[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #11906: support 1D and 3D arrays in MKLDNN.

[GitBox]

pengzhao-intel commented on issue #11906: support 1D and 3D arrays in MKLDNN.
URL: 
https://github.com/apache/incubator-mxnet/issues/11906#issuecomment-472726885
 
 
   The request features are implemented :)
   Closing now. Feel free to reopen if there is another issue.


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 closed issue #11906: support 1D and 3D arrays in MKLDNN.

[GitBox]

pengzhao-intel closed issue #11906: support 1D and 3D arrays in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/issues/11906
 
 
   


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 #13221: Flaky tests: test_gluon_model_zoo_gpu.test_training

[GitBox]

pengzhao-intel commented on issue #13221: Flaky tests: 
test_gluon_model_zoo_gpu.test_training
URL: 
https://github.com/apache/incubator-mxnet/issues/13221#issuecomment-472726554
 
 
   @pengxin99 could you check this test again?


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 #12411: MKLDNN fallback when not recording gradients and calling backwards

[GitBox]

pengzhao-intel commented on issue #12411: MKLDNN fallback when not recording 
gradients and calling backwards
URL: https://github.com/apache/incubator-mxnet/pull/12411#issuecomment-472725837
 
 
   ping, any update @azai91 


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 #13598: More fine-grained operator implementation dispatch & memory planning flow

[GitBox]

pengzhao-intel commented on issue #13598: More fine-grained operator 
implementation dispatch & memory planning flow 
URL: 
https://github.com/apache/incubator-mxnet/issues/13598#issuecomment-472725241
 
 
   @danithace Tao has answered the question in the forum :) thanks to raising 
the question.


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 #13881: Cmake : could not find mkl (missing MKLDNN_LIBRARY)

[GitBox]

pengzhao-intel commented on issue #13881: Cmake : could not find mkl (missing 
MKLDNN_LIBRARY)
URL: 
https://github.com/apache/incubator-mxnet/issues/13881#issuecomment-472724536
 
 
   @juliusshufan 


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] pengzhao-intel commented on issue #14093: MKLDNN Unsupported Dimension Bug

[GitBox]

pengzhao-intel commented on issue #14093: MKLDNN Unsupported Dimension Bug
URL: 
https://github.com/apache/incubator-mxnet/issues/14093#issuecomment-472723967
 
 
   Resolved by the latest master, so I am closing this issue.
   Feel free to re-open if there is any other open issue.
   


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 closed issue #14093: MKLDNN Unsupported Dimension Bug

[GitBox]

pengzhao-intel closed issue #14093: MKLDNN Unsupported Dimension Bug
URL: https://github.com/apache/incubator-mxnet/issues/14093
 
 
   


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 #14112: CMakeLists.txt

[GitBox]

pengzhao-intel commented on issue #14112: CMakeLists.txt
URL: 
https://github.com/apache/incubator-mxnet/issues/14112#issuecomment-472723691
 
 
   @juliusshufan 
   
   @mxnet-label-bot add [windows]
   


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wkcn commented on issue #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn commented on issue #14427: Copy the gradient to the new context when 
as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#issuecomment-472722154
 
 
   @szha There is a discussion about whether to copy the gradient automatically.
   https://github.com/apache/incubator-mxnet/issues/14391


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 #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265417769
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   There is a case:
   ```python
   a = mx.nd.array([1,2,3],ctx=mx.cpu())
   a.attach_grad()
   b = a.as_in_context(mx.gpu(0))
   print(b.grad) # return a gradient which is equal to a.grad, namely 
b._old_grad
   b.attach_grad()
   print(b.grad) # return a new gradient
   ```
   We could not overwrite the attribute grad, since the returned ndarray may be 
attached grad. It is necessary to add the attribute `_old_grad` to store the 
gradient.
   Is there a better solution?


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 #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265417769
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   There is a case:
   ```python
   a = mx.nd.array([1,2,3],ctx=mx.cpu())
   a.attach_grad()
   b = a.as_in_context(mx.gpu(0))
   print(b.grad) # return a gradient which is equal to a.grad, namely 
b._old_grad
   b.attach_grad()
   print(b.grad) # return a new gradient
   ```
   It is necessary to add the attribute `_old_grad` to store the gradient.
   Is there a better solution?


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] szha commented on issue #14420: v1.4.x: Backport of Windows CI Fixes

[GitBox]

szha commented on issue #14420: v1.4.x: Backport of Windows CI Fixes
URL: https://github.com/apache/incubator-mxnet/pull/14420#issuecomment-472703869
 
 
   Could you please link the related commit/PR 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] wkcn commented on a change in pull request #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265417769
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   There is a case:
   ```python
   a = mx.nd.array([1,2,3],ctx=mx.cpu())
   a.attach_grad()
   b = a.as_in_context(mx.gpu(0))
   print(b.grad) # return a gradient which is equal to a.grad, namely 
b._old_grad
   b.attach_grad()
   print(b.grad) # return a new gradient
   ```
   I need the attribute `_old_grad` to store the gradient.


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 #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265417769
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   There is a case:
   ```python
   a = mx.nd.array([1,2,3],ctx=mx.cpu())
   a.attach_grad()
   b = a.as_in_context()
   print(b.grad) # return a gradient which is equal to a.grad
   b.attach_grad()
   print(b.grad) # return a new gradient
   ```
   I need the attribute `_old_grad` to store the gradient.


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] yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray

[GitBox]

yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray
URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-472700764
 
 
   > Do we still around size 0 to page size?
   
   Yes. I am not sure the reason behind the logic so I do not change it in this 
PR. We also use aligned alloc in CPU which allocates 16/64 bytes when size is 0.


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] szha commented on a change in pull request #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

szha commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265416734
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   This seems rather obscure as users aren't aware that this field exists. If 
user explicitly chose to do the copy, I think it's ok to set it in `.grad` 
attribute


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] szha commented on a change in pull request #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

szha commented on a change in pull request #14427: Copy the gradient to the new 
context when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427#discussion_r265417009
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -2086,12 +2088,17 @@ def copyto(self, other):
 if other.handle is self.handle:
 warnings.warn('You are attempting to copy an array to itself', 
RuntimeWarning)
 return False
-return _internal._copyto(self, out=other)
+data = _internal._copyto(self, out=other)
 elif isinstance(other, Context):
 hret = NDArray(_new_alloc_handle(self.shape, other, True, 
self.dtype))
-return _internal._copyto(self, out=hret)
+data = _internal._copyto(self, out=hret)
 else:
 raise TypeError('copyto does not support type ' + str(type(other)))
+if copy_grad:
+grad = self.grad
+if grad is not None:
+setattr(data, '_old_grad', grad.copyto(other, copy_grad=False))
 
 Review comment:
   What's less clear to me is if copy_grad is true, the new grad array should 
be marked as the gradient of the new data array in the target context. Again, 
this seems to implicit.


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] yuxihu commented on issue #14365: Fix memory leak for size-zero ndarray

[GitBox]

yuxihu commented on issue #14365: Fix memory leak for size-zero ndarray
URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-472700764
 
 
   > Do we still around size 0 to page size?
   
   Yes. I am not sure the reason behind the logic so I do not change it in this 
PR. We also use aligned alloc in CPU which also allocates 16/64 bytes when size 
is 0.


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] yuxihu commented on a change in pull request #14365: Fix memory leak for size-zero ndarray

[GitBox]

yuxihu commented on a change in pull request #14365: Fix memory leak for 
size-zero ndarray
URL: https://github.com/apache/incubator-mxnet/pull/14365#discussion_r265416688
 
 

 ##
 File path: include/mxnet/ndarray.h
 ##
 @@ -1055,9 +1054,12 @@ class NDArray {
   }
   size_t aux_bytes = shape.Size() * mshadow::mshadow_sizeof(aux_types[i]);
   if (aux_handles[i].size < aux_bytes) {
-// free storage if necessary and alloc again
-if (aux_handles[i].size > 0) Storage::Get()->Free(aux_handles[i]);
-// init aux storage
+// set the context for aux handle to make sure we free it to the right 
device.
+if (aux_handles[i].ctx.dev_type != ctx.dev_type) {
 
 Review comment:
   Yes, test_operator_gpu:test_sparse_nd_elemwise_add failed if context is not 
set here. A new aux handle is created 
[here](https://github.com/apache/incubator-mxnet/blob/master/include/mxnet/ndarray.h#L1054)
 with the default context (CPU) and size(0). Currently we have the size > 0 
check so we do not call Free. After we remove the size > 0 check, it will call 
Free with CPU context, which caused [this 
failure](https://github.com/apache/incubator-mxnet/blob/master/src/storage/storage.cc#L134).
 


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 opened a new pull request #14427: Copy the gradient to the new context when as_in_context() is called

[GitBox]

wkcn opened a new pull request #14427: Copy the gradient to the new context 
when as_in_context() is called
URL: https://github.com/apache/incubator-mxnet/pull/14427
 
 
   ## Description ##
   Hi! there.
   Fixed: https://github.com/apache/incubator-mxnet/issues/14391
   Now, ndarray.as_in_context() copies the gradient to the new context.
   
   ## 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 ###
   - [x] Add a new arguments `copy_grad` in `mx.nd.as_in_context` and 
`mx.nd.copyto`
   - [x] Add unit-test in `test_ndarray.py`
   
   ## 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] juliusshufan commented on issue #14376: I failed to install the mxnet-mkl version on Windows 10

[GitBox]

juliusshufan commented on issue #14376: I  failed to install the mxnet-mkl 
version on Windows 10
URL: 
https://github.com/apache/incubator-mxnet/issues/14376#issuecomment-472692009
 
 
   @wufengqian 
   Can you please try the below steps to make it works?
   1. Download and install the latest full MKL version following instructions 
on the intel website. https://software.intel.com/en-us/mkl
   2. Add the mkl dll path to Windows PATH, the local is similiar to: (X 
standing for the partition you installing the full MKL)
   : X:\Program Files 
(x86)\IntelSWTools\compilers_and_libraries_2019.3.203\windows\redist\intel64_win\mkl
   
   Thanks.
   


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] ThomasDelteil commented on issue #14391: as_in_context() does not copy the gradient to the new context

[GitBox]

ThomasDelteil commented on issue #14391: as_in_context() does not copy the 
gradient to the new context
URL: 
https://github.com/apache/incubator-mxnet/issues/14391#issuecomment-472688222
 
 
   In what case would you not want to copy gradients that has been attached?
   Most as in context happened on data without gradient, I only discovered the
   bug on some niche use case that requires data with gradient attached.
   
   If the data does not have gradient attached the runtime would be the same
   since no gradient would need moving
   
   On Wed, Mar 13, 2019, 19:00 JackieWu  wrote:
   
   > I'm worry copy_grad=True will drop the performance. Assume that it takes
   > 1s to copy the data of a large tensor, it will takes extra 1s to copy the
   > gradient.
   >
   > —
   > 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] haojin2 commented on issue #13958: Windows builds running out of heap space in CI

[GitBox]

haojin2 commented on issue #13958: Windows builds running out of heap space in 
CI
URL: 
https://github.com/apache/incubator-mxnet/issues/13958#issuecomment-472685473
 
 
   I'm running into this issue for #14359 too. Any updates?


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] iblis17 commented on issue #14422: [WIP] [DO NOT MERGE] Support scalar and zero-size tensors

[GitBox]

iblis17 commented on issue #14422: [WIP] [DO NOT MERGE] Support scalar and 
zero-size tensors
URL: https://github.com/apache/incubator-mxnet/pull/14422#issuecomment-472684190
 
 
   Just FYI: https://github.blog/2019-02-14-introducing-draft-pull-requests/


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 #14426: mx.random.seed with ctx failures on a gpu build when run with cpu context

[GitBox]

mxnet-label-bot commented on issue #14426: mx.random.seed with ctx failures on 
a gpu build when run with cpu context
URL: 
https://github.com/apache/incubator-mxnet/issues/14426#issuecomment-472678265
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended labels: 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] anirudh2290 opened a new issue #14426: mx.random.seed with ctx failures on a gpu build when run with cpu context

[GitBox]

anirudh2290 opened a new issue #14426: mx.random.seed with ctx failures on a 
gpu build when run with cpu context
URL: https://github.com/apache/incubator-mxnet/issues/14426
 
 
   Minimum reproducible example. 
   ```
   import mxnet as mx
   
   def set_seed_variously_for_context(ctx, init_seed, num_init_seeds, 
final_seed):
   end_seed = init_seed + num_init_seeds
   for seed in range(init_seed, end_seed):
   mx.random.seed(seed, ctx=ctx)
   z = mx.random.seed(seed, ctx=ctx)
   return end_seed
   
   dtype = 'float32'
   
   samples_imp = []
   samples_sym = []
   ctx = mx.cpu()
   shape = (200, 200)
   params = {'low': -1.5, 'high': 3.0}
   params.update(shape=shape, dtype=dtype)
   seed = set_seed_variously_for_context(ctx, 1, 1, 1234)
   mx.nd.waitall()
   ```
   
   To reproduce, this will require the exception handling support for waitall 
in this PR: https://github.com/apache/incubator-mxnet/pull/14397 . This issue 
was found because of CI failures when running test_random.py on windows. It was 
hidden earlier because waitall didnt have exception rethrow support. This issue 
may have been around since the PR was added: #10367 
   
   Currently working on fixing 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] szha commented on a change in pull request #14423: Add repr for SymbolBlock

[GitBox]

szha commented on a change in pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#discussion_r265398834
 
 

 ##
 File path: python/mxnet/gluon/block.py
 ##
 @@ -1024,6 +1024,11 @@ def imports(symbol_file, input_names, param_file=None, 
ctx=None):
 ret.collect_params().load(param_file, ctx=ctx)
 return ret
 
+def __repr__(self):
+s = '{name}(\n{modstr}\n)'
+modstr = 
'\n'.join(['{block}'.format(block=self.__dict__['_cached_graph'][-1])])
 
 Review comment:
   Printing the inner structure might be too complicated and also not readable, 
thus I suggested printing the count.
   
   Why would `cached_graph[1]` ever hit index out of bounds?


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] szha commented on a change in pull request #14425: Fix crashes on visualization

[GitBox]

szha commented on a change in pull request #14425: Fix crashes on visualization
URL: https://github.com/apache/incubator-mxnet/pull/14425#discussion_r265398504
 
 

 ##
 File path: python/mxnet/visualization.py
 ##
 @@ -339,11 +339,17 @@ def looks_like_weight(name):
 elif op == "BatchNorm":
 attr["fillcolor"] = cm[3]
 elif op in ('Activation', 'LeakyReLU'):
-label = r"%s\n%s" % (op, node["attrs"]["act_type"])
+if op == "LeakyReLU":
 
 Review comment:
   if special handling is needed for each case, then the outer `elif op in ...` 
is not necessary.


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] szha commented on a change in pull request #14425: Fix crashes on visualization

[GitBox]

szha commented on a change in pull request #14425: Fix crashes on visualization
URL: https://github.com/apache/incubator-mxnet/pull/14425#discussion_r265398555
 
 

 ##
 File path: python/mxnet/visualization.py
 ##
 @@ -339,11 +339,17 @@ def looks_like_weight(name):
 elif op == "BatchNorm":
 attr["fillcolor"] = cm[3]
 elif op in ('Activation', 'LeakyReLU'):
-label = r"%s\n%s" % (op, node["attrs"]["act_type"])
+if op == "LeakyReLU":
+attrs = node.get("attrs")
+act_type = attrs.get("act_type", "Leaky") if attrs else "Leaky"
+else:
+act_type = node["attrs"]["act_type"]
+label = r"%s\n%s" % (op, str(act_type))
 
 Review comment:
   this usage is being deprecated. use string.format instead.


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wkcn commented on issue #14391: as_in_context() does not copy the gradient to the new context

[GitBox]

wkcn commented on issue #14391: as_in_context() does not copy the gradient to 
the new context
URL: 
https://github.com/apache/incubator-mxnet/issues/14391#issuecomment-472674519
 
 
   I'm worry `copy_grad=True` will drop the performance. Assume that it takes 
1s to copy the data of a large tensor, it will takes extra 1s to copy the 
gradient.


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] vandanavk commented on issue #14425: Fix crashes on visualization

[GitBox]

vandanavk commented on issue #14425: Fix crashes on visualization
URL: https://github.com/apache/incubator-mxnet/pull/14425#issuecomment-472674003
 
 
   Test for plot network has been disabled in the code
   
   ```
   @unittest.skipIf(not graphviz_exists(), "Skip test_plot_network as Graphviz 
could not be imported")
   def test_plot_network():
   ```


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] vandanavk commented on issue #13485: Found bug in caffe convert tool and solved

[GitBox]

vandanavk commented on issue #13485: Found bug in caffe convert tool and solved
URL: 
https://github.com/apache/incubator-mxnet/issues/13485#issuecomment-472671056
 
 
   @mxnet-label-bot add [Converter]


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] vandanavk commented on issue #14093: MKLDNN Unsupported Dimension Bug

[GitBox]

vandanavk commented on issue #14093: MKLDNN Unsupported Dimension Bug
URL: 
https://github.com/apache/incubator-mxnet/issues/14093#issuecomment-472670802
 
 
   @mxnet-label-bot add [MKLDNN]


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] vandanavk commented on issue #14234: Backwards compatibility out of bounds for 1.2.1

[GitBox]

vandanavk commented on issue #14234: Backwards compatibility out of bounds for 
1.2.1
URL: 
https://github.com/apache/incubator-mxnet/issues/14234#issuecomment-472670660
 
 
   @mxnet-label-bot add [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] vandanavk commented on issue #14425: Check for kernel in Pooling

[GitBox]

vandanavk commented on issue #14425: Check for kernel in Pooling
URL: https://github.com/apache/incubator-mxnet/pull/14425#issuecomment-472670455
 
 
   @mxnet-label-bot add [Visualization, pr-awaiting-review]


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] vandanavk opened a new pull request #14425: Check for kernel in Pooling

[GitBox]

vandanavk opened a new pull request #14425: Check for kernel in Pooling
URL: https://github.com/apache/incubator-mxnet/pull/14425
 
 
   ## Description ##
   Fixes https://github.com/apache/incubator-mxnet/issues/14245
   
   ## 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 ###
   - Add a check for the presence of kernel in attrs
   
   ## Comments ##
   Test code:
   ```
   import mxnet as mx
   data=mx.sym.Variable('data')
   symbol = mx.sym.Pooling(data = data, global_pool = True, pool_type = 'avg')
   graph = mx.viz.plot_network(symbol)
   graph
   ```
   


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] vandanavk commented on a change in pull request #14423: Add repr for SymbolBlock

[GitBox]

vandanavk commented on a change in pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#discussion_r265392063
 
 

 ##
 File path: python/mxnet/gluon/block.py
 ##
 @@ -1024,6 +1024,11 @@ def imports(symbol_file, input_names, param_file=None, 
ctx=None):
 ret.collect_params().load(param_file, ctx=ctx)
 return ret
 
+def __repr__(self):
+s = '{name}(\n{modstr}\n)'
+modstr = 
'\n'.join(['{block}'.format(block=self.__dict__['_cached_graph'][-1])])
 
 Review comment:
   I think showing all layers like how HybridSequential prints, would be most 
useful, but not sure how to do that for Symbol API and SymbolBlock models.


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] vandanavk commented on a change in pull request #14423: Add repr for SymbolBlock

[GitBox]

vandanavk commented on a change in pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#discussion_r265391945
 
 

 ##
 File path: python/mxnet/gluon/block.py
 ##
 @@ -1024,6 +1024,11 @@ def imports(symbol_file, input_names, param_file=None, 
ctx=None):
 ret.collect_params().load(param_file, ctx=ctx)
 return ret
 
+def __repr__(self):
+s = '{name}(\n{modstr}\n)'
+modstr = 
'\n'.join(['{block}'.format(block=self.__dict__['_cached_graph'][-1])])
 
 Review comment:
   corrected __dict__ to self._cached_graph. 
   -1 index as the last element of cached_graph is the output. I could have 
used cached_graph[1], but did not want to hit index out of bounds (in case it 
ever happens).
   
   


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.

[zhasheng]

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

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


The following commit(s) were added to refs/heads/asf-site by this push:
 new 3087405  Bump the publish timestamp.
3087405 is described below

commit 308740566067d5ce131dfa2341911ea9c91d3bf7
Author: mxnet-ci 
AuthorDate: Thu Mar 14 01:17:49 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..7e2d98a
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Thu Mar 14 01:17:49 UTC 2019

[GitHub] [incubator-mxnet] vandanavk commented on issue #4887: multi-node training hangs occasionally

[GitBox]

vandanavk commented on issue #4887: multi-node training hangs occasionally
URL: 
https://github.com/apache/incubator-mxnet/issues/4887#issuecomment-472664509
 
 
   @mxnet-label-bot add [Distributed]


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] vandanavk commented on issue #7664: index out of bound error when update eval metric

[GitBox]

vandanavk commented on issue #7664: index out of bound error when update eval 
metric
URL: 
https://github.com/apache/incubator-mxnet/issues/7664#issuecomment-472663377
 
 
   @mxnet-label-bot add [Metric]


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] vandanavk commented on issue #8646: Bug in FullyConnected???

[GitBox]

vandanavk commented on issue #8646: Bug in FullyConnected???
URL: 
https://github.com/apache/incubator-mxnet/issues/8646#issuecomment-472663154
 
 
   @mxnet-label-bot add [Operator]


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] vandanavk commented on issue #9159: Symbol.bind() do not warn or error on dtype mismatch between Symbol and binding Data

[GitBox]

vandanavk commented on issue #9159: Symbol.bind() do not warn or error on dtype 
mismatch between Symbol and binding Data
URL: 
https://github.com/apache/incubator-mxnet/issues/9159#issuecomment-472662517
 
 
   @mxnet-label-bot update [Feature Request, Symbol]


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] vandanavk commented on issue #9624: problem when set fix_gamma=True in batchnorm

[GitBox]

vandanavk commented on issue #9624: problem when set fix_gamma=True in batchnorm
URL: 
https://github.com/apache/incubator-mxnet/issues/9624#issuecomment-472662105
 
 
   @mxnet-label-bot add [Operator]


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] vandanavk commented on issue #9967: Error in output.shape[[output.names]]

[GitBox]

vandanavk commented on issue #9967: Error in output.shape[[output.names]]
URL: 
https://github.com/apache/incubator-mxnet/issues/9967#issuecomment-472661644
 
 
   @mxnet-label-bot add [R]


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] vandanavk commented on issue #10173: Unpredictable nan in Array

[GitBox]

vandanavk commented on issue #10173: Unpredictable nan in Array
URL: 
https://github.com/apache/incubator-mxnet/issues/10173#issuecomment-472661558
 
 
   @mxnet-label-bot add [NDArray]


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 issue #10389: Report clear errors when opencv::imdecode fails.

[GitBox]

anirudh2290 commented on issue #10389: Report clear errors when 
opencv::imdecode fails.
URL: 
https://github.com/apache/incubator-mxnet/issues/10389#issuecomment-472661483
 
 
   Working on a PR for 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] vandanavk commented on issue #10389: Report clear errors when opencv::imdecode fails.

[GitBox]

vandanavk commented on issue #10389: Report clear errors when opencv::imdecode 
fails.
URL: 
https://github.com/apache/incubator-mxnet/issues/10389#issuecomment-472660943
 
 
   @mxnet-label-bot add [Exception Handling]


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] vandanavk commented on issue #10357: Write Inplace isn't enabled in activation forward during inference.

[GitBox]

vandanavk commented on issue #10357: Write Inplace isn't enabled in activation 
forward during inference.
URL: 
https://github.com/apache/incubator-mxnet/issues/10357#issuecomment-472661040
 
 
   @mxnet-label-bot add [Operator]


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] vandanavk commented on issue #12444: Error in IdentityAttachKLSparseReg

[GitBox]

vandanavk commented on issue #12444: Error in IdentityAttachKLSparseReg
URL: 
https://github.com/apache/incubator-mxnet/issues/12444#issuecomment-472660812
 
 
   @mxnet-label-bot add [Symbol]


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 #14313: compatibility with opencv4

[GitBox]

wkcn commented on a change in pull request #14313: compatibility with opencv4
URL: https://github.com/apache/incubator-mxnet/pull/14313#discussion_r265384977
 
 

 ##
 File path: Makefile
 ##
 @@ -146,11 +146,33 @@ endif
 
 # setup opencv
 ifeq ($(USE_OPENCV), 1)
-   CFLAGS += -DMXNET_USE_OPENCV=1 $(shell pkg-config --cflags opencv)
-   LDFLAGS += $(filter-out -lopencv_ts, $(shell pkg-config --libs opencv))
+   CFLAGS += -DMXNET_USE_OPENCV=1
+   ifneq ($(USE_OPENCV_INC_PATH), NONE)
 
 Review comment:
   @reminisce @szha Hi! I have fixed it in the PR 
https://github.com/apache/incubator-mxnet/pull/14424


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 opened a new pull request #14424: fix Makefile bug when USE_OPENCV_INC_PATH is not defined

[GitBox]

wkcn opened a new pull request #14424: fix Makefile bug when 
USE_OPENCV_INC_PATH is not defined
URL: https://github.com/apache/incubator-mxnet/pull/14424
 
 
   ## Description ##
   fix Makefile bug when USE_OPENCV_INC_PATH is not defined
   
   ## 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] ptrendx commented on a change in pull request #14377: Correct update count with Gluon trainer and update_on_kvstore=False

[GitBox]

ptrendx commented on a change in pull request #14377: Correct update count with 
Gluon trainer and update_on_kvstore=False
URL: https://github.com/apache/incubator-mxnet/pull/14377#discussion_r265384330
 
 

 ##
 File path: python/mxnet/optimizer/optimizer.py
 ##
 @@ -380,6 +381,11 @@ def set_wd_mult(self, args_wd_mult):
 self.wd_mult[name] = float(attr[name]['__wd_mult__'])
 self.wd_mult.update(args_wd_mult)
 
+def set_current_context(self, ctx):
+if ctx not in self._all_index_update_counts:
 
 Review comment:
   Will do


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] eric-haibin-lin commented on a change in pull request #14365: Fix memory leak for size-zero ndarray

[GitBox]

eric-haibin-lin commented on a change in pull request #14365: Fix memory leak 
for size-zero ndarray
URL: https://github.com/apache/incubator-mxnet/pull/14365#discussion_r265383417
 
 

 ##
 File path: include/mxnet/ndarray.h
 ##
 @@ -1055,9 +1054,12 @@ class NDArray {
   }
   size_t aux_bytes = shape.Size() * mshadow::mshadow_sizeof(aux_types[i]);
   if (aux_handles[i].size < aux_bytes) {
-// free storage if necessary and alloc again
-if (aux_handles[i].size > 0) Storage::Get()->Free(aux_handles[i]);
-// init aux storage
+// set the context for aux handle to make sure we free it to the right 
device.
+if (aux_handles[i].ctx.dev_type != ctx.dev_type) {
 
 Review comment:
   This usually do not happen. Have you observed such mismatch? Would it be 
more appropriate to add CHECK_EQ instead of modifying the ctx? 


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] andrewfayres commented on issue #14421: Updating mxnet from 1.0.0, networks give different outputs

[GitBox]

andrewfayres commented on issue #14421: Updating mxnet from 1.0.0, networks 
give different outputs
URL: 
https://github.com/apache/incubator-mxnet/issues/14421#issuecomment-472658140
 
 
   Do you have a sample model which can be shared to reproduce this? There is a 
backwards compatibility checker for models as part of the nightly pipeline. It 
trains models on earlier releases and checks for consistency on the latest 
builds. It's possible there is an edge case which is being missed.
   
   @mxnet-label-bot add [Pending Requestor Info, 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] szha commented on a change in pull request #14423: Add repr for SymbolBlock

[GitBox]

szha commented on a change in pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#discussion_r265381409
 
 

 ##
 File path: python/mxnet/gluon/block.py
 ##
 @@ -1024,6 +1024,11 @@ def imports(symbol_file, input_names, param_file=None, 
ctx=None):
 ret.collect_params().load(param_file, ctx=ctx)
 return ret
 
+def __repr__(self):
+s = '{name}(\n{modstr}\n)'
+modstr = 
'\n'.join(['{block}'.format(block=self.__dict__['_cached_graph'][-1])])
 
 Review comment:
   would it be more useful to show also the input count and output count?


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] szha commented on a change in pull request #14423: Add repr for SymbolBlock

[GitBox]

szha commented on a change in pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#discussion_r265381261
 
 

 ##
 File path: python/mxnet/gluon/block.py
 ##
 @@ -1024,6 +1024,11 @@ def imports(symbol_file, input_names, param_file=None, 
ctx=None):
 ret.collect_params().load(param_file, ctx=ctx)
 return ret
 
+def __repr__(self):
+s = '{name}(\n{modstr}\n)'
+modstr = 
'\n'.join(['{block}'.format(block=self.__dict__['_cached_graph'][-1])])
 
 Review comment:
   why use `__dict__`? or -1 index?


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] andrewfayres commented on issue #14415: [Test Failure] Clojure Integration

[GitBox]

andrewfayres commented on issue #14415: [Test Failure] Clojure Integration
URL: 
https://github.com/apache/incubator-mxnet/issues/14415#issuecomment-472656045
 
 
   Yep, this was due to the Scala NDArray arange test failing. Going to go 
ahead and get this one closed.
   
   @mxnet-label-bot [Clojure, Test, Flaky, Scala]


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] andrewfayres commented on issue #14415: [Test Failure] Clojure Integration

[GitBox]

andrewfayres commented on issue #14415: [Test Failure] Clojure Integration
URL: 
https://github.com/apache/incubator-mxnet/issues/14415#issuecomment-472656083
 
 
   @mxnet-label-bot add [Clojure, Test, Flaky, Scala]


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] andrewfayres commented on issue #14411: cpp-package: test failure TestStochasticTiming_2D

[GitBox]

andrewfayres commented on issue #14411: cpp-package: test failure 
TestStochasticTiming_2D
URL: 
https://github.com/apache/incubator-mxnet/issues/14411#issuecomment-47266
 
 
   @mxnet-label-bot add [C++, 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] vandanavk commented on issue #14423: Add repr for SymbolBlock

[GitBox]

vandanavk commented on issue #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#issuecomment-472655509
 
 
   @szha @eric-haibin-lin added a test in test_import() of test_gluon.py


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] andrewfayres commented on issue #14408: Multi-threaded execution leads to high CPU load

[GitBox]

andrewfayres commented on issue #14408: Multi-threaded execution leads to high 
CPU load
URL: 
https://github.com/apache/incubator-mxnet/issues/14408#issuecomment-472655377
 
 
   @mxnet-label-bot add [Performance, c++, question]
   
   Adding labels for better visibility. You might try asking this question on 
https://discuss.mxnet.io/ instead. Questions there tend to get better 
visibility as github is primarily used to track issues.


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wkcn commented on issue #14409: [Numpy] Change semantics of ndim for operators in `src/operator/contrib`

[GitBox]

wkcn commented on issue #14409: [Numpy] Change semantics of ndim for operators 
in `src/operator/contrib`
URL: https://github.com/apache/incubator-mxnet/pull/14409#issuecomment-472653316
 
 
   Could we add a cast to be compatible with nnvm::TShape,Tuple?


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 issue #14406: resnet cpp-package test is broken

[GitBox]

anirudh2290 commented on issue #14406: resnet cpp-package test is broken
URL: 
https://github.com/apache/incubator-mxnet/issues/14406#issuecomment-472651873
 
 
   the last i observed it was around 11GB. For now I am going to use smaller 
batch_size for tests and later @leleamol will revisit and improve the cpp tests.


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] wkcn commented on issue #14406: resnet cpp-package test is broken

[GitBox]

wkcn commented on issue #14406: resnet cpp-package test is broken
URL: 
https://github.com/apache/incubator-mxnet/issues/14406#issuecomment-472651055
 
 
   Could someone please look the GPU memory used by the model?


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] szha commented on a change in pull request #14377: Correct update count with Gluon trainer and update_on_kvstore=False

[GitBox]

szha commented on a change in pull request #14377: Correct update count with 
Gluon trainer and update_on_kvstore=False
URL: https://github.com/apache/incubator-mxnet/pull/14377#discussion_r265376237
 
 

 ##
 File path: python/mxnet/optimizer/optimizer.py
 ##
 @@ -380,6 +381,11 @@ def set_wd_mult(self, args_wd_mult):
 self.wd_mult[name] = float(attr[name]['__wd_mult__'])
 self.wd_mult.update(args_wd_mult)
 
+def set_current_context(self, ctx):
+if ctx not in self._all_index_update_counts:
 
 Review comment:
   rename to _set_current_context, and document the type of ctx?


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 issue #14406: resnet cpp-package test is broken

[GitBox]

anirudh2290 commented on issue #14406: resnet cpp-package test is broken
URL: 
https://github.com/apache/incubator-mxnet/issues/14406#issuecomment-472649147
 
 
   hi @leleamol . to reproduce you will have to use g3.8xlarge. I was able to 
reproduce on a g3.8xlarge.


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] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

[GitBox]

apeforest commented on a change in pull request #14286: Add examples of running 
MXNet with Horovod
URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r265369218
 
 

 ##
 File path: example/distributed_training-horovod/README.md
 ##
 @@ -0,0 +1,207 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+# MXNet + Horovod
+[Horovod](https://github.com/horovod/horovod) is a distributed training 
framework that demonstrates 
+excellent scaling efficiency for dense models running on a large number of 
nodes. It currently 
+supports mainstream deep learning frameworks such as MXNet, TensorFlow, Keras, 
and PyTorch. 
+It is created at Uber and currently hosted by the [Linux Foundation Deep 
Learning](https://lfdl.io)(LF DL). 
+
+MXNet is recently supported in Horovod 0.16.0 
[release](https://eng.uber.com/horovod-pyspark-apache-mxnet-support/).
+
+## What's New?
+Compared with the standard distributed training script in MXNet which uses 
parameter server to 
+distribute and aggregate parameters, Horovod uses ring allreduce algorithm to 
communicate parameters 
 
 Review comment:
   thanks for the review. updated.


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] vandanavk commented on issue #14423: Add repr for SymbolBlock

[GitBox]

vandanavk commented on issue #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#issuecomment-472643669
 
 
   Thanks for inputs @szha @eric-haibin-lin. Added an example in the PR 
description. Will add a test and submit shortly


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] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

[GitBox]

apeforest commented on a change in pull request #14286: Add examples of running 
MXNet with Horovod
URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r265368832
 
 

 ##
 File path: example/distributed_training-horovod/README.md
 ##
 @@ -0,0 +1,201 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+# Distributed Training using MXNet with Horovod 
+[Horovod](https://github.com/horovod/horovod) is a distributed training 
framework that demonstrates 
+excellent scaling efficiency for dense models running on a large number of 
nodes. It currently 
+supports mainstream deep learning frameworks such as MXNet, TensorFlow, Keras, 
and PyTorch. 
+It is created at Uber and currently hosted by the [Linux Foundation Deep 
Learning](https://lfdl.io)(LF DL). 
+
+MXNet is supported in Horovod 0.16.0 
[release](https://eng.uber.com/horovod-pyspark-apache-mxnet-support/).
+
+## What's New?
+Compared with the standard distributed training script in MXNet which uses 
parameter server to 
+distribute and aggregate parameters, Horovod uses ring allreduce algorithm to 
communicate parameters 
+between workers. There is no dedicated server and the communication data size 
+between workers does not depend on the number of workers. Therefore, it scales 
well in the case where 
+there are a large number of workers and network bandwidth is the bottleneck.
+
+# Install
+## Install MXNet
+```bash
+$ pip install mxnet
 
 Review comment:
   updated.


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] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

[GitBox]

apeforest commented on a change in pull request #14286: Add examples of running 
MXNet with Horovod
URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r265368779
 
 

 ##
 File path: example/distributed_training-horovod/README.md
 ##
 @@ -0,0 +1,201 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+# Distributed Training using MXNet with Horovod 
+[Horovod](https://github.com/horovod/horovod) is a distributed training 
framework that demonstrates 
+excellent scaling efficiency for dense models running on a large number of 
nodes. It currently 
+supports mainstream deep learning frameworks such as MXNet, TensorFlow, Keras, 
and PyTorch. 
+It is created at Uber and currently hosted by the [Linux Foundation Deep 
Learning](https://lfdl.io)(LF DL). 
+
+MXNet is supported in Horovod 0.16.0 
[release](https://eng.uber.com/horovod-pyspark-apache-mxnet-support/).
+
+## What's New?
+Compared with the standard distributed training script in MXNet which uses 
parameter server to 
+distribute and aggregate parameters, Horovod uses ring allreduce algorithm to 
communicate parameters 
+between workers. There is no dedicated server and the communication data size 
+between workers does not depend on the number of workers. Therefore, it scales 
well in the case where 
+there are a large number of workers and network bandwidth is the bottleneck.
+
+# Install
+## Install MXNet
+```bash
+$ pip install mxnet
+```
+**Note**: There is a [known 
issue](https://github.com/horovod/horovod/issues/884) when running Horovod with 
MXNet on a Linux system with GCC version 5.X and above. We recommend users to 
build MXNet from source following this 
[guide](https://mxnet.incubator.apache.org/install/build_from_source.html) as a 
workaround for now.
+
+## Install Horovod
+```bash
+$ pip install horovod
+```
+
+This basic installation is good for laptops and for getting to know Horovod.
+If you're installing Horovod on a server with GPUs, read the [Horovod on 
GPU](https://github.com/horovod/horovod/blob/master/docs/gpus.md) page.
+If you want to use Docker, read the [Horovod in 
Docker](https://github.com/horovod/horovod/blob/master/docs/docker.md) page.
+
+## Install Open MPI
 
 Review comment:
   updated


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] eric-haibin-lin commented on issue #14423: Add repr for SymbolBlock

[GitBox]

eric-haibin-lin commented on issue #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#issuecomment-472641287
 
 
   +1 on szha's comment


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] vandanavk commented on issue #14423: Add repr for SymbolBlock

[GitBox]

vandanavk commented on issue #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423#issuecomment-472636678
 
 
   @mxnet-label-bot add [Gluon, pr-awaiting-review]


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] vandanavk opened a new pull request #14423: Add repr for SymbolBlock

[GitBox]

vandanavk opened a new pull request #14423: Add repr for SymbolBlock
URL: https://github.com/apache/incubator-mxnet/pull/14423
 
 
   ## Description ##
   SymbolBlock uses Block.__repr__ when print(SymbolBlock) is executed. In 
Block.__repr__, data is printed only if it is of type Block. Adding a __repr__ 
to SymbolBlock class that prints the output symbol. This behavior is similar to 
printing pure Symbol API.
   
   Fixes https://github.com/apache/incubator-mxnet/issues/13616
   
   ## 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 ###
   - Add a __repr__ for SymbolBlock
   
   ## Comments ##
   @eric-haibin-lin @Ishitori 
   


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] leleamol commented on issue #14406: resnet cpp-package test is broken

[GitBox]

leleamol commented on issue #14406: resnet cpp-package test is broken
URL: 
https://github.com/apache/incubator-mxnet/issues/14406#issuecomment-472635606
 
 
   I tried these examples with the recent code change in "WaitAll()" on p2.16x 
instances and c5.18x instances. I did not see the crash.
   
   However,  we still need to add missing exception handling in the example so 
that we can prevent the crashes due to unhandled exceptions.


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] ifeherva edited a comment on issue #14321: moveaxis operator now accepts negative indices and sequence of ints as well.

[GitBox]

ifeherva edited a comment on issue #14321: moveaxis operator now accepts 
negative indices and sequence of ints as well.
URL: https://github.com/apache/incubator-mxnet/pull/14321#issuecomment-472614064
 
 
   Strange that the same GPU tests fail though I am puzzled how these changes 
contribute to CUDA errors.


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] ifeherva commented on issue #14321: moveaxis operator now accepts negative indices and sequence of ints as well.

[GitBox]

ifeherva commented on issue #14321: moveaxis operator now accepts negative 
indices and sequence of ints as well.
URL: https://github.com/apache/incubator-mxnet/pull/14321#issuecomment-472614064
 
 
   Strange that the same GPU test fail though I am puzzled how these changes 
contribute to CUDA errors.


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] reminisce opened a new pull request #14422: [WIP] [DO NOT MERGE] Support scalar and zero-size tensors

[GitBox]

reminisce opened a new pull request #14422: [WIP] [DO NOT MERGE] Support scalar 
and zero-size tensors
URL: https://github.com/apache/incubator-mxnet/pull/14422
 
 
   ## Description ##
   Only used for the purpose of code review. Will close the PR and submit a new 
one against the numpy branch.
   
   @eric-haibin-lin @szha @junrushao1994 @zheng-da @yzhliu 


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.

[zhasheng]

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

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


The following commit(s) were added to refs/heads/asf-site by this push:
 new 869446c  Bump the publish timestamp.
869446c is described below

commit 869446ccb1f465bac6a06028c29b45d1ae10a004
Author: mxnet-ci 
AuthorDate: Wed Mar 13 21:06:28 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..b7d68e4
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Wed Mar 13 21:06:28 UTC 2019

[GitHub] [incubator-mxnet] lanking520 commented on issue #14349: installed mxnet using pip install mxnet==1.3.1. Using numpy 1.14.3. But the import mxnet fails

[GitBox]

lanking520 commented on issue #14349: installed mxnet using pip install 
mxnet==1.3.1. Using numpy 1.14.3. But the import mxnet fails
URL: 
https://github.com/apache/incubator-mxnet/issues/14349#issuecomment-472602875
 
 
   @aChatterjee13  It can happened in a way that pip does not linked to the 
python you used. Recommended to create a conda environment or pip environment 
to install and run mxnet. Please let me know if you are facing any problems by 
doing so.


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: [MXNET-1226] add Docs update for MXNet Java (#14395)

[lanking]

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

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


The following commit(s) were added to refs/heads/master by this push:
 new 4432af1  [MXNET-1226] add Docs update for MXNet Java (#14395)
4432af1 is described below

commit 4432af1f47a439517eff9a21bef23ef7ae5e4aa4
Author: Lanking 
AuthorDate: Wed Mar 13 13:07:19 2019 -0700

[MXNET-1226] add Docs update for MXNet Java (#14395)

* add Docs update for MXNet Java

* fixed based on comments

* Update docs/install/java_setup.md

Co-Authored-By: lanking520 

* Update docs/tutorials/java/mxnet_java_on_intellij.md

Co-Authored-By: lanking520 
---
 docs/install/index.md |  6 ++---
 docs/install/java_setup.md| 32 ++-
 docs/tutorials/java/mxnet_java_on_intellij.md | 17 +++---
 3 files changed, 13 insertions(+), 42 deletions(-)

diff --git a/docs/install/index.md b/docs/install/index.md
index eab1c82..f1e959e 100644
--- a/docs/install/index.md
+++ b/docs/install/index.md
@@ -505,7 +505,7 @@ You can use the Maven packages defined in the following 
dependency to include MX
 
 You can use the Maven packages defined in the following dependency to include 
MXNet in your Java project. The Java API is provided as a subset of the Scala 
API and is intended for inference only. Please refer to the MXNet-Java setup guide for a detailed set of 
instructions to help you with the setup process.
 
-https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0-SNAPSHOT~~";>https://img.shields.io/badge/org.apache.mxnet-linux gpu-green.svg" 
alt="maven badge"/>
+https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0~~";>https://img.shields.io/badge/org.apache.mxnet-linux gpu-green.svg" 
alt="maven badge"/>
 
 ```html
 
@@ -522,7 +522,7 @@ You can use the Maven packages defined in the following 
dependency to include MX
 
 You can use the Maven packages defined in the following dependency to include 
MXNet in your Java project. The Java API is provided as a subset of the Scala 
API and is intended for inference only. Please refer to the MXNet-Java setup guide for a detailed set of 
instructions to help you with the setup process.
 
-https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0-SNAPSHOT~~";>https://img.shields.io/badge/org.apache.mxnet-linux cpu-green.svg" 
alt="maven badge"/>
+https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0~~";>https://img.shields.io/badge/org.apache.mxnet-linux cpu-green.svg" 
alt="maven badge"/>
 
 ```html
 
@@ -810,7 +810,7 @@ Not available at this time. 
 
 You can use the Maven packages defined in the following dependency to include 
MXNet in your Java project. The Java API is provided as a subset of the Scala 
API and is intended for inference only. Please refer to the MXNet-Java setup guide for a detailed set of 
instructions to help you with the setup process.
 
-https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0-SNAPSHOT~~";>https://img.shields.io/badge/org.apache.mxnet-mac cpu-green.svg" 
alt="maven badge"/>
+https://repository.apache.org/#nexus-search;gav~org.apache.mxnet~~1.4.0~~";>https://img.shields.io/badge/org.apache.mxnet-mac cpu-green.svg" 
alt="maven badge"/>
 
 ```html
 
diff --git a/docs/install/java_setup.md b/docs/install/java_setup.md
index 569f439..ce4a4ab 100644
--- a/docs/install/java_setup.md
+++ b/docs/install/java_setup.md
@@ -35,21 +35,15 @@ The following instructions are provided for macOS and 
Ubuntu. Windows is not yet
 brew update
 brew tap caskroom/versions
 brew cask install java8
-brew install opencv
 brew install maven
 ```
 
 **Ubuntu Steps**
 
-These scripts will install Maven and its dependencies. You will be running the 
Scala scripts because the MXNet-Java project has a dependency on the 
MXNet-Scala project.
+Please run the following lines:
 
 ```bash
-wget 
https://raw.githubusercontent.com/apache/incubator-mxnet/master/ci/docker/install/ubuntu_core.sh
-wget 
https://raw.githubusercontent.com/apache/incubator-mxnet/master/ci/docker/install/ubuntu_scala.sh
-chmod +x ubuntu_core.sh
-chmod +x ubuntu_scala.sh
-sudo ./ubuntu_core.sh
-sudo ./ubuntu_scala.sh
+sudo apt-get install openjdk-8-java maven
 ```
 
 **Step 2.** Run the demo MXNet-Java project.
@@ -58,25 +52,14 @@ Go to the [MXNet-Java demo project's 
README](https://github.com/apache/incubator
 
  Maven Repository
 
-MXNet-Java can be easily included in your Maven managed project. The Java 
packages are currently available as nightly builds on Maven. Add the following 
Maven repository to your `pom.xml` to fetch the Java packages :
-
-```html
-
-
-  Apache Snapshot
-  https://repository.apache.org/content/groups/snapshots
-
-
-```
-
-Also, add the dependency which corresponds to your platform to the 

[GitHub] [incubator-mxnet] lanking520 merged pull request #14395: [MXNET-1226] add Docs update for MXNet Java

[GitBox]

lanking520 merged pull request #14395: [MXNET-1226] add Docs update for MXNet 
Java
URL: https://github.com/apache/incubator-mxnet/pull/14395
 
 
   


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] szha commented on a change in pull request #14313: compatibility with opencv4

[GitBox]

szha commented on a change in pull request #14313: compatibility with opencv4
URL: https://github.com/apache/incubator-mxnet/pull/14313#discussion_r265309054
 
 

 ##
 File path: Makefile
 ##
 @@ -146,11 +146,33 @@ endif
 
 # setup opencv
 ifeq ($(USE_OPENCV), 1)
-   CFLAGS += -DMXNET_USE_OPENCV=1 $(shell pkg-config --cflags opencv)
-   LDFLAGS += $(filter-out -lopencv_ts, $(shell pkg-config --libs opencv))
+   CFLAGS += -DMXNET_USE_OPENCV=1
+   ifneq ($(USE_OPENCV_INC_PATH), NONE)
 
 Review comment:
   @reminisce good catch. @wkcn how about we add a default value for empty 
value here to help with cases where people already have old config.mk without 
the new 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] reminisce commented on a change in pull request #14313: compatibility with opencv4

[GitBox]

reminisce commented on a change in pull request #14313: compatibility with 
opencv4
URL: https://github.com/apache/incubator-mxnet/pull/14313#discussion_r265308339
 
 

 ##
 File path: Makefile
 ##
 @@ -146,11 +146,33 @@ endif
 
 # setup opencv
 ifeq ($(USE_OPENCV), 1)
-   CFLAGS += -DMXNET_USE_OPENCV=1 $(shell pkg-config --cflags opencv)
-   LDFLAGS += $(filter-out -lopencv_ts, $(shell pkg-config --libs opencv))
+   CFLAGS += -DMXNET_USE_OPENCV=1
+   ifneq ($(USE_OPENCV_INC_PATH), NONE)
 
 Review comment:
   This can be problematic when `USE_OPENCV_INC_PATH` is not defined in 
`config.mk`, because it's an empty string at the time. In that situation, no 
correct paths to headers and libs of opencv are actually added to `CFLAGS` and 
`LDFLAGS`.


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.

[zhasheng]

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

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


The following commit(s) were added to refs/heads/asf-site by this push:
 new 4f27d9f  Bump the publish timestamp.
4f27d9f is described below

commit 4f27d9f3b852023162e038585a3c010077f39b7e
Author: mxnet-ci 
AuthorDate: Wed Mar 13 19:24:50 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..5788aa6
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Wed Mar 13 19:24:50 UTC 2019

[incubator-mxnet] branch master updated: Cudnn conv dgrad algo filtering (#14310)

[marcoabreu]

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

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


The following commit(s) were added to refs/heads/master by this push:
 new ce99e49  Cudnn conv dgrad algo filtering (#14310)
ce99e49 is described below

commit ce99e49e86ee5f187a3d8cc6b029a71652a53af4
Author: Dick Carter 
AuthorDate: Wed Mar 13 12:17:05 2019 -0700

Cudnn conv dgrad algo filtering (#14310)

* Add test exposing issue with conv dgrad algo 3 for some cudnn's.

* Add test temporarily to tests run with tensorrt CI build (cuda10, 
cudnn7.4.2)

* Relax tol of new test.

* Fix for problematic conv dgrad algo 3 for some cuDNNs.

* Add algo exclusion term to cudnnFind result processing.

* Revert "Add test temporarily to tests run with tensorrt CI build (cuda10, 
cudnn7.4.2)"

This reverts commit 1cb743bd2313ca5d1845dbdda6cccd9dcaa5d30c.

* Trigger CI.

* Add link to cuDNN release notes.

* Trigger CI.
---
 src/operator/nn/cudnn/cudnn_convolution-inl.h   | 25 
 src/operator/nn/cudnn/cudnn_deconvolution-inl.h | 24 ---
 tests/python/gpu/test_operator_gpu.py   | 26 +
 3 files changed, 68 insertions(+), 7 deletions(-)

diff --git a/src/operator/nn/cudnn/cudnn_convolution-inl.h 
b/src/operator/nn/cudnn/cudnn_convolution-inl.h
index 66df82e..55b2638 100644
--- a/src/operator/nn/cudnn/cudnn_convolution-inl.h
+++ b/src/operator/nn/cudnn/cudnn_convolution-inl.h
@@ -521,7 +521,19 @@ class CuDNNConvolutionOp {
 wshape[1],
 wshape[2],
 wshape[3]));
-
+#if CUDNN_VERSION >= 7301 && CUDNN_VERSION < 7500
+  auto kernel_h = wshape[2];
+  auto kernel_w = wshape[3];
+  auto stride_h = stride[0];
+  auto stride_w = stride[1];
+  auto pad_h = pad[0];
+  auto pad_w = pad[1];
+  if (param_.layout.value() == kNCHW &&
+  (((stride_h == 2) && (kernel_h % 2 == 0) && (pad_h % 2 == 0)) ||
+   ((stride_w == 2) && (kernel_w % 2 == 0) && (pad_w % 2 == 0 {
+exclude_dgrad_algo_ = CUDNN_CONVOLUTION_BWD_DATA_ALGO_FFT_TILING;
+  }
+#endif
 } else if (param_.kernel.ndim() == 3) {
   // 3d conv
   #if CUDNN_MAJOR >= 5
@@ -714,7 +726,7 @@ class CuDNNConvolutionOp {
 bwd_data_results.resize(actual_bwd_data_algos);
 AlgoFinalSelect(bwd_data_results, 
"backprop-to-data",
-   workspace_byte, bwd);
+   workspace_byte, bwd, 
exclude_dgrad_algo_);
 #else
 // CUDNN_MAJOR < 7
 const int kMaxAlgos = 10;
@@ -910,12 +922,14 @@ class CuDNNConvolutionOp {
   // workspace constraints.
   template 
   void AlgoFinalSelect(const std::vector &perf_results, std::string 
kernel_name,
-   size_t workspace_byte, CuDNNAlgo *algo) {
+   size_t workspace_byte, CuDNNAlgo *algo,
+   int32_t algo_exclude = -1) {
 // Determine the fastest acceptable algo that matches the algo_preference 
(-1 = any),
 // regardless of mathType.
 bool enforce_determinism = dmlc::GetEnv("MXNET_ENFORCE_DETERMINISM", 
false);
 for (decltype(perf_results.size()) i = 0; i != perf_results.size(); ++i) {
   const auto &result = perf_results[i];
+  bool algo_exclusion = static_cast(result.algo) == algo_exclude;
   bool algo_is_tensor_core = false;
   #if CUDNN_MAJOR >= 7
 algo_is_tensor_core = result.mathType == CUDNN_TENSOR_OP_MATH;
@@ -924,7 +938,8 @@ class CuDNNConvolutionOp {
 #if CUDNN_MAJOR >= 7
   (!enforce_determinism || result.determinism == 
cudnnDeterminism_t::CUDNN_DETERMINISTIC) &&
 #endif
-  (param_.cudnn_tune.value() == conv::kFastest || result.memory <= 
workspace_byte)) {
+  (param_.cudnn_tune.value() == conv::kLimited || result.memory <= 
workspace_byte) &&
+  !algo_exclusion) {
 algo->Set(result.algo, algo_is_tensor_core);
 return;
   }
@@ -1104,6 +1119,8 @@ class CuDNNConvolutionOp {
   bool cudnn_tensor_core_;
   // Is req[kWeight] == conv::kAddTo ?
   bool add_to_weight_;
+  // Is there a dgrad algo that should be avoided (-1 == none)?
+  int32_t exclude_dgrad_algo_ = -1;
   ConvolutionParam param_;
 };
 #endif  // __CUDACC__ && CUDNN
diff --git a/src/operator/nn/cudnn/cudnn_deconvolution-inl.h 
b/src/operator/nn/cudnn/cudnn_deconvolution-inl.h
index ec95d2b..47f688c 100644
--- a/src/operator/nn/cudnn/cudnn_deconvolution-inl.h
+++ b/src/operator/nn/cudnn/cudnn_deconvolution-inl.h
@@ -446,6 +446,19 @@ class CuDNNDeconvolutionOp {
 wshape[1],
 wshape[2],
 

[GitHub] [incubator-mxnet] marcoabreu merged pull request #14310: Cudnn conv dgrad algo filtering

[GitBox]

marcoabreu merged pull request #14310: Cudnn conv dgrad algo filtering
URL: https://github.com/apache/incubator-mxnet/pull/14310
 
 
   


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 #14421: Updating mxnet from 1.0.0, network give different ouputs

[GitBox]

mxnet-label-bot commented on issue #14421: Updating mxnet from 1.0.0, network 
give different ouputs
URL: 
https://github.com/apache/incubator-mxnet/issues/14421#issuecomment-472560478
 
 
   Hey, this is the MXNet Label Bot. 
Thank you for submitting the issue! I will try and suggest some labels so 
that the appropriate MXNet community members can help resolve it. 
Here are my recommended labels: 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] jmerkow opened a new issue #14421: Updating mxnet from 1.0.0, network give different ouputs

[GitBox]

jmerkow opened a new issue #14421: Updating mxnet from 1.0.0, network give 
different ouputs
URL: https://github.com/apache/incubator-mxnet/issues/14421
 
 
   I am working in a production environment, where have some networks 
implemented in mxnet 1.0.0. I am working updating our systems and trying to 
push to the latest mxnet (1.4.x as of now) but when we upgrade our networks 
produce different outputs.
   We are using symbols saved to json files and arg/aux_params stored in 
.params files. these were all produced by mxnet 1.0.0 or earlier.
   
   When using the latest mxnet (or 1.4.x) we are getting different outputs for 
the same inputs, with our saved models. I have been trying to use git bisect or 
slowly upgrading versions to figure out where this breaking change occurred but 
there are issues with your git history and/or some strange (compiler??) 
incompatibilities which prevent getting a clean checkout/build for nearly all 
of the intermediate versions…
   
   These are VERY different outputs, im rounding to about the 5 decimal point, 
or higher, so these aren’t numerical differences.
   And we are using modules, not gluons (part of the point is upgrade and move 
towards gluons)
   
   Does anyone have any idea what could be causing this? Or what to look into 
to solve it?
   
   


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 numpy updated (d8e9235 -> b4497e7)

[reminisce]

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

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


 discard d8e9235  [numpy] Shape support scalar tensor (#14315)
 new f2497aa  Updated news.md with the latest mkldnn submodule version 
(#14298)
 new 49d7fc6  Enhance gpu quantization (#14094)
 new d6eafca  Bypass ThreadedEngine in 
test_operator_gpu.py:test_convolution_multiple_streams. (#14338)
 new 111b881  Limit workspace for cudnnGet results (#14326)
 new 19d737f  [MXNET-1331] Removal of non-MXNET classes from JAR (#14303)
 new 184c2a5  fix render issue in NDArray linalg docs (#14258)
 new fccce20  Add more support for mxnet_to_coreml (#14222)
 new a0f3f92   Add default parameters for Scala NDArray.arange (#13816)
 new b486594  Register fake grad to subgraph and quantized operators 
(#14275)
 new 39412b3  corrected a spellign (#14247)
 new 83d2c2d  [MXNET-1324] Add NaiveRunGraph to imperative utils (#14192)
 new 8ab7998  Updates build_lib.sh to copy the cub library license (#14347)
 new efb8823  Add MKLDNN headers to pip package (#14339)
 new 7b8e3a9  compatibility with opencv4 (#14313)
 new f4ab2d7  [MXNET-1291] solve pylint errors in examples with issue 
no.12205 (#13848)
 new 49932fa  #14199: catch subprocess.CalledProcessError in get_gpus() 
(#14212)
 new 6caaa38  print error message for mxnet::cpp::Operator::Invoke when 
failed (#14318)
 new 8beea18  Bulked op segments to allow Variable nodes (#14200)
 new e703694  Fixes #14181, validate model output shape for ObjectDetector. 
(#14215)
 new 12c41e6  Optimizer MXKVStoreUpdater bug fix in serializeState method 
(#14337)
 new 2b7d57d  Installs qemu pip requirements from qemu requirements file 
(#14355)
 new 838e256  Optimize NMS part 2 (#14352)
 new 30b1cbc  add exception (#14362)
 new 8668db7  MKLDNN based Quantized FullyConnected Operator and its fusion 
(#14128)
 new ce9e3cf  add pos_weight for SigmoidBinaryCrossEntropyLoss (#13612)
 new ed83071  Julia: split symbolic-node.jl into several snippets (#14024)
 new c645591  Fix NaN value comparisons in relu, max and min ops (#14262)
 new 8be97d7  [clojure-package][wip] add `->nd-vec` function in 
`ndarray.clj` (#14308)
 new 35098b8  support leading dimension of -1 in ravel/unravel (#14356)
 new 4f5cba5  fix engine crash in shutdown phase (#14382)
 new 47d4d66  Flaky test 
https://github.com/apache/incubator-mxnet/issues/14189 (#14190)
 new af41af5  Julia: rename `mx.clip` to `clamp` for `NDArray` (#14027)
 new 2df5756  add backgroud class in box_nms (#14058)
 new 0e8c270  CI Changes for Codified Windows AMIs (#14336)
 new a4b9802  [Clojure] Helper function for n-dim vector to ndarray (#14305)
 new 89bebd1  [DOC] fix sym.arange doc (#14237)
 new 73b29fa  Julia: add binding for runtime feature detection (#13992)
 new ab0ca86  [MXNET-1093] Add python3 Docker images for each MXNet release 
(#12791)
 new 66c74cc  Enable bulking test on windows (#14392)
 new 6aa8c27  [MXNET-1327] Allow RNN Layers to be initialized to fp16 
(#14219)
 new c4cae6e  Disables flaky test_operator.test_sgld test (#14410)
 new 82504ad  Fix relative difference scala (#14417)
 new b4497e7  [numpy] Shape support scalar tensor (#14315)

This update added new revisions after undoing existing revisions.
That is to say, some revisions that were in the old version of the
branch are not in the new version.  This situation occurs
when a user --force pushes a change and generates a repository
containing something like this:

 * -- * -- B -- O -- O -- O   (d8e9235)
\
 N -- N -- N   refs/heads/numpy (b4497e7)

You should already have received notification emails for all of the O
revisions, and so the following emails describe only the N revisions
from the common base, B.

Any revisions marked "omit" are not gone; other references still
refer to them.  Any revisions marked "discard" are gone forever.

The 9389 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails.  The revisions
listed as "add" were already present in the repository and have only
been added to this reference.


Summary of changes:
 CONTRIBUTORS.md|   5 +-
 Makefile   |  30 +-
 NEWS.md|   2 +-
 benchmark/python/control_flow/rnn.py   |   7 +-
 ci/build_windows.py|   6 +-
 ci/docker/qemu/runtime_functions.py|   4 +-
 ci/docker/qemu/vmcontrol.py|   1 +
 ci/jenkins/Jenkins_steps.groovy|  32 +-
 ci/jenkins/Jenkinsfile_windows_cpu |  10 +-
 ci/jenkins/Jenkinsfile_windows_gpu |  10 +-
 ci/qemu/README.md  

[GitHub] [incubator-mxnet] karan6181 commented on issue #14403: Fix slice op issues #13760 and #14105

[GitBox]

karan6181 commented on issue #14403: Fix slice op issues #13760 and #14105
URL: https://github.com/apache/incubator-mxnet/pull/14403#issuecomment-472554335
 
 
   @mxnet-label-bot add [Operator, pr-awaiting-review]


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] karan6181 commented on issue #14420: v1.4.x: Backport of Windows CI Fixes

[GitBox]

karan6181 commented on issue #14420: v1.4.x: Backport of Windows CI Fixes
URL: https://github.com/apache/incubator-mxnet/pull/14420#issuecomment-472553316
 
 
   @perdasilva Thank you for the contribution!. Could you please look into it 
why CI if failing?
   Thanks!


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


With regards,
Apache Git Services

[GitHub] [incubator-mxnet] karan6181 commented on issue #14420: v1.4.x: Backport of Windows CI Fixes

[GitBox]

karan6181 commented on issue #14420: v1.4.x: Backport of Windows CI Fixes
URL: https://github.com/apache/incubator-mxnet/pull/14420#issuecomment-472552899
 
 
   @mxnet-label-bot add [CI, Windows, pr-awaiting-testing]


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] karan6181 commented on issue #14412: Disables flaky TestStochasticTiming_2D test

[GitBox]

karan6181 commented on issue #14412: Disables flaky TestStochasticTiming_2D test
URL: https://github.com/apache/incubator-mxnet/pull/14412#issuecomment-472551337
 
 
   @mxnet-label-bot add [C++, Test, pr-awaiting-review]


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] karan6181 commented on issue #14409: [Numpy] Change semantics of ndim for operators in `src/operator/contrib`

[GitBox]

karan6181 commented on issue #14409: [Numpy] Change semantics of ndim for 
operators in `src/operator/contrib`
URL: https://github.com/apache/incubator-mxnet/pull/14409#issuecomment-472550377
 
 
   @mxnet-label-bot add [Numpy, Operator, pr-awaiting-review]


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] karan6181 commented on issue #14407: Update imagenet quantization script for MKL-DNN

[GitBox]

karan6181 commented on issue #14407: Update imagenet quantization script for 
MKL-DNN
URL: https://github.com/apache/incubator-mxnet/pull/14407#issuecomment-472549136
 
 
   @mxnet-label-bot add [Quantization, MKLDNN]


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 issue #14406: resnet cpp-package test is broken

[GitBox]

anirudh2290 commented on issue #14406: resnet cpp-package test is broken
URL: 
https://github.com/apache/incubator-mxnet/issues/14406#issuecomment-472548635
 
 
   as i said this happened in waitall change. waitall earlier used to hide 
exceptions, but with the PR: 
https://github.com/apache/incubator-mxnet/pull/14397 it is thrown. These 
problems would have been there from before but surfacing 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] abhinavs95 commented on a change in pull request #14405: [MXNet-1343][WIP][Fit API]Add CNN integration test for fit() API

[GitBox]

abhinavs95 commented on a change in pull request #14405: [MXNet-1343][WIP][Fit 
API]Add CNN integration test for fit() API
URL: https://github.com/apache/incubator-mxnet/pull/14405#discussion_r265270422
 
 

 ##
 File path: tests/nightly/estimator/test_estimator_cnn_gpu.py
 ##
 @@ -0,0 +1,96 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Test gluon estimator on GPU using ResNet18
+
+import os
+import sys
+import mxnet as mx
+from mxnet import gluon
+from mxnet.gluon import data
+from mxnet.gluon.estimator import estimator, event_handler
+from mxnet.gluon.model_zoo import vision
+
+def load_data_mnist(batch_size, resize=None, num_workers=None,
+root=os.path.join('~', '.mxnet', 'datasets', 'mnist')):
+'''
+Load MNIST dataset
+'''
+root = os.path.expanduser(root)  # Expand the user path '~'.
+transformer = []
+if resize:
+transformer += [data.vision.transforms.Resize(resize)]
+transformer += [data.vision.transforms.ToTensor()]
+transformer = data.vision.transforms.Compose(transformer)
+mnist_train = data.vision.MNIST(root=root, train=True)
+mnist_test = data.vision.MNIST(root=root, train=False)
+
+if num_workers is None:
+num_workers = 0 if sys.platform.startswith('win32') else 4
+
+train_iter = data.DataLoader(
+mnist_train.transform_first(transformer), batch_size, shuffle=True,
+num_workers=num_workers)
+test_iter = data.DataLoader(
+mnist_test.transform_first(transformer), batch_size, shuffle=False,
+num_workers=num_workers)
+return train_iter, test_iter
+
+def test_estimator():
+'''
+Test estimator by training resnet18_v1 for 5 epochs on MNIST and verify 
accuracy
+'''
+model_name = 'resnet18_v1'
+batch_size = 128
+num_epochs = 5
+input_size = 224
+lr = 0.001
+# Set context
+if mx.context.num_gpus() > 0:
+context = mx.gpu(0)
+else:
+context = mx.cpu()
+# Get model
+net = vision.get_model(model_name, classes=10)
+# Load train and validation data
+train_data, test_data = load_data_mnist(batch_size, resize=input_size)
+# Define loss and evaluation metrics
+loss = gluon.loss.SoftmaxCrossEntropyLoss()
+acc = mx.metric.Accuracy()
+# Hybridize and initialize net
+net.hybridize()
+net.initialize(mx.init.MSRAPrelu(), ctx=context)
+# Define trainer
+trainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': lr})
+# Define estimator
+est = estimator.Estimator(net=net,
+  loss=loss,
+  metrics=acc,
+  trainers=trainer,
+  context=context)
+# Call fit() to begin training
+logging_handler = event_handler.LoggingHandler(est, model_name+'_log', 
model_name+'_log')
 
 Review comment:
   same as above


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] abhinavs95 commented on a change in pull request #14405: [MXNet-1343][WIP][Fit API]Add CNN integration test for fit() API

[GitBox]

abhinavs95 commented on a change in pull request #14405: [MXNet-1343][WIP][Fit 
API]Add CNN integration test for fit() API
URL: https://github.com/apache/incubator-mxnet/pull/14405#discussion_r265270334
 
 

 ##
 File path: tests/nightly/estimator/test_estimator_cnn_cpu.py
 ##
 @@ -0,0 +1,106 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Test gluon estimator on CPU using CNN models
+
+import numpy as np
+import mxnet as mx
+from mxnet import gluon, init, nd
+from mxnet.gluon.estimator import estimator, event_handler
+from mxnet.gluon.model_zoo import vision
+
+def bilinear_kernel(in_channels, out_channels, kernel_size):
+'''
+Bilinear interpolation using transposed convolution
+'''
+factor = (kernel_size + 1) // 2
+if kernel_size % 2 == 1:
+center = factor - 1
+else:
+center = factor - 0.5
+og = np.ogrid[:kernel_size, :kernel_size]
+filt = (1 - abs(og[0] - center) / factor) * (1 - abs(og[1] - center) / 
factor)
+weight = np.zeros((in_channels, out_channels, kernel_size, kernel_size), 
dtype='float32')
+weight[range(in_channels), range(out_channels), :, :] = filt
+return nd.array(weight)
+
+def FCN(num_classes=21, ctx=None):
+'''
+FCN model for semantic segmentation
+'''
+pretrained_net = vision.resnet18_v2(pretrained=True, ctx=ctx)
+
+net = gluon.nn.HybridSequential()
+for layer in pretrained_net.features[:-2]:
+net.add(layer)
+
+net.add(gluon.nn.Conv2D(num_classes, kernel_size=1),
+gluon.nn.Conv2DTranspose(num_classes, kernel_size=64, padding=16, 
strides=32))
+return net
+
+def test_estimator():
+'''
+Test estimator by doing one pass over each model with synthetic data
+'''
+models = ['resnet18_v1',
+  'alexnet',
+  'FCN'
+  ]
+context = mx.cpu()
+for model_name in models:
+batch_size = 1
+num_epochs = 1
+lr = 0.001
+# Get model
+if model_name is 'FCN':
+num_classes = 21
+net = FCN(num_classes=num_classes, ctx=context)
+dataset = 
gluon.data.dataset.ArrayDataset(mx.nd.random.uniform(shape=(batch_size, 3, 320, 
480)),
+  
mx.nd.zeros(shape=(batch_size, 320, 480)))
+loss = gluon.loss.SoftmaxCrossEntropyLoss(axis=1)
+net[-1].initialize(init.Constant(bilinear_kernel(num_classes, 
num_classes, 64)), ctx=context)
+net[-2].initialize(init=init.Xavier(), ctx=context)
+else:
+net = vision.get_model(model_name, classes=10)
+dataset = 
gluon.data.dataset.ArrayDataset(mx.nd.random.uniform(shape=(batch_size, 1, 224, 
224)),
+  mx.nd.zeros(batch_size))
+loss = gluon.loss.SoftmaxCrossEntropyLoss()
+net.initialize(mx.init.MSRAPrelu(), ctx=context)
+
+train_data = gluon.data.DataLoader(dataset, batch_size=batch_size)
+# Define loss and evaluation metrics
+acc = mx.metric.Accuracy()
+# Hybridize and initialize net
+net.hybridize()
+# Define trainer
+trainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 
lr})
+# Define estimator
+est = estimator.Estimator(net=net,
+  loss=loss,
+  metrics=acc,
+  trainers=trainer,
+  context=context)
+# Call fit() to begin training
+logging_handler = event_handler.LoggingHandler(est, model_name+'_log', 
model_name+'_log')
 
 Review comment:
   Thanks for pointing it out. Removing logging handler from integration tests 
for now as its undergoing changes. It will be covered in unit tests.


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