --Python Info--
   Version  : 3.7.4
   Compiler : Clang 10.0.1 (clang-1001.0.46.4)
   Build: ('default', 'Jul  9 2019 18:13:23')
   Arch : ('64bit', '')
   Pip Info---
   Version  : 19.0.3
   Directory: 
/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/pip-19.0.3-py3.7.egg/pip
   --MXNet Info---
   Version  : 1.5.0
   Directory: 
/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet
   Commit Hash   : 75a9e187d00a8b7ebc71412a02ed0e3ae489d91f
   Library  : 
['/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/libmxnet.so']
   Build features:
   ✖ CUDA
   ✖ CUDNN
   ✖ NCCL
   ✖ CUDA_RTC
   ✖ TENSORRT
   ✔ CPU_SSE
   ✔ CPU_SSE2
   ✔ CPU_SSE3
   ✔ CPU_SSE4_1
   ✔ CPU_SSE4_2
   ✖ CPU_SSE4A
   ✔ CPU_AVX
   ✖ CPU_AVX2
   ✖ OPENMP
   ✖ SSE
   ✖ F16C
   ✖ JEMALLOC
   ✖ BLAS_OPEN
   ✖ BLAS_ATLAS
   ✖ BLAS_MKL
   ✖ BLAS_APPLE
   ✔ LAPACK
   ✖ MKLDNN
   ✔ OPENCV
   ✖ CAFFE
   ✖ PROFILER
   ✔ DIST_KVSTORE
   ✖ CXX14
   ✖ INT64_TENSOR_SIZE
   ✔ SIGNAL_HANDLER
   ✖ DEBUG
   --System Info--
   Platform : Darwin-18.7.0-x86_64-i386-64bit
   system   : Darwin
   node : XXX
   release  : 18.7.0
   version  : Darwin Kernel Version 18.7.0: Tue Aug 20 16:57:14 PDT 2019; 
root:xnu-4903.271.2~2/RELEASE_X86_64
   --Hardware Info--
   machine  : x86_64
   processor: i386
   b'machdep.cpu.brand_string: Intel(R) Core(TM) i7-7660U CPU @ 2.50GHz'
   b'machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE 
MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ 
DTES64 MON DSCPL VMX SMX EST TM2 SSSE3 FMA CX16 TPR PDCM SSE4.1 SSE4.2 x2APIC 
MOVBE POPCNT AES PCID XSAVE OSXSAVE SEGLIM64 TSCTMR AVX1.0 RDRAND F16C'
   b'machdep.cpu.leaf7_features: RDWRFSGS TSC_THREAD_OFFSET SGX BMI1 HLE AVX2 
SMEP BMI2 ERMS INVPCID RTM FPU_CSDS MPX RDSEED ADX SMAP CLFSOPT IPT MDCLEAR 
TSXFA IBRS STIBP L1DF SSBD'
   b'machdep.cpu.extfeatures: SYSCALL XD 1GBPAGE EM64T LAHF LZCNT PREFETCHW 
RDTSCP TSCI'
   --Network Test--
   Setting timeout: 10
   Timing for MXNet: https://github.com/apache/incubator-mxnet, DNS: 0.0137 
sec, LOAD: 0.5112 sec.
   Timing for Gluon Tutorial(en): http://gluon.mxnet.io, DNS: 0.0180 sec, LOAD: 
0.4525 sec.
   Timing for Gluon Tutorial(cn): https://zh.gluon.ai, DNS: 0.0198 sec, LOAD: 
0.8612 sec.
   Timing for FashionMNIST: 
https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/dataset/fashion-mnist/train-labels-idx1-ubyte.gz,
 DNS: 0.0233 sec, LOAD: 0.1894 sec.
   Timing for PYPI: https://pypi.python.org/pypi/pip, DNS: 0.0120 sec, LOAD: 
0.3173 sec.
   Timing for Conda: https://repo.continuum.io/pkgs/free/, DNS: 0.0105 sec, 
LOAD: 0.0961 sec.
   --Environment--
   
   ```
   
   I'm using Pyton
   
   ## Build info (Required if built from source)
   N/A
   
   ## Error Message:
   ```
   Traceback (most recent call last):
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/block.py",
 line 811, in _call_cached_op
   for is_arg, i in self._cached_op_args]
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/block.py",
 line 811, in 
   for is_arg, i in self._cached_op_args]
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/parameter.py",
 line 543, in data
   return self._check_and_get(self._data, ctx)
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/parameter.py",
 line 234, in _check_and_get
   "num_features, etc., for network layers."%(self.name))
   mxnet.gluon.parameter.DeferredInitializationError: Parameter 
'mlp0_dense0_weight' has not been initialized yet because initialization was 
deferred. Actual initialization happens during the first forward pass. Please 
pass one batch of data through the network before accessing Parameters. You can 
also avoid deferred initialization by specifying in_units, num_features, etc., 
for network layers.
   
   During handling of the above exception, another exception occurred:
   
   Traceback (most recent call last):
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/block.py",
 line 797, in _deferred_infer_shape
   self.infer_shape(*args)
 File 
"/Users/XX/PycharmProjects/XX/venv/lib/python3.7/site-packages/mxnet/gluon/block.py",
 line 870, in infer_shape
   self._infer_attrs('infer_shape', 'shape', *args)
 File 

[GitHub] [incubator-mxnet] sad- commented on issue #16181: disabled 3 tutorials tests

[GitBox]

sad- commented on issue #16181: disabled 3 tutorials tests
URL: 
https://github.com/apache/incubator-mxnet/issues/16181#issuecomment-531967993
 
 
   profiler fix should be in this pr: 
https://github.com/apache/incubator-mxnet/pull/16160
   
   But yeah we can disable until after the fixes for mkldnn_quantization and 
gluon_performance are done. I think we may have to whitelist 
mkldnn_quantization from the test suite actually in the long run.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16181: disabled 3 tutorials tests

[GitBox]

Vikas-kum commented on issue #16181: disabled 3 tutorials tests
URL: 
https://github.com/apache/incubator-mxnet/issues/16181#issuecomment-531967661
 
 
   @mxnet-label-bot [Test]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix

[GitBox]

anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#discussion_r324895908
 
 

 ##
 File path: tests/tutorials/test_tutorials.py
 ##
 @@ -114,10 +113,10 @@ def test_gluon_save_load_params():
 
 def test_gluon_hybrid():
 assert _test_tutorial_nb('gluon/hybrid')
-
+"""
 
 Review comment:
   can we add a comment linking the related issue here. same for other 
commented code.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix

[GitBox]

anirudh2290 commented on a change in pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#discussion_r324895780
 
 

 ##
 File path: tests/nightly/JenkinsfileForBinaries
 ##
 @@ -39,7 +39,7 @@ core_logic: {
   utils.pack_lib('gpu', mx_lib)
 }
   }
-},
+}/*,
 
 Review comment:
   can we add a comment tagging the related issue here.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16159: fixing test for model compatibility checker

[GitBox]

Vikas-kum commented on issue #16159: fixing test for model compatibility checker
URL: https://github.com/apache/incubator-mxnet/pull/16159#issuecomment-531966484
 
 
   @anirudh2290 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16179: Tutorials nighly fix

[GitBox]

Vikas-kum commented on issue #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#issuecomment-531966172
 
 
   github issue created - 
https://github.com/apache/incubator-mxnet/issues/16181 for reenabling  tests 
once we have fixes.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum commented on issue #16179: Tutorials nighly fix

[GitBox]

Vikas-kum commented on issue #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179#issuecomment-531965545
 
 
   @anirudh2290 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16181: disabled 3 tutorials tests

[GitBox]

mxnet-label-bot commented on issue #16181: disabled 3 tutorials tests
URL: 
https://github.com/apache/incubator-mxnet/issues/16181#issuecomment-531965494
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Test


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum opened a new issue #16181: disabled 3 tutorials tests

[GitBox]

Vikas-kum opened a new issue #16181: disabled 3 tutorials tests
URL: https://github.com/apache/incubator-mxnet/issues/16181
 
 
   Nighltly tests are failing due to some tutorial tests. We fixed some. 
   3 tests were disabled form this file : tests/tutorials/test_tutorials.py 
   test_gluon_performance 
   test_python_profiler
   test_mkldnn_quantization 
   
   We need to uncomment the test after fixes are done in tutorials.
   
   https://github.com/apache/incubator-mxnet/pull/16179/files


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.

[anirudh2290]

This is an automated email from the ASF dual-hosted git repository.

anirudh2290 pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git


The following commit(s) were added to refs/heads/asf-site by this push:
 new 4dc2959  Bump the publish timestamp.
4dc2959 is described below

commit 4dc2959f3db34cc2f4ede494d8df1f6c709d09b0
Author: mxnet-ci 
AuthorDate: Mon Sep 16 21:15:55 2019 +

Bump the publish timestamp.
---
 date.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/date.txt b/date.txt
new file mode 100644
index 000..23fcfa4
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Mon Sep 16 21:15:55 UTC 2019

[GitHub] [incubator-mxnet] marcoabreu commented on a change in pull request #16175: [Dataset] add shard API

[GitBox]

marcoabreu commented on a change in pull request #16175: [Dataset] add shard API
URL: https://github.com/apache/incubator-mxnet/pull/16175#discussion_r324882558
 
 

 ##
 File path: tests/python/unittest/test_gluon_data.py
 ##
 @@ -296,6 +296,24 @@ def test_dataset_filter():
 for idx, sample in enumerate(a_xform_filtered):
 assert sample % 10 == 0
 
+def test_dataset_shard():
+length = 9
+a = mx.gluon.data.SimpleDataset([i for i in range(length)])
+shard_0 = a.shard(4, 0)
+shard_1 = a.shard(4, 1)
+shard_2 = a.shard(4, 2)
+shard_3 = a.shard(4, 3)
+assert len(shard_0) + len(shard_1) + len(shard_2) + len(shard_3) == length
+assert len(shard_0) > 0
 
 Review comment:
   Use == instead since the size should be known?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] frankfliu closed pull request #15946: Cherry pick scala fix into 1.5.x branch

[GitBox]

frankfliu closed pull request #15946: Cherry pick scala fix into 1.5.x branch
URL: https://github.com/apache/incubator-mxnet/pull/15946
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] frankfliu closed pull request #14249: Update inference sample make file.

[GitBox]

frankfliu closed pull request #14249: Update inference sample make file.
URL: https://github.com/apache/incubator-mxnet/pull/14249
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated (7fc1d84 -> 692f49f)

[apeforest]

This is an automated email from the ASF dual-hosted git repository.

apeforest pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


from 7fc1d84  Adds dynamic libmxnet to CD pipeline (#16163)
 add 692f49f  Sequence last fix (#16156)

No new revisions were added by this update.

Summary of changes:
 src/operator/sequence_last-inl.h   | 12 ++--
 src/operator/sequence_last.cc  | 10 --
 tests/nightly/test_large_vector.py |  4 +++-
 3 files changed, 17 insertions(+), 9 deletions(-)

[GitHub] [incubator-mxnet] apeforest merged pull request #16156: Sequence last fix

[GitBox]

apeforest merged pull request #16156: Sequence last fix
URL: https://github.com/apache/incubator-mxnet/pull/16156
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] sxjscience commented on issue #16180: Add exception check for numpy reshape

[GitBox]

sxjscience commented on issue #16180: Add exception check for numpy reshape
URL: https://github.com/apache/incubator-mxnet/pull/16180#issuecomment-531944224
 
 
   @samskalicky @zachgk 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] sxjscience opened a new pull request #16180: Add exception check for numpy reshape

[GitBox]

sxjscience opened a new pull request #16180: Add exception check for numpy 
reshape
URL: https://github.com/apache/incubator-mxnet/pull/16180
 
 
   ## Description ##
   
   Try to add exception to avoid error-message downgrade issues reported in 
https://github.com/apache/incubator-mxnet/issues/16102
   
   ## 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 ###
   
   ## 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 to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16168: group norm tests are failing

[GitBox]

mxnet-label-bot commented on issue #16168: group norm tests are failing
URL: 
https://github.com/apache/incubator-mxnet/issues/16168#issuecomment-531941818
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Test


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16173: Saving and loading cudNN autotune and graph optimization

[GitBox]

mxnet-label-bot commented on issue #16173: Saving and loading cudNN autotune 
and graph optimization
URL: 
https://github.com/apache/incubator-mxnet/issues/16173#issuecomment-531941860
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Cuda, Feature


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16176: When convert mxnet model (simple pose) to onnx , the deconvolution’s convert function miss?

[GitBox]

mxnet-label-bot commented on issue #16176: When convert mxnet model (simple 
pose) to onnx , the deconvolution’s convert function miss?
URL: 
https://github.com/apache/incubator-mxnet/issues/16176#issuecomment-531941783
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): ONNX, Bug


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16174: How to build new operators?

[GitBox]

mxnet-label-bot commented on issue #16174: How to build new operators?
URL: 
https://github.com/apache/incubator-mxnet/issues/16174#issuecomment-531941549
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Build


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16162: [CI] [Flaky] [Test] Unix Cpu test failed on Clojure cases

[GitBox]

mxnet-label-bot commented on issue #16162: [CI] [Flaky] [Test] Unix Cpu test 
failed on Clojure cases
URL: 
https://github.com/apache/incubator-mxnet/issues/16162#issuecomment-531941629
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Test, Flaky, CI


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap

[GitBox]

mxnet-label-bot commented on issue #16167: [RFC] Apache MXNet 2.0 Roadmap
URL: 
https://github.com/apache/incubator-mxnet/issues/16167#issuecomment-531940931
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended label(s): Feature


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zachgk commented on issue #16108: Subgraph API creates duplicate inputs and outputs

[GitBox]

zachgk commented on issue #16108: Subgraph API creates duplicate inputs and 
outputs
URL: 
https://github.com/apache/incubator-mxnet/issues/16108#issuecomment-531941216
 
 
   @mxnet-label-bot add [Performance]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated (5ed5689 -> 7fc1d84)

[zachgk]

This is an automated email from the ASF dual-hosted git repository.

zachgk pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


from 5ed5689  numpy operator ravel, derive from reshape (#16016)
 add 7fc1d84  Adds dynamic libmxnet to CD pipeline (#16163)

No new revisions were added by this update.

Summary of changes:
 cd/Jenkinsfile_cd_pipeline |   8 +-
 cd/Jenkinsfile_release_job |   4 +-
 .../{static => dynamic}/Jenkins_pipeline.groovy|  16 ++--
 ci/docker/runtime_functions.sh | 101 +
 4 files changed, 117 insertions(+), 12 deletions(-)
 copy cd/mxnet_lib/{static => dynamic}/Jenkins_pipeline.groovy (70%)

[GitHub] [incubator-mxnet] zachgk merged pull request #16163: Adds dynamic libmxnet to CD pipeline

[GitBox]

zachgk merged pull request #16163: Adds dynamic libmxnet to CD pipeline
URL: https://github.com/apache/incubator-mxnet/pull/16163
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #15921: [WIP] dynamic custom operator support

[GitBox]

samskalicky commented on a change in pull request #15921: [WIP] dynamic custom 
operator support
URL: https://github.com/apache/incubator-mxnet/pull/15921#discussion_r324862929
 
 

 ##
 File path: include/mxnet/lib_api.h
 ##
 @@ -18,33 +18,601 @@
  */
 
 /*!
- * Copyright (c) 2015 by Contributors
+ * Copyright (c) 2019 by Contributors
  * \file lib_api.h
  * \brief APIs to interact with libraries
+ * This API specifies function prototypes to
+ * register custom ops for library authors
  */
+
 #ifndef MXNET_LIB_API_H_
 #define MXNET_LIB_API_H_
 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#define MX_LIBRARY_VERSION 1
+
+/*!
+ * \brief External Tensor data types
+ */
+enum MXDType {
+  kFloat32 = 0,
+  kFloat64 = 1,
+  kFloat16 = 2,
+  kUint8 = 3,
+  kInt32 = 4,
+  kInt8  = 5,
+  kInt64 = 6,
+};
+
+enum MXReturnValue {
+  MX_FAIL = 0,
+  MX_SUCCESS = 1,
+};
+
+/*!
+ * \brief External Tensor data structure
+ */
+struct MXTensor {
+  MXTensor() : data(nullptr) {}
+
+  MXTensor(void *data, const std::vector , MXDType dtype)
+  : data{data}, shape{shape}, dtype{dtype} {}
+
+  /*!
+   * \brief helper function to cast data pointer
+   */
+  template
+  data_type* getData() {
+return reinterpret_cast(data);
+  }
+
+  void *data;  // not owned
+  std::vector shape;
+  MXDType dtype;
+};
+
+/*!
+ * \brief resource malloc function to allocate memory inside Forward/Backward 
functions
+ */
+typedef void* (*xpu_malloc_t)(void*, int);
+
+/*!
+ * \brief Class to provide resource APIs to Forward/Backward functions
+ */
+class OpResource {
+ public:
+  OpResource(xpu_malloc_t xm, void* _xm) : xpu_malloc(xm), _xpu_malloc(_xm) {}
+
+  /*!
+   * \brief allocate memory controlled by MXNet
+   */
+  void* alloc(int size) {
+return xpu_malloc(_xpu_malloc, size);
+  }
+ private:
+  xpu_malloc_t xpu_malloc;
+  void* _xpu_malloc;
+};
+
+/*!
+ * \brief Macro to help passing serialized subgraph through attribute dict
+ */
+#define SUBGRAPH_SYM_JSON "subgraph_sym_json"
+
+/*!
+ * \brief An abstract class for library author creating stateful op
+ * custom library should override Forward and destructor, and has an
+ * option to implement Backward
+ */
+class CustomStatefulOp {
+ public:
+  virtual MXReturnValue Forward(std::vector inputs,
+std::vector outputs,
+OpResource op_res) = 0;
+  virtual MXReturnValue Backward(std::vector inputs,
+ std::vector outputs,
+ OpResource op_res) {
+std::cout << "Error! Operator does not support backward" << std::endl;
+return MX_FAIL;
+  }
+  virtual ~CustomStatefulOp() = 0;
 
 Review comment:
   lets not force the user to implement this. it also solves the weirdness 
around pure virtual destructor issue. I suggest we remove this line.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.

[anirudh2290]

This is an automated email from the ASF dual-hosted git repository.

anirudh2290 pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git


The following commit(s) were added to refs/heads/asf-site by this push:
 new 3ea5b8b  Bump the publish timestamp.
3ea5b8b is described below

commit 3ea5b8b7d87691e47f289ffecbad4f0ad10dc27c
Author: mxnet-ci 
AuthorDate: Mon Sep 16 19:42:39 2019 +

Bump the publish timestamp.
---
 date.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/date.txt b/date.txt
new file mode 100644
index 000..0b373f8
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Mon Sep 16 19:42:39 UTC 2019

[GitHub] [incubator-mxnet] Vikas-kum closed pull request #16159: fixing test for model compatibility checker

[GitBox]

Vikas-kum closed pull request #16159: fixing test for model compatibility 
checker
URL: https://github.com/apache/incubator-mxnet/pull/16159
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum opened a new pull request #16159: fixing test for model compatibility checker

[GitBox]

Vikas-kum opened a new pull request #16159: fixing test for model compatibility 
checker
URL: https://github.com/apache/incubator-mxnet/pull/16159
 
 
   ## Description ##
   adding libtvm.so to test
   
   ## 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 to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16106: src/executor/graph_executor.cc:1847: Check failed: arg_names.size() == in_args_map.size() (2 vs. 1)

[GitBox]

samskalicky commented on issue #16106: src/executor/graph_executor.cc:1847: 
Check failed: arg_names.size() == in_args_map.size() (2 vs. 1)
URL: 
https://github.com/apache/incubator-mxnet/issues/16106#issuecomment-531913166
 
 
   @zachgk assign [@samskalicky  ]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum opened a new pull request #16179: Tutorials nighly fix

[GitBox]

Vikas-kum opened a new pull request #16179: Tutorials nighly fix
URL: https://github.com/apache/incubator-mxnet/pull/16179
 
 
   ## Description ##
   Fixing tutorials test. 
   Disabling some tutorials test until we get a fix for them. Explicitly 
disabling gluon perf test, profiler test and mkldnn tutorial test
   
   
   ## 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 to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zhreshold commented on a change in pull request #16175: [Dataset] add shard API

[GitBox]

zhreshold commented on a change in pull request #16175: [Dataset] add shard API
URL: https://github.com/apache/incubator-mxnet/pull/16175#discussion_r324829139
 
 

 ##
 File path: python/mxnet/gluon/data/dataset.py
 ##
 @@ -64,6 +64,37 @@ def filter(self, fn):
 from . import FilterSampler
 return _SampledDataset(self, FilterSampler(fn, self))
 
+def shard(self, num_shards, index):
+"""Returns a new dataset includes only 1/num_shards of this dataset.
+
+For distributed training, be sure to shard before you randomize the 
dataset
+(such as shuffle), if you want each worker to reach a unique subset.
+
+Parameters
+--
+num_shards : int
+A integer representing the number of data shards.
+index : int
+A integer representing the index of the current shard.
+
+Returns
+---
+Dataset
+The result dataset.
+"""
+assert index < num_shards, 'Shard index of out bound: %d out of 
%d'%(index, num_shards)
+assert num_shards > 0, 'Number of shards must be greater than 0'
+assert index >= 0, 'Index must be non-negative'
+length = len(self)
+shard_len = length // num_shards
+rest = length % num_shards
+# Compute the start index for this partition
+start = shard_len * index + min(index, rest)
 
 Review comment:
   don't quite get it, why start and end are computed like this?


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16102: Usability degradation

[GitBox]

samskalicky commented on issue #16102: Usability degradation
URL: 
https://github.com/apache/incubator-mxnet/issues/16102#issuecomment-531912115
 
 
   @zachgk assign [@sxjscience]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16099: [Bug] NDArray asscalar error after set npx.set_np()

[GitBox]

samskalicky commented on issue #16099: [Bug] NDArray asscalar error after set 
npx.set_np()
URL: 
https://github.com/apache/incubator-mxnet/issues/16099#issuecomment-531911611
 
 
   @zachgk assign [@reminisce ]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16098: [Bug] Unrecognize parameter shape after npx.set_up()

[GitBox]

samskalicky commented on issue #16098: [Bug] Unrecognize parameter shape after 
npx.set_up()
URL: 
https://github.com/apache/incubator-mxnet/issues/16098#issuecomment-531911361
 
 
   @zachgk assign [@reminisce ]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch ir-patch created (now 5ed5689)

[junrushao]

This is an automated email from the ASF dual-hosted git repository.

junrushao pushed a change to branch ir-patch
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


  at 5ed5689  numpy operator ravel, derive from reshape (#16016)

No new revisions were added by this update.

[GitHub] [incubator-mxnet] samskalicky commented on issue #16098: [Bug] Unrecognize parameter shape after npx.set_up()

[GitBox]

samskalicky commented on issue #16098: [Bug] Unrecognize parameter shape after 
npx.set_up()
URL: 
https://github.com/apache/incubator-mxnet/issues/16098#issuecomment-531909162
 
 
   @zachgk add [bug]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16087: Error memory leak in multi thread use Threading library in python, when predict from same a model for multi thread

[GitBox]

samskalicky commented on issue #16087: Error memory leak in multi thread use 
Threading library in python, when predict from same a model for multi thread
URL: 
https://github.com/apache/incubator-mxnet/issues/16087#issuecomment-531907923
 
 
   @anirudh2290 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16087: Error memory leak in multi thread use Threading library in python, when predict from same a model for multi thread

[GitBox]

samskalicky commented on issue #16087: Error memory leak in multi thread use 
Threading library in python, when predict from same a model for multi thread
URL: 
https://github.com/apache/incubator-mxnet/issues/16087#issuecomment-531907672
 
 
   @mxnet-label-bot add [bug]
   
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] samskalicky commented on issue #16087: Error memory leak in multi thread use Threading library in python, when predict from same a model for multi thread

[GitBox]

samskalicky commented on issue #16087: Error memory leak in multi thread use 
Threading library in python, when predict from same a model for multi thread
URL: 
https://github.com/apache/incubator-mxnet/issues/16087#issuecomment-531907514
 
 
   @mxnet-label-bot add [bug,pending-requester-response]


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Vikas-kum closed issue #16168: group norm tests are failing

[GitBox]

Vikas-kum closed issue #16168: group norm tests are failing
URL: https://github.com/apache/incubator-mxnet/issues/16168
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] DickJC123 edited a comment on issue #15657: Eliminate common expressions

[GitBox]

DickJC123 edited a comment on issue #15657: Eliminate common expressions
URL: https://github.com/apache/incubator-mxnet/pull/15657#issuecomment-531360012
 
 
   Another area to spend some extra thought on involves the primary outputs and 
the gradients of primary inputs.  If you combined two output NDArrays via CSE, 
would there be any problem with someone wanting to index through the outputs?  
Ditto for input grads.  If a CSE-optimized model generated NDArrays for another 
model (and that model had mutable inputs), would functionality be preserved?  I 
think a downstream model gets a copy of the NDArray as the input, so it should 
work, but do we ever want to optimize away that copy?
   
   CSE may eliminate downstream in-place options for node outputs that have 
been combined.  Not sure if there's a case where using CSE is a net perf loss, 
but I doubt it.
   
   And finally, here's a fairly obscure case where the operator state may not 
be contained entirely in the node attributes (we should probably outlaw this 
usage, since it assumes no caching of the env var lookup):
   ```
   sym = ...
   os.environ['MXNET_MY_OP_BEHAVIOR']=X
   sym2 = MyOp(data=sym,...)# op behavior affected by 
captured env var setting
   os.environ['MNXET_MY_OP_BEHAVIOR']=Y
   sym3 = MyOp(data=sym,...)# op behavior affected by 
captured env var setting
   ```
   Perhaps shifting from the node attr dict to the param struct would catch 
this case.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] hzfan commented on issue #15855: [Numpy] cumprod

[GitBox]

hzfan commented on issue #15855: [Numpy] cumprod
URL: https://github.com/apache/incubator-mxnet/pull/15855#issuecomment-531888996
 
 
   Another minor modification: Rename 
`src/operator/contrib/tvmop/core/fromnumeric.cc` as 
`src/operator/numpy/np_cumprod.cc` (following the naming convention of 
`cumsum.cc`)


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] zhreshold commented on issue #16114: improve dataloader signals and messages

[GitBox]

zhreshold commented on issue #16114: improve dataloader signals and messages
URL: https://github.com/apache/incubator-mxnet/pull/16114#issuecomment-531885989
 
 
   @wkcn It's mandatory to have timeout in order to catch excetions in 
subprocess due to a python bug. 


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] access2rohit opened a new pull request #16178: [WIP]improving argmax perf

[GitBox]

access2rohit opened a new pull request #16178: [WIP]improving argmax perf
URL: https://github.com/apache/incubator-mxnet/pull/16178
 
 
   ## Description ##
   Removes mshadow implementation of argmax and add fresh implementation for it
   
   ## 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 to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] hzfan commented on issue #15855: [Numpy] cumprod

[GitBox]

hzfan commented on issue #15855: [Numpy] cumprod
URL: https://github.com/apache/incubator-mxnet/pull/15855#issuecomment-531877059
 
 
   Have Rebased. And a minor change: move the dispatch of req to func `set_req` 
in `src/operator/tvmop/utils.h` so it can be reused.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] leezu edited a comment on issue #16097: [numpy] array ufunc and array function protocols

[GitBox]

leezu edited a comment on issue #16097: [numpy] array ufunc and array function 
protocols
URL: https://github.com/apache/incubator-mxnet/pull/16097#issuecomment-530725409
 
 
   @szha this assumes that people can't upgrade to numpy 1.17 if they are using 
1.16, even though there are no breaking changes 
https://docs.scipy.org/doc/numpy/release.html
   
   If we properly declare the 1.17 dependency, users that install mxnet via pip 
will automatically get numpy 1.17 and everything will work. If we don't bump 
the dependency, users have to read through error messages to understand they 
need to upgrade..


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] leezu edited a comment on issue #16097: [numpy] array ufunc and array function protocols

[GitBox]

leezu edited a comment on issue #16097: [numpy] array ufunc and array function 
protocols
URL: https://github.com/apache/incubator-mxnet/pull/16097#issuecomment-530725409
 
 
   @szha this assumes that people can't upgrade to numpy 1.17 if they are using 
1.16, even though there is no breaking changes 
https://docs.scipy.org/doc/numpy/release.html
   
   If we properly declare the 1.17 dependency, users that install mxnet via pip 
will automatically get numpy 1.17 and everything will work. If we don't bump 
the dependency, users have to read through error messages to understand they 
need to upgrade..


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] shoubhik opened a new issue #16177: How to dump quantized weights from MKLDNN as Ndarray

[GitBox]

shoubhik opened a new issue #16177: How to dump quantized weights from MKLDNN 
as Ndarray
URL: https://github.com/apache/incubator-mxnet/issues/16177
 
 
   Hi,
   I have a use case where i need to debug the weights of Convolution layer 
when quantized by MKLDNN in MXNet 
[here](https://github.com/apache/incubator-mxnet/blob/master/src/operator/subgraph/mkldnn/mkldnn_conv.cc#L189-L190).
 The problem I am facing is that the model saved has the original fp32 weights 
and it is converted to int8 during runtime. When I do try to dump the 
cached_weight_ into a File as NDArray it is computed asynchronously and I get 
an error saying the weight is unavailable. The code I am using is something 
like this
   
   
   ```
   LOG(INFO) << "Writing cached integer weight file";
 char cwd[200];
 getcwd(cwd, sizeof(cwd));
 std::string current_dir(cwd);
 LOG(INFO) << "The current dir is : " << current_dir;
 LOG(INFO) << "Writing qunatized cached weights to file";
 int num_args = 1;
 std::vector data(num_args);
 std::vector names(num_args);
 data[0] = cached_weight_;
 names[0] = op_name;
 std::string cached_weight_file_path = current_dir + 
"/quantized-conv-weights/" + op_name;
 std::unique_ptr 
fo(dmlc::Stream::Create(cached_weight_file_path.c_str(), "w"));
 mxnet::NDArray::Save(fo.get(), data, names);
 LOG(INFO) << "Written weights to " + cached_weight_file_path;
   ```


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] Laurawly commented on issue #15901: [Numpy] operator hypot

[GitBox]

Laurawly commented on issue #15901: [Numpy] operator hypot
URL: https://github.com/apache/incubator-mxnet/pull/15901#issuecomment-531810964
 
 
   @tingying2020 Please fix the branch conflicts.


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated (18d145c -> 5ed5689)

[sxjscience]

This is an automated email from the ASF dual-hosted git repository.

sxjscience pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.


from 18d145c  use 1E-4 in groupnorm test(#16169)
 add 5ed5689  numpy operator ravel, derive from reshape (#16016)

No new revisions were added by this update.

Summary of changes:
 python/mxnet/ndarray/numpy/_op.py  | 56 +-
 python/mxnet/numpy/multiarray.py   | 49 -
 python/mxnet/symbol/numpy/_symbol.py   | 42 -
 src/operator/numpy/np_matrix_op.cc |  1 +
 tests/python/unittest/test_numpy_op.py | 33 
 5 files changed, 178 insertions(+), 3 deletions(-)

[GitHub] [incubator-mxnet] sxjscience merged pull request #16016: [numpy] operator ravel, derive from reshape

[GitBox]

sxjscience merged pull request #16016: [numpy] operator ravel, derive from 
reshape
URL: https://github.com/apache/incubator-mxnet/pull/16016
 
 
   


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] QueensGambit commented on issue #16173: Saving and loading cudNN autotune and graph optimization

[GitBox]

QueensGambit commented on issue #16173: Saving and loading cudNN autotune and 
graph optimization
URL: 
https://github.com/apache/incubator-mxnet/issues/16173#issuecomment-531757648
 
 
   @ZhennanQin This sounds reasonable.
   Maybe it would be good to create a cwiki page for the top level design:
   * https://cwiki.apache.org/confluence/display/MXNET/Proposals


This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services