[GitHub] [incubator-mxnet] szha commented on issue #18751: gluon.nn.BatchNorm seems to swap updated values of moving_mean and moving_var on GPU.

2020-07-17 Thread GitBox
szha commented on issue #18751: URL: https://github.com/apache/incubator-mxnet/issues/18751#issuecomment-660421440 @gilbertfrancois thanks for the report. Does it also happen with the 2.0 on master branch? You can install mxnet-cu102 v2.0.0b20200716 from https://dist.mxnet.io/python

[GitHub] [incubator-mxnet] szha commented on a change in pull request #18748: Fixup move gluon.metric api docs

2020-07-17 Thread GitBox
szha commented on a change in pull request #18748: URL: https://github.com/apache/incubator-mxnet/pull/18748#discussion_r456745543 ## File path: docs/python_docs/python/api/gluon/metric/index.rst ## @@ -0,0 +1,23 @@ +.. Licensed to the Apache Software Foundation (ASF) under

[GitHub] [incubator-mxnet] szha commented on pull request #18721: Remove Makefile

2020-07-17 Thread GitBox
szha commented on pull request #18721: URL: https://github.com/apache/incubator-mxnet/pull/18721#issuecomment-660420287 Developers may still need to rely on some of the Makefile targets for development purpose. Make sure that they are documented prominently, as we can't ask developers to

[GitHub] [incubator-mxnet] szha commented on pull request #18749: Refactor Gluon parameter serialization format

2020-07-17 Thread GitBox
szha commented on pull request #18749: URL: https://github.com/apache/incubator-mxnet/pull/18749#issuecomment-660419990 Also, I'd like to see backend serialization/deserialization happen since it likely will be needed. If we merge this for python only first and there's any issue in adding

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18721: Remove Makefile

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18721: URL: https://github.com/apache/incubator-mxnet/pull/18721#issuecomment-660418477 Jenkins CI successfully triggered : [edge, sanity, unix-cpu, miscellaneous, unix-gpu, centos-gpu, windows-cpu, clang, website, centos-cpu, windows-gpu]

[GitHub] [incubator-mxnet] leezu commented on pull request #18721: Remove Makefile

2020-07-17 Thread GitBox
leezu commented on pull request #18721: URL: https://github.com/apache/incubator-mxnet/pull/18721#issuecomment-660418464 @mxnet-bot run ci [all] This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18748: Fixup move gluon.metric api docs

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18748: URL: https://github.com/apache/incubator-mxnet/pull/18748#issuecomment-660418433 Jenkins CI successfully triggered : [edge, sanity, unix-cpu, miscellaneous, unix-gpu, centos-gpu, windows-cpu, clang, website, centos-cpu, windows-gpu]

[GitHub] [incubator-mxnet] leezu commented on pull request #18748: Fixup move gluon.metric api docs

2020-07-17 Thread GitBox
leezu commented on pull request #18748: URL: https://github.com/apache/incubator-mxnet/pull/18748#issuecomment-660418425 @mxnet-bot run ci [all] This is an automated message from the Apache Git Service. To respond to the

[incubator-mxnet] branch master updated: Revert "Add qr backward for wide matrices with m < n (#18197)" (#18750)

2020-07-17 Thread dickjc123
This is an automated email from the ASF dual-hosted git repository. dickjc123 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 444a7ee Revert "Add qr backward

[incubator-mxnet] branch master updated: Revert "Add qr backward for wide matrices with m < n (#18197)" (#18750)

2020-07-17 Thread dickjc123
This is an automated email from the ASF dual-hosted git repository. dickjc123 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 444a7ee Revert "Add qr backward

[GitHub] [incubator-mxnet] DickJC123 merged pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
DickJC123 merged pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [incubator-mxnet] Kh4L commented on a change in pull request #18749: Refactor Gluon parameter serialization format

2020-07-17 Thread GitBox
Kh4L commented on a change in pull request #18749: URL: https://github.com/apache/incubator-mxnet/pull/18749#discussion_r456731407 ## File path: python/mxnet/gluon/block.py ## @@ -336,41 +337,43 @@ def _collect_params_with_prefix(self, prefix='', select=None):

[GitHub] [incubator-mxnet] leezu commented on pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
leezu commented on pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750#issuecomment-660399925 @mxnet-bot run ci [unix-cpu] This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750#issuecomment-660399950 Jenkins CI successfully triggered : [unix-cpu] This is an automated message from the Apache Git

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

2020-07-17 Thread aaronmarkham
This is an automated email from the ASF dual-hosted git repository. aaronmarkham 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 f392469 Bump the

[incubator-mxnet-site] branch asf-site updated: Publish triggered by CI

2020-07-17 Thread aaronmarkham
This is an automated email from the ASF dual-hosted git repository. aaronmarkham 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 289ba46 Publish

[GitHub] [incubator-mxnet] ksindwan commented on issue #14387: problem in metric.update(0, record) when train with multi GPU

2020-07-17 Thread GitBox
ksindwan commented on issue #14387: URL: https://github.com/apache/incubator-mxnet/issues/14387#issuecomment-660392074 Facing the same issue This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] gilbertfrancois commented on issue #18100: curand error in tests

2020-07-17 Thread GitBox
gilbertfrancois commented on issue #18100: URL: https://github.com/apache/incubator-mxnet/issues/18100#issuecomment-660388871 Could the failing `test_operator.py::test_batchnorm` be related to issue https://github.com/apache/incubator-mxnet/issues/18751 ?

[GitHub] [incubator-mxnet] DickJC123 opened a new issue #18753: PRs that have passed CI, but have "aged" considerably w.r.t. master, can still be merged.

2020-07-17 Thread GitBox
DickJC123 opened a new issue #18753: URL: https://github.com/apache/incubator-mxnet/issues/18753 ## Description Could our system be changed to prevent merging of a CI-passing PR after a "merge window" has elapsed? I guess this is not normally a problem since after a time an auto-merge

[GitHub] [incubator-mxnet] access2rohit commented on pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
access2rohit commented on pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#issuecomment-660388094 @josephevans Thanks for contribution to Large Tensor Support project. Also paste the output of your test runs as a comment or in the description in the PR.

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#discussion_r456721056 ## File path: src/operator/tensor/la_op-inl.h ## @@ -36,7 +36,7 @@ using namespace mshadow; // Copies lower/upper triangular part to

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#discussion_r456721056 ## File path: src/operator/tensor/la_op-inl.h ## @@ -36,7 +36,7 @@ using namespace mshadow; // Copies lower/upper triangular part to

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#discussion_r456721100 ## File path: src/operator/tensor/la_op-inl.h ## @@ -36,7 +36,7 @@ using namespace mshadow; // Copies lower/upper triangular part to

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#discussion_r456721100 ## File path: src/operator/tensor/la_op-inl.h ## @@ -36,7 +36,7 @@ using namespace mshadow; // Copies lower/upper triangular part to

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752#issuecomment-660387316 Hey @josephevans , Thanks for submitting the PR All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the

[GitHub] [incubator-mxnet] josephevans opened a new pull request #18752: [WIP] Fix linalg_potri operator for large tensor.

2020-07-17 Thread GitBox
josephevans opened a new pull request #18752: URL: https://github.com/apache/incubator-mxnet/pull/18752 ## Description ## Fix linalg_potri operator when using large tensor size. Unit tests to come. This is an

[GitHub] [incubator-mxnet] gilbertfrancois opened a new issue #18751: gluon.nn.BatchNorm seems to swap updated values of moving_mean and moving_var on GPU.

2020-07-17 Thread GitBox
gilbertfrancois opened a new issue #18751: URL: https://github.com/apache/incubator-mxnet/issues/18751 ## Description The implementation of `gluon.nn.BatchNorm` behaves different on CPU and GPU context. I suspect that during the update of running_mean and running_var, the updated

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 edited a comment on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660381271 > Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html > Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 edited a comment on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660381271 > Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html > Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660382958 > > Yes, my plan is to copy the get_started.html directly from master branch to v1.6 static artifact (maybe need some change in the header) > > Thank you. Do you

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 edited a comment on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660381271 > Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html > Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 edited a comment on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660381271 > Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html > Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660381271 > Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html > Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] leezu commented on pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
leezu commented on pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750#issuecomment-660381244 @hzfan please don't merge PRs without verifying that the last CI run is recent. In this case you merged a PR where the CI has not verified it for multiple months.

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750#issuecomment-660380906 Hey @leezu , Thanks for submitting the PR All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following

[GitHub] [incubator-mxnet] leezu opened a new pull request #18750: Revert "[Numpy] Add qr backward part 2 for wide matrices with m < n"

2020-07-17 Thread GitBox
leezu opened a new pull request #18750: URL: https://github.com/apache/incubator-mxnet/pull/18750 Reverts apache/incubator-mxnet#18197 as test_np_linalg_qr does not pass and it blocks all development. @D-Roberts feel free to submit a fix if you have one ready; alternatively I

[incubator-mxnet] 01/01: Revert "Add qr backward for wide matrices with m < n (#18197)"

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a commit to branch revert-18197-qr_back_two in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git commit bd52fb75af3406e7618440fc39d44c7ceca62bf9 Author: Leonard Lausen AuthorDate: Fri Jul 17

[incubator-mxnet] branch revert-18197-qr_back_two created (now bd52fb7)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch revert-18197-qr_back_two in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at bd52fb7 Revert "Add qr backward for wide matrices with m < n (#18197)" This branch

[GitHub] [incubator-mxnet] leezu commented on pull request #18749: Refactor Gluon parameter serialization format

2020-07-17 Thread GitBox
leezu commented on pull request #18749: URL: https://github.com/apache/incubator-mxnet/pull/18749#issuecomment-660379339 @ZheyuYe please verify if this addresses https://github.com/apache/incubator-mxnet/issues/18717 This

[GitHub] [incubator-mxnet] leezu opened a new pull request #18749: Refactor Gluon parameter serialization format

2020-07-17 Thread GitBox
leezu opened a new pull request #18749: URL: https://github.com/apache/incubator-mxnet/pull/18749 Switch to npz serialization format https://numpy.org/devdocs/reference/generated/numpy.lib.format.html At this point in time, only Python bindings exist and no functionality is

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18749: Refactor Gluon parameter serialization format

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18749: URL: https://github.com/apache/incubator-mxnet/pull/18749#issuecomment-660379004 Hey @leezu , Thanks for submitting the PR All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following

[incubator-mxnet] branch leezu-patch-3 updated (33698e4 -> 620e070)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. discard 33698e4 Update index.rst add 620e070 Fix metric API page This update added new revisions

[GitHub] [incubator-mxnet] leezu commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
leezu commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660377875 > Yes, my plan is to copy the get_started.html directly from master branch to v1.6 static artifact (maybe need some change in the header) Thank you. Do you mean

[incubator-mxnet] branch leezu-patch-3 updated (33698e4 -> 620e070)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. discard 33698e4 Update index.rst add 620e070 Fix metric API page This update added new revisions

[incubator-mxnet] branch leezu-patch-3 updated (33698e4 -> 620e070)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. discard 33698e4 Update index.rst add 620e070 Fix metric API page This update added new revisions

[incubator-mxnet] branch leezu-patch-3 created (now 33698e4)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at 33698e4 Update index.rst No new revisions were added by this update.

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18748: Fixup move gluon.metric api docs

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18748: URL: https://github.com/apache/incubator-mxnet/pull/18748#issuecomment-660376897 Hey @leezu , Thanks for submitting the PR All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following

[GitHub] [incubator-mxnet] DickJC123 opened a new issue #18747: unittests using @retry decorator can segfault if they fail

2020-07-17 Thread GitBox
DickJC123 opened a new issue #18747: URL: https://github.com/apache/incubator-mxnet/issues/18747 ## Description This is a problem I ran into in the development of PR https://github.com/apache/incubator-mxnet/pull/18694, and **I have include a fix** in commit

[GitHub] [incubator-mxnet] aaronmarkham commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
aaronmarkham commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660376969 Not sure, but this didn't work: http://ec2-34-219-134-42.us-west-2.compute.amazonaws.com/community/contribute.html Shouldn't the dropdown read 1.6 if the

[GitHub] [incubator-mxnet] leezu opened a new pull request #18748: Fixup move gluon.metric api docs

2020-07-17 Thread GitBox
leezu opened a new pull request #18748: URL: https://github.com/apache/incubator-mxnet/pull/18748 Fixup https://github.com/apache/incubator-mxnet/commit/cec86add654fccda2f2112d4487738854e7663ca This is an automated message

[incubator-mxnet] 01/01: Update index.rst

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a commit to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git commit 33698e429d8bb10058c5ab96bfc0c7fb0dcb8d82 Author: Leonard Lausen AuthorDate: Fri Jul 17 16:15:17 2020

[incubator-mxnet] branch leezu-patch-3 created (now 33698e4)

2020-07-17 Thread lausen
This is an automated email from the ASF dual-hosted git repository. lausen pushed a change to branch leezu-patch-3 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at 33698e4 Update index.rst This branch includes the following new commits: new 33698e4 Update

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660372395 > Instead of just testing redirects on the contribute page in the preview, shouldn't we be testing redirects across all pages? This PR is to ensure the redirect

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
ys2843 commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660371019 > If `/versions/1.6/get_started` becomes the default, we need to backport the license fixes to that page. Is it possible? Or do we require using the 1.x branch given

[GitHub] [incubator-mxnet] szha commented on pull request #18574: Update the onnx-tensorrt submodule

2020-07-17 Thread GitBox
szha commented on pull request #18574: URL: https://github.com/apache/incubator-mxnet/pull/18574#issuecomment-660367098 cc @leezu on the build. I think the tests were removed as we are deprecating NDArray in 2.0 and there's additional work needed for supporting numpy operator for

[GitHub] [incubator-mxnet] leezu opened a new issue #17810: GraphExecutor + Numpy + Dynamic shape crash

2020-07-17 Thread GitBox
leezu opened a new issue #17810: URL: https://github.com/apache/incubator-mxnet/issues/17810 ## Description There seem to be some issues with GraphExecutor + Numpy + Dynamic shape. Consider the example below, which leads to a crash if `mx.npx.set_np()` is present and works fine if

[GitHub] [incubator-mxnet] szha commented on issue #17810: GraphExecutor + Numpy + Dynamic shape crash

2020-07-17 Thread GitBox
szha commented on issue #17810: URL: https://github.com/apache/incubator-mxnet/issues/17810#issuecomment-660366181 This is still a valid combination on 1.7 This is an automated message from the Apache Git Service. To

[GitHub] [incubator-mxnet] Kh4L commented on pull request #18574: Update the onnx-tensorrt submodule

2020-07-17 Thread GitBox
Kh4L commented on pull request #18574: URL: https://github.com/apache/incubator-mxnet/pull/18574#issuecomment-660354283 > Thanks, Serge. How is this tested at the moment? It was supposed to be tested in the tensorrt tests deleted here in this PR

[GitHub] [incubator-mxnet] MoisesHer commented on pull request #18608: Cherry-pick #18310 #18355

2020-07-17 Thread GitBox
MoisesHer commented on pull request #18608: URL: https://github.com/apache/incubator-mxnet/pull/18608#issuecomment-660351179 > Hi @ChaiBapchya , I took a look at the failure of [unix-gpu] which showed TVM compilation error and GPU OOM runtime error, seems it's more like the CI side issue.

[GitHub] [incubator-mxnet] leezu opened a new issue #18746: Numpy Op fallback does not work for functions without return value

2020-07-17 Thread GitBox
leezu opened a new issue #18746: URL: https://github.com/apache/incubator-mxnet/issues/18746 https://github.com/apache/incubator-mxnet/blob/cec86add654fccda2f2112d4487738854e7663ca/python/mxnet/numpy/multiarray.py#L332-L365 Will fail in line 359 if `out is None`. Example

[GitHub] [incubator-mxnet] leezu commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
leezu commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660339978 If `/versions/1.6/get_started` becomes the default, we need to backport the license fixes to that page. Is it possible? Or do we require using the 1.x branch given this

[GitHub] [incubator-mxnet] connorgoggins commented on pull request #18738: set website default version - test redirect

2020-07-17 Thread GitBox
connorgoggins commented on pull request #18738: URL: https://github.com/apache/incubator-mxnet/pull/18738#issuecomment-660340224 Instead of just testing redirects on the contribute page in the preview, shouldn't we be testing redirects across all pages?

[GitHub] [incubator-mxnet] leezu commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
leezu commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660325168 As mentioned, I don't think 30 minutes is an issue here as this is a default feature, will run in parallel to other pipelines which already take 60 minutes.

[GitHub] [incubator-mxnet] leezu edited a comment on issue #18740: test_sparse_operator.py::test_elemwise_binary_ops

2020-07-17 Thread GitBox
leezu edited a comment on issue #18740: URL: https://github.com/apache/incubator-mxnet/issues/18740#issuecomment-660324478 The pytest workers are all in separate processes. I'm only aware of the difference that `OMP_NUM_THREADS=$(expr $(nproc) / 4)` is exported before running the

[GitHub] [incubator-mxnet] leezu commented on issue #18740: test_sparse_operator.py::test_elemwise_binary_ops

2020-07-17 Thread GitBox
leezu commented on issue #18740: URL: https://github.com/apache/incubator-mxnet/issues/18740#issuecomment-660324478 The pytest workers are all in separate processes. I'm only aware of the difference that `OMP_NUM_THREADS=$(expr $(nproc) / 4)` is exported before running the parallel

[GitHub] [incubator-mxnet] access2rohit edited a comment on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
access2rohit edited a comment on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660324294 We can split up and parallelize the jobs to run on multiple workers as a part of multiple test suites in CI. That will speed things up but will consume more

[GitHub] [incubator-mxnet] access2rohit edited a comment on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
access2rohit edited a comment on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660324294 We can split up and parallelize the jobs to run on multiple workers as a part of multiple test suites in CI. That will speed things up but will consume more

[GitHub] [incubator-mxnet] access2rohit commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
access2rohit commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660324294 We can split up and parallelize the jobs to run on multiple works as a part of multiple test suites in CI. That will speed things up but will consume more CI

[GitHub] [incubator-mxnet] leezu commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
leezu commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660323091 @access2rohit there's only one long-running test: ``` ] [2020-02-25T20:36:21.873Z] OK (SKIP=1) [2020-02-25T20:36:21.873Z] + nosetests-3.4

[GitHub] [incubator-mxnet] DickJC123 commented on issue #18740: test_sparse_operator.py::test_elemwise_binary_ops

2020-07-17 Thread GitBox
DickJC123 commented on issue #18740: URL: https://github.com/apache/incubator-mxnet/issues/18740#issuecomment-660322294 I had some success with marking this test with `@pytest.mark.serial` without understanding the underlying issue, or why this action fixed it. Could someone enlighten

[GitHub] [incubator-mxnet] access2rohit commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
access2rohit commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660322071 > Why does the run linked above only take 26m then? I think last successful run should be considered. And if we do then :

[GitHub] [incubator-mxnet] waytrue17 commented on a change in pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
waytrue17 commented on a change in pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#discussion_r456657714 ## File path: src/operator/subgraph/static_shape_subgraph_property.cc ## @@ -0,0 +1,80 @@ +/* + * Licensed to the Apache Software Foundation

[GitHub] [incubator-mxnet] leezu commented on pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
leezu commented on pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#issuecomment-660321410 Backtrace of the issue below. I use the following patch to obtain the backtrace ``` diff modified src/c_api/c_api_ndarray.cc @@ -394,7 +394,6 @@ int

[GitHub] [incubator-mxnet] leezu commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
leezu commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660319898 Why does the run linked above only take 26m then? This is an automated message from the Apache Git Service.

[GitHub] [incubator-mxnet] leezu edited a comment on pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
leezu edited a comment on pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#issuecomment-660317004 This isn't working yet. You can try the following slight extension of the `test_dynamic_shape.py`: ``` diff modified

[GitHub] [incubator-mxnet] leezu commented on pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
leezu commented on pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#issuecomment-660317004 This is broken. You can try the following slight extension of the `test_dynamic_shape.py`: ``` diff modified tests/python/unittest/test_dynamic_shape.py

[GitHub] [incubator-mxnet] access2rohit commented on issue #18745: Run Large Tensor Tests as part of PR tests

2020-07-17 Thread GitBox
access2rohit commented on issue #18745: URL: https://github.com/apache/incubator-mxnet/issues/18745#issuecomment-660315952 @leezu simply running all tests within test_large_array.py takes more than 90mins. I don't think its practical to enable them on CI. It will simply timeout on all

[GitHub] [incubator-mxnet] leezu commented on a change in pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
leezu commented on a change in pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#discussion_r456648950 ## File path: src/operator/subgraph/static_shape_subgraph_property.cc ## @@ -0,0 +1,80 @@ +/* + * Licensed to the Apache Software Foundation

[GitHub] [incubator-mxnet] leezu commented on a change in pull request #18690: [WIP] optimize graph in presence of dynamic shape ops

2020-07-17 Thread GitBox
leezu commented on a change in pull request #18690: URL: https://github.com/apache/incubator-mxnet/pull/18690#discussion_r456645837 ## File path: src/operator/subgraph/static_shape_subgraph_property.cc ## @@ -0,0 +1,80 @@ +/* + * Licensed to the Apache Software Foundation

[GitHub] [incubator-mxnet] leezu closed issue #17810: GraphExecutor + Numpy + Dynamic shape crash

2020-07-17 Thread GitBox
leezu closed issue #17810: URL: https://github.com/apache/incubator-mxnet/issues/17810 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

[GitHub] [incubator-mxnet] leezu commented on issue #17810: GraphExecutor + Numpy + Dynamic shape crash

2020-07-17 Thread GitBox
leezu commented on issue #17810: URL: https://github.com/apache/incubator-mxnet/issues/17810#issuecomment-660298253 GraphExecutor was removed This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] szha commented on pull request #18197: [Numpy] Add qr backward part 2 for wide matrices with m < n

2020-07-17 Thread GitBox
szha commented on pull request #18197: URL: https://github.com/apache/incubator-mxnet/pull/18197#issuecomment-660296403 @D-Roberts @hzfan could you look into the issue that @ptrendx mentioned? If it can't be fixed in a couple of hours let's revert the change first.

[GitHub] [incubator-mxnet] szha commented on pull request #18574: Update the onnx-tensorrt submodule

2020-07-17 Thread GitBox
szha commented on pull request #18574: URL: https://github.com/apache/incubator-mxnet/pull/18574#issuecomment-660295550 Thanks, Serge. How is this tested at the moment? This is an automated message from the Apache Git

[GitHub] [incubator-mxnet] szha commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456629856 ## File path: python/mxnet/numpy/multiarray.py ## @@ -40,8 +40,9 @@ get_oshape_of_gather_nd_op from ..ndarray._internal

[GitHub] [incubator-mxnet] access2rohit commented on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660294002 > @sandeep-krishnamurthy I'm not asking about MKL as feature request. I'm saying that the current build is broken and it needs fix. Basically I will

[GitHub] [incubator-mxnet] szha commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456629049 ## File path: python/mxnet/numpy/multiarray.py ## @@ -92,6 +93,16 @@ _NDARRAY_NO_ZERO_DIM_BOOL_ARRAY = -1 _NDARRAY_ZERO_DIM_BOOL_ARRAY_FALSE = 0

[GitHub] [incubator-mxnet] szha commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456628889 ## File path: tests/python/unittest/test_np_large_array.py ## @@ -0,0 +1,101 @@ +# Licensed to the Apache Software Foundation (ASF) under one +#

[GitHub] [incubator-mxnet] szha commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456629152 ## File path: python/mxnet/numpy/multiarray.py ## @@ -92,6 +93,16 @@ _NDARRAY_NO_ZERO_DIM_BOOL_ARRAY = -1 _NDARRAY_ZERO_DIM_BOOL_ARRAY_FALSE = 0

[GitHub] [incubator-mxnet] szha commented on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660292641 @access2rohit will do. We still need to fix the CI for enabling large tensor first. This is an

[GitHub] [incubator-mxnet] access2rohit commented on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660292618 Overall functionality seems fine. Can you run all the tests in the file `tests/python/unittest/test_np_large_array.py` using pytest and paste the output in the

[GitHub] [incubator-mxnet] szha commented on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
szha commented on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660292353 @sandeep-krishnamurthy I'm not asking about MKL as feature request. I'm saying that the current build is broken and it needs fix.

[GitHub] [incubator-mxnet] access2rohit edited a comment on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit edited a comment on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660291731 @szha Can we move the file: `tests/python/unittest/test_np_large_array.py` to nightly folder instead? Keeping them in unittest will significantly

[GitHub] [incubator-mxnet] access2rohit commented on pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#issuecomment-660291731 @szha Can we move the file: tests/python/unittest/test_np_large_array.py to nightly folder instead? Keeping them in unittest will significantly slowdown the

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456624909 ## File path: tests/python/unittest/test_np_large_array.py ## @@ -0,0 +1,101 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456620176 ## File path: python/mxnet/numpy/multiarray.py ## @@ -92,6 +93,16 @@ _NDARRAY_NO_ZERO_DIM_BOOL_ARRAY = -1

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18490: MXNet-TRT: Add PrePartition param caching - move init_tensorrt_params logic

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18490: URL: https://github.com/apache/incubator-mxnet/pull/18490#issuecomment-660283994 Jenkins CI successfully triggered : [centos-cpu] This is an automated message from the Apache Git

[GitHub] [incubator-mxnet] Kh4L commented on pull request #18490: MXNet-TRT: Add PrePartition param caching - move init_tensorrt_params logic

2020-07-17 Thread GitBox
Kh4L commented on pull request #18490: URL: https://github.com/apache/incubator-mxnet/pull/18490#issuecomment-660283973 @mxnet-bot run ci [centos-cpu] This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-mxnet] access2rohit commented on a change in pull request #18368: [WIP] enable large tensor in np

2020-07-17 Thread GitBox
access2rohit commented on a change in pull request #18368: URL: https://github.com/apache/incubator-mxnet/pull/18368#discussion_r456618261 ## File path: python/mxnet/numpy/multiarray.py ## @@ -92,6 +93,16 @@ _NDARRAY_NO_ZERO_DIM_BOOL_ARRAY = -1

[GitHub] [incubator-mxnet] Kh4L commented on pull request #18574: Update the onnx-tensorrt submodule

2020-07-17 Thread GitBox
Kh4L commented on pull request #18574: URL: https://github.com/apache/incubator-mxnet/pull/18574#issuecomment-66028 @mxnet-bot run ci [unix-gpu] This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18574: Update the onnx-tensorrt submodule

2020-07-17 Thread GitBox
mxnet-bot commented on pull request #18574: URL: https://github.com/apache/incubator-mxnet/pull/18574#issuecomment-660282257 Jenkins CI successfully triggered : [unix-gpu] This is an automated message from the Apache Git

  1   2   >