[GitHub] kobenaxie commented on issue #8634: How to use label smoothing and set 'ignore_label' with gluon like mx.symSoftmaxOutput ?

kobenaxie commented on issue #8634: How to use label smoothing and set 'ignore_label' with gluon like mx.symSoftmaxOutput ? URL: https://github.com/apache/incubator-mxnet/issues/8634#issuecomment-345191383 When computing CE loss in

[GitHub] YeWenting opened a new issue #8694: Error when running "/rcnn/script/additional_deps.sh"

YeWenting opened a new issue #8694: Error when running "/rcnn/script/additional_deps.sh" URL: https://github.com/apache/incubator-mxnet/issues/8694 ## Description I tried to run "/example/rcnn/script/additional_deps.sh". However I get the following **error**:

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345225233 The proper link is now: https://github.com/apache/incubator-mxnet/blob/master/docs/install/ubuntu_setup.md

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345206553 Still struggling to install AI::FANN. For which I need this apparently. > On 16 Nov 2017, at 23:52, Sergey Kolychev

[GitHub] szha commented on issue #7516: [Question] Use LSTM for Sine Wave prediction

szha commented on issue #7516: [Question] Use LSTM for Sine Wave prediction URL: https://github.com/apache/incubator-mxnet/issues/7516#issuecomment-345230962 This issue is closed due to lack of activity in the last 90 days. Feel free to ping me to reopen if this is still an active issue.

[GitHub] szha closed issue #7516: [Question] Use LSTM for Sine Wave prediction

szha closed issue #7516: [Question] Use LSTM for Sine Wave prediction URL: https://github.com/apache/incubator-mxnet/issues/7516 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] sbodenstein commented on issue #8603: Contrib operators for object-detection bounding box related stuffs

sbodenstein commented on issue #8603: Contrib operators for object-detection bounding box related stuffs URL: https://github.com/apache/incubator-mxnet/pull/8603#issuecomment-345217999 > @sbodenstein Functionally, yes, but not identical api. This one provide more options. > Do you have

[GitHub] JeanKossaifi commented on issue #8246: Continued Work on Advanced Indexing

JeanKossaifi commented on issue #8246: Continued Work on Advanced Indexing URL: https://github.com/apache/incubator-mxnet/pull/8246#issuecomment-345268196 Thanks for the great work @reminisce ! This is an automated message

[GitHub] arundasan91 commented on issue #8465: TypeError: zip argument #2 must support iteration

arundasan91 commented on issue #8465: TypeError: zip argument #2 must support iteration URL: https://github.com/apache/incubator-mxnet/issues/8465#issuecomment-345275414 Hi @FCInter, I've had similar issues and found that I missed to pass proper `data_name` and `label_name` to my

[GitHub] JeanKossaifi commented on issue #8084: Support for fancy indexing

JeanKossaifi commented on issue #8084: Support for fancy indexing URL: https://github.com/apache/incubator-mxnet/issues/8084#issuecomment-345268007 Addressed by #8246. This is an automated message from the Apache Git

[GitHub] JeanKossaifi closed issue #8084: Support for fancy indexing

JeanKossaifi closed issue #8084: Support for fancy indexing URL: https://github.com/apache/incubator-mxnet/issues/8084 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub

[GitHub] analog-cbarber commented on issue #8640: CustomOp with more than one auxiliary state segfaults

analog-cbarber commented on issue #8640: CustomOp with more than one auxiliary state segfaults URL: https://github.com/apache/incubator-mxnet/issues/8640#issuecomment-345279473 Local build of master as of ca4b683935529197cc7f7971765794065d67016c does not show this bug on my machine, but

[GitHub] piiswrong commented on a change in pull request #8689: Mark tests that should only be run nightly.

piiswrong commented on a change in pull request #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#discussion_r151759822 ## File path: tests/python/unittest/test_sparse_operator.py ## @@ -1447,7 +1458,9 @@ def

[GitHub] szha commented on issue #8689: Mark tests that should only be run nightly.

szha commented on issue #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#issuecomment-345336009 I understand the motivation for moving tests to nightly. The concern I still have is if we disable certain tests on PR-level and only

[GitHub] doctaweeks commented on issue #8653: error: identifier "__float128" is undefined

doctaweeks commented on issue #8653: error: identifier "__float128" is undefined URL: https://github.com/apache/incubator-mxnet/issues/8653#issuecomment-345348866 I've just tried rebuilding the last commit I built successfully. It failed in the same manor. I suspect this build

[GitHub] eric-haibin-lin commented on issue #7229: How to visualize intermediate layer output that are learnt during training process?

eric-haibin-lin commented on issue #7229: How to visualize intermediate layer output that are learnt during training process? URL: https://github.com/apache/incubator-mxnet/issues/7229#issuecomment-345349024 @zihaolucky is the tensorboard-lite ready?

[GitHub] cjolivier01 commented on issue #8697: Fix linker undefined reference errors

cjolivier01 commented on issue #8697: Fix linker undefined reference errors URL: https://github.com/apache/incubator-mxnet/pull/8697#issuecomment-345350298 Already fixed in another PR This is an automated message from the

[GitHub] cjolivier01 closed pull request #8697: Fix linker undefined reference errors

cjolivier01 closed pull request #8697: Fix linker undefined reference errors URL: https://github.com/apache/incubator-mxnet/pull/8697 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake of provenance: As this is a

[GitHub] tdomhan opened a new pull request #8696: Fixing the monitor callback of the bucketing module.

tdomhan opened a new pull request #8696: Fixing the monitor callback of the bucketing module. URL: https://github.com/apache/incubator-mxnet/pull/8696 ## Description ## install_monitor was broken for the bucketing module, as it would only register the monitor with the buckets present

[GitHub] sergeykolychev commented on issue #6379: Error installing perl bindings

sergeykolychev commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345324898 @davehodg you can run (for example) the calculator.pl from the examples dir by ``` ## see

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345325511 I?m having to revise docker. I haven?t done it for a while. The perils of being in PHP land most of this year. > On 17 Nov

[GitHub] KellenSunderland commented on a change in pull request #8689: Mark tests that should only be run nightly.

KellenSunderland commented on a change in pull request #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#discussion_r151760710 ## File path: tests/python/unittest/test_sparse_operator.py ## @@ -1447,7 +1458,9 @@

[GitHub] KellenSunderland commented on a change in pull request #8689: Mark tests that should only be run nightly.

KellenSunderland commented on a change in pull request #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#discussion_r151760779 ## File path: tests/python/unittest/test_operator.py ## @@ -1266,6 +1267,7 @@ def

[GitHub] piiswrong closed pull request #8643: independent test-related changes pulled from broadcast-tuning branch

piiswrong closed pull request #8643: independent test-related changes pulled from broadcast-tuning branch URL: https://github.com/apache/incubator-mxnet/pull/8643 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake

[GitHub] KellenSunderland commented on issue #8689: Mark tests that should only be run nightly.

KellenSunderland commented on issue #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#issuecomment-345338272 @szha For what it's worth I agree with all of your points. I hope we can work together on updating tests soon (I plan to

[GitHub] cjolivier01 opened a new pull request #8697: Fix linker undefined reference errors

cjolivier01 opened a new pull request #8697: Fix linker undefined reference errors URL: https://github.com/apache/incubator-mxnet/pull/8697 ## Description ## (Brief description on what this PR is about) ## Checklist ## ### Essentials ### - [ ] Passed code style checking

[GitHub] sergeykolychev commented on issue #6379: Error installing perl bindings

sergeykolychev commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345287361 @davehodg AI::FANN ? You've got me confused, how's related to AI::MXNet ?

[GitHub] ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#discussion_r151738034 ## File path: example/sparse/matrix_fact_parallel_model.py ## @@

[GitHub] ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#discussion_r151738034 ## File path: example/sparse/matrix_fact_parallel_model.py ## @@

[GitHub] ptrendx commented on a change in pull request #8294: NCCL integration

ptrendx commented on a change in pull request #8294: NCCL integration URL: https://github.com/apache/incubator-mxnet/pull/8294#discussion_r151744458 ## File path: python/mxnet/model.py ## @@ -104,15 +105,18 @@ def _initialize_kvstore(kvstore, param_arrays, arg_params,

[GitHub] piiswrong commented on a change in pull request #8689: Mark tests that should only be run nightly.

piiswrong commented on a change in pull request #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#discussion_r151760109 ## File path: tests/python/unittest/test_operator.py ## @@ -1266,6 +1267,7 @@ def test_bneq(a,

[incubator-mxnet] branch master updated: Refreshed branch quick_pr (#8643)

This is an automated email from the ASF dual-hosted git repository. jxie pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 3107326 Refreshed branch quick_pr

[GitHub] szha commented on issue #8689: Mark tests that should only be run nightly.

szha commented on issue #8689: Mark tests that should only be run nightly. URL: https://github.com/apache/incubator-mxnet/pull/8689#issuecomment-345338868 @KellenSunderland thanks. The information you collected in this PR while analyzing the tests are valuable. It would be great if we

[GitHub] ptrendx commented on a change in pull request #8294: NCCL integration

ptrendx commented on a change in pull request #8294: NCCL integration URL: https://github.com/apache/incubator-mxnet/pull/8294#discussion_r151759354 ## File path: src/kvstore/comm.h ## @@ -53,13 +74,31 @@ class Comm { */ virtual const NDArray& Reduce( int

[GitHub] sergeykolychev commented on issue #6379: Error installing perl bindings

sergeykolychev commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345301322 @davehodg Cool, thanks. Just in a case you can also get working mxnet/perl via docker from https://hub.docker.com/r/mxnet/perl/

[GitHub] ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#discussion_r151738034 ## File path: example/sparse/matrix_fact_parallel_model.py ## @@

[GitHub] sergeykolychev commented on issue #6379: Error installing perl bindings

sergeykolychev commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345304388 yeah I get it, just wanted to let you know that with already precompiled docker image you do not need to go through trouble of

[GitHub] ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#discussion_r151738034 ## File path: example/sparse/matrix_fact_parallel_model.py ## @@

[GitHub] amithr1 opened a new issue #8695: Hangs training on P100

amithr1 opened a new issue #8695: Hangs training on P100 URL: https://github.com/apache/incubator-mxnet/issues/8695 I am trying to train imagenet using the default resnet on a single node having upto 4 P100s.. When I use the master branch, I see hangs. When I attached gdb I see the

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345301778 I?m still trying to piece things together! > On 17 Nov 2017, at 17:00, Sergey Kolychev wrote: >

[GitHub] ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on a change in pull request #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#discussion_r151738034 ## File path: example/sparse/matrix_fact_parallel_model.py ## @@

[GitHub] hen commented on issue #8688: Restored some copyright attribution that were accidentally removed.

hen commented on issue #8688: Restored some copyright attribution that were accidentally removed. URL: https://github.com/apache/incubator-mxnet/pull/8688#issuecomment-345313429 Looks good to me. +1. This is an automated

[GitHub] ZiyueHuang commented on issue #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

ZiyueHuang commented on issue #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#issuecomment-345322939 Hi @mbaijal Yes. `group2context` seems problematic and we need debug it.

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345317009 What dep am I missing? # docker run mxnet/perl perl -e 'use AI::MXNet qw(mx); my $ar = mx->nd->ones([2,2]); print (($ar*2)->aspdl)'

[GitHub] sergeykolychev commented on issue #6379: Error installing perl bindings

sergeykolychev commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345320043 @davehodg It's just a harmless warning from OpenCV (a prerequisite of mxnet) that does not affect anything, but if you want to get

[GitHub] davehodg commented on issue #6379: Error installing perl bindings

davehodg commented on issue #6379: Error installing perl bindings URL: https://github.com/apache/incubator-mxnet/issues/6379#issuecomment-345288328 It?s AI::MXNet. Please hold. > On 17 Nov 2017, at 16:13, Sergey Kolychev wrote: > >

[GitHub] mbaijal commented on issue #8632: [WIP] a user friendly way to use g2c in module and an example of g2c

mbaijal commented on issue #8632: [WIP] a user friendly way to use g2c in module and an example of g2c URL: https://github.com/apache/incubator-mxnet/pull/8632#issuecomment-345319759 Hi @ZiyueHuang I see this PR is WIP. Does that mean it is not meant for 1.0? It is failing lint

[GitHub] doctaweeks commented on issue #8653: error: identifier "__float128" is undefined

doctaweeks commented on issue #8653: error: identifier "__float128" is undefined URL: https://github.com/apache/incubator-mxnet/issues/8653#issuecomment-345368348 This is definitely a glibc 2.26 issue:

[GitHub] lialie commented on issue #6773: Deadlock and crashes during shutdown

lialie commented on issue #6773: Deadlock and crashes during shutdown URL: https://github.com/apache/incubator-mxnet/pull/6773#issuecomment-345368346 @cjolivier01 situation is much like #8675 . blocking at: ++nwait_consumer_; cv_.wait(lock, [this] {<--

[GitHub] mbaijal commented on issue #8180: Add wide and deep model into sparse example

mbaijal commented on issue #8180: Add wide and deep model into sparse example URL: https://github.com/apache/incubator-mxnet/pull/8180#issuecomment-345374355 @ZiyueHuang This PR is failing in a test. Is this caused by the changes here?

[GitHub] tqchen opened a new pull request #8698: [COREML] Update the json getter

tqchen opened a new pull request #8698: [COREML] Update the json getter URL: https://github.com/apache/incubator-mxnet/pull/8698 ## Description ## (Brief description on what this PR is about) ## Checklist ## ### Essentials ### - [x] Passed code style checking (`make lint`)

[GitHub] tqchen commented on issue #8698: [COREML] Update the json getter

tqchen commented on issue #8698: [COREML] Update the json getter URL: https://github.com/apache/incubator-mxnet/pull/8698#issuecomment-345366272 cc @piiswrong @pracheer This is an automated message from the Apache Git

[GitHub] lialie commented on issue #8675: Sparse tensors causing a deadlock

lialie commented on issue #8675: Sparse tensors causing a deadlock URL: https://github.com/apache/incubator-mxnet/issues/8675#issuecomment-345368918 @marcoabreu Have figured it out? This is an automated message from the

[GitHub] cjolivier01 commented on issue #6773: Deadlock and crashes during shutdown

cjolivier01 commented on issue #6773: Deadlock and crashes during shutdown URL: https://github.com/apache/incubator-mxnet/pull/6773#issuecomment-345378692 What is holding that mutex? From the code, it doesn't look like anything that holds that mutex does anything that could block (mostly

[GitHub] piiswrong commented on issue #8698: [COREML] Update the json getter

piiswrong commented on issue #8698: [COREML] Update the json getter URL: https://github.com/apache/incubator-mxnet/pull/8698#issuecomment-345369763 I think attr was also referenced in the other files? This is an automated

[GitHub] piiswrong commented on issue #8698: [COREML] Update the json getter

piiswrong commented on issue #8698: [COREML] Update the json getter URL: https://github.com/apache/incubator-mxnet/pull/8698#issuecomment-345369903 Also this needs to support both attr and attrs because we need to handle json saved by previous versions

[GitHub] cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast

cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#discussion_r151828803 ## File path: src/operator/tensor/elemwise_binary_broadcast_op.h ## @@ -134,24 +137,122 @@

[GitHub] cjolivier01 commented on issue #8686: Kernel operator tuning including broadcast

cjolivier01 commented on issue #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#issuecomment-345419483 OMP overhead this run, ~8500 ns. unary and binary op times: 1-200 ns OperatorTuneBase::duration_t

[incubator-mxnet] branch master updated: Fix sparse dot test failure due to launching kernel when nnr = 0 and bug of square_sum (#8470)

This is an automated email from the ASF dual-hosted git repository. jxie pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git The following commit(s) were added to refs/heads/master by this push: new 33698ea Fix sparse dot test failure due

[GitHub] piiswrong closed pull request #8470: Fix sparse dot test failure due to launching kernel when nnr = 0 and bug of square_sum

piiswrong closed pull request #8470: Fix sparse dot test failure due to launching kernel when nnr = 0 and bug of square_sum URL: https://github.com/apache/incubator-mxnet/pull/8470 This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed

[GitHub] cjolivier01 commented on issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake?

cjolivier01 commented on issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake? URL: https://github.com/apache/incubator-mxnet/issues/8702#issuecomment-345423131 One potential solution would be a script (python or bash) that fetches, builds and installs CMake. An

[GitHub] marcoabreu commented on issue #8675: Sparse tensors causing a deadlock

marcoabreu commented on issue #8675: Sparse tensors causing a deadlock URL: https://github.com/apache/incubator-mxnet/issues/8675#issuecomment-345425375 @lialie Our current research shows that this is caused by ubuntu 14.04 (in our case inside docker) on the new Skylake Processors with

[GitHub] piiswrong commented on issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake?

piiswrong commented on issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake? URL: https://github.com/apache/incubator-mxnet/issues/8702#issuecomment-345422711 One potential problem is it adds another dependency. Its particularly painful for people using centos or

[GitHub] piiswrong commented on issue #8703: [DISCUSSION] (when) Should we deprecate support for python2?

piiswrong commented on issue #8703: [DISCUSSION] (when) Should we deprecate support for python2? URL: https://github.com/apache/incubator-mxnet/issues/8703#issuecomment-345422656 Let's worry about that when 2020 comes. This

[GitHub] cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast

cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#discussion_r151825923 ## File path: src/operator/tensor/elemwise_binary_broadcast_op.h ## @@ -134,24 +137,122 @@

[GitHub] anirudh2290 commented on issue #8699: Temporary workspace for operators can only be allocated once

anirudh2290 commented on issue #8699: Temporary workspace for operators can only be allocated once URL: https://github.com/apache/incubator-mxnet/issues/8699#issuecomment-345425165 Thanks a lot for your help. I will try to use malloc and free(implementation only for CPU currently) in the

[GitHub] javelinjs commented on issue #8682: Cannot build scala package for GPU

javelinjs commented on issue #8682: Cannot build scala package for GPU URL: https://github.com/apache/incubator-mxnet/issues/8682#issuecomment-345422770 you do not need to pass `USE_BLAS=openblas USE_LAPACK=1 USE_CUDA=1 USE_CUDA_PATH=/opt/cuda USE_CUDNN=1

[GitHub] rahul003 commented on issue #8527: send _send_command_to_servers must be string

rahul003 commented on issue #8527: send _send_command_to_servers must be string URL: https://github.com/apache/incubator-mxnet/pull/8527#issuecomment-345172903 @azai91 Try pushing empty commit to trigger CI again. We are facing some issues with CI currently.

[GitHub] sxjscience commented on issue #8699: Temporary workspace for operators can only be allocated once

sxjscience commented on issue #8699: Temporary workspace for operators can only be allocated once URL: https://github.com/apache/incubator-mxnet/issues/8699#issuecomment-345408198 @anirudh2290 In that case, you may try to allocate an upper bound of the temporary storage.

[GitHub] sxjscience commented on issue #8690: [bug] dropout backward with openmp

sxjscience commented on issue #8690: [bug] dropout backward with openmp URL: https://github.com/apache/incubator-mxnet/issues/8690#issuecomment-345408969 @zhenlinluo @piiswrong It seems to be a bug. This is an automated

[GitHub] sxjscience closed issue #8690: [bug] dropout backward with openmp

sxjscience closed issue #8690: [bug] dropout backward with openmp URL: https://github.com/apache/incubator-mxnet/issues/8690 This is an automated message from the Apache Git Service. To respond to the message, please log on

[GitHub] sxjscience commented on issue #8690: [bug] dropout backward with openmp

sxjscience commented on issue #8690: [bug] dropout backward with openmp URL: https://github.com/apache/incubator-mxnet/issues/8690#issuecomment-345408969 ~~@zhenlinluo @piiswrong It seems to be a bug.~~ I'm wrong. The current version is correct. In fact, the masks are updated in the

[GitHub] szha commented on a change in pull request #8701: add inline for cached op & fixed a bug when calling backward on variable

szha commented on a change in pull request #8701: add inline for cached op & fixed a bug when calling backward on variable URL: https://github.com/apache/incubator-mxnet/pull/8701#discussion_r151825336 ## File path: python/mxnet/_ctypes/ndarray.py ## @@ -105,10 +105,13 @@

[GitHub] piiswrong commented on a change in pull request #8686: Kernel operator tuning including broadcast

piiswrong commented on a change in pull request #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#discussion_r151821026 ## File path: src/operator/tensor/elemwise_binary_broadcast_op.h ## @@ -134,24 +137,122 @@

[GitHub] anirudh2290 commented on issue #8699: Temporary workspace for operators can only be allocated once

anirudh2290 commented on issue #8699: Temporary workspace for operators can only be allocated once URL: https://github.com/apache/incubator-mxnet/issues/8699#issuecomment-345404067 @zhreshold the amount of memory I want to allocate for t2 depends on the results in t1. So I cannot

[GitHub] piiswrong commented on a change in pull request #8686: Kernel operator tuning including broadcast

piiswrong commented on a change in pull request #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#discussion_r151821567 ## File path: src/operator/tensor/elemwise_binary_broadcast_op.h ## @@ -134,24 +137,122 @@

[GitHub] jeremiedb closed issue #8120: [R] Bug introduced in initializer between 0.10.1 and 0.11.1

jeremiedb closed issue #8120: [R] Bug introduced in initializer between 0.10.1 and 0.11.1 URL: https://github.com/apache/incubator-mxnet/issues/8120 This is an automated message from the Apache Git Service. To respond to

[GitHub] cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast

cjolivier01 commented on a change in pull request #8686: Kernel operator tuning including broadcast URL: https://github.com/apache/incubator-mxnet/pull/8686#discussion_r151821987 ## File path: src/operator/tensor/elemwise_binary_broadcast_op.h ## @@ -134,24 +137,122 @@

[GitHub] jeremiedb closed issue #5488: [WIP] RNN with bucketing and mask - R package

jeremiedb closed issue #5488: [WIP] RNN with bucketing and mask - R package URL: https://github.com/apache/incubator-mxnet/issues/5488 This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] piiswrong opened a new pull request #8701: add inline for cached op & fixed a bug when calling backward on variable

piiswrong opened a new pull request #8701: add inline for cached op & fixed a bug when calling backward on variable URL: https://github.com/apache/incubator-mxnet/pull/8701 ## Description ## (Brief description on what this PR is about) ## Checklist ## ### Essentials ### - [

[GitHub] piiswrong opened a new issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake?

piiswrong opened a new issue #8702: [DISCUSSION] Should we deprecate Makefile and only use CMake? URL: https://github.com/apache/incubator-mxnet/issues/8702 1. Maintaining two build pipelines has been a lot of trouble. 2. We are planning to pull in dependencies and build/link them

[GitHub] ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST

ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST URL: https://github.com/apache/incubator-mxnet/issues/8545#issuecomment-345402150 Fantastic, it works great! I was a bit confused about its version 0.12.0, while the official version went to 0.12.1, which still

[GitHub] ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST

ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST URL: https://github.com/apache/incubator-mxnet/issues/8545#issuecomment-345402443 Fantastic, the most recent version with #8121 works great! I was a bit confused about its version 0.12.0, while the official

[GitHub] ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST

ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST URL: https://github.com/apache/incubator-mxnet/issues/8545#issuecomment-345402150 Fantastic, it works great! I was a bit confused about its version 0.12.0, while the official version went to 0.12.1, which still

[GitHub] piiswrong commented on issue #8700: running Gluon's SymbolBlock seems to give a List type instead of Nd.array type

piiswrong commented on issue #8700: running Gluon's SymbolBlock seems to give a List type instead of Nd.array type URL: https://github.com/apache/incubator-mxnet/issues/8700#issuecomment-345402589 Because you specified output to be `outputs =

[GitHub] ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST

ironfrown commented on issue #8545: Incorrect results from R 3.4.2 in MNIST URL: https://github.com/apache/incubator-mxnet/issues/8545#issuecomment-345402443 Fantastic, the most recent version with #8121 works great! I was a bit

[GitHub] piiswrong commented on issue #8687: Autograd y = x and y = 1*x gives different gradient

piiswrong commented on issue #8687: Autograd y = x and y = 1*x gives different gradient URL: https://github.com/apache/incubator-mxnet/issues/8687#issuecomment-345403398 https://github.com/apache/incubator-mxnet/pull/8701 Here is the fix.

[GitHub] cjolivier01 commented on issue #6773: Deadlock and crashes during shutdown

cjolivier01 commented on issue #6773: Deadlock and crashes during shutdown URL: https://github.com/apache/incubator-mxnet/pull/6773#issuecomment-345379960 Why is SignalQueuesForKill() commented out? That could cause a deadlock, actually.

[GitHub] anirudh2290 opened a new issue #8699: Temporary workspace for operators can only be allocated once

anirudh2290 opened a new issue #8699: Temporary workspace for operators can only be allocated once URL: https://github.com/apache/incubator-mxnet/issues/8699 ## Description Temporary workspace in operator can only be allocated once. This becomes a problem if you have some computed

[GitHub] eric-haibin-lin commented on issue #5601: running lazy evaluation code example outputs "Illegal instruction"

eric-haibin-lin commented on issue #5601: running lazy evaluation code example outputs "Illegal instruction" URL: https://github.com/apache/incubator-mxnet/issues/5601#issuecomment-345380910 What version of openblas did you use? I've seen similar error when a buggy version of openblas

[GitHub] eric-haibin-lin commented on issue #5134: Discussion and troubleshooting on PyPI (pip) installation (the newest 0.9.3 release)

eric-haibin-lin commented on issue #5134: Discussion and troubleshooting on PyPI (pip) installation (the newest 0.9.3 release) URL: https://github.com/apache/incubator-mxnet/issues/5134#issuecomment-345380712 Re: Illegal Instruction, I think the reason is that the DLAMI was using the

[GitHub] data4science commented on issue #5717: [QA] why not all examples run out of the box?

data4science commented on issue #5717: [QA] why not all examples run out of the box? URL: https://github.com/apache/incubator-mxnet/issues/5717#issuecomment-345389223 @eric-haibin-lin Sorry I got side tracked from the MKL examples. I?ll re-focus on them.

[GitHub] tz-hmc opened a new issue #8700: running Gluon's SymbolBlock seems to give a List type instead of Nd.array type

tz-hmc opened a new issue #8700: running Gluon's SymbolBlock seems to give a List type instead of Nd.array type URL: https://github.com/apache/incubator-mxnet/issues/8700 Hi, I'm trying to use Resnet as a feature extractor, but for some reason, it returns a list instead of a ndarray.

[GitHub] tqchen commented on issue #8698: [COREML] Update the json getter

tqchen commented on issue #8698: [COREML] Update the json getter URL: https://github.com/apache/incubator-mxnet/pull/8698#issuecomment-345395143 @pracheer updated, This is an automated message from the Apache Git Service. To

[GitHub] eric-haibin-lin commented on issue #8625: Attempting to write a relation network in Gluon

eric-haibin-lin commented on issue #8625: Attempting to write a relation network in Gluon URL: https://github.com/apache/incubator-mxnet/issues/8625#issuecomment-345396332 @reminisce can you help take a look at this issue?

[GitHub] eric-haibin-lin commented on issue #7653: ipython, The kernel appears to have died. It will restart automatically.

eric-haibin-lin commented on issue #7653: ipython, The kernel appears to have died. It will restart automatically. URL: https://github.com/apache/incubator-mxnet/issues/7653#issuecomment-345396552 What version of openblas did you use? I've seen similar issues like this when a buggy

[GitHub] zhreshold commented on issue #8699: Temporary workspace for operators can only be allocated once

zhreshold commented on issue #8699: Temporary workspace for operators can only be allocated once URL: https://github.com/apache/incubator-mxnet/issues/8699#issuecomment-345403868 If you have t1 and t2 temp space to allocate, you can allocate a t1 + t2 space once and have the second

[GitHub] eric-haibin-lin commented on issue #5717: [QA] why not all examples run out of the box?

eric-haibin-lin commented on issue #5717: [QA] why not all examples run out of the box? URL: https://github.com/apache/incubator-mxnet/issues/5717#issuecomment-345382556 @Bumblebee1964 any update for the cpp examples? :)

[GitHub] eric-haibin-lin commented on issue #5717: [QA] why not all examples run out of the box?

eric-haibin-lin commented on issue #5717: [QA] why not all examples run out of the box? URL: https://github.com/apache/incubator-mxnet/issues/5717#issuecomment-345382556 @Bumblebee1964 any update for the cpp examples? :) Someone is asking about batch mode in cpp

[GitHub] eric-haibin-lin commented on issue #8459: builde error

eric-haibin-lin commented on issue #8459: builde error URL: https://github.com/apache/incubator-mxnet/issues/8459#issuecomment-345393864 @szha cuda 6.5 should be supported, right? @994374821 did you try an older version of MXNet?

  1   2   >