[GitHub] [incubator-mxnet] ys2843 commented on pull request #18288: Website global search feature

[GitBox]

ys2843 commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629589289


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18288: Website global search feature

[GitBox]

mxnet-bot commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629589299


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] connorgoggins commented on pull request #18340: [Website 2.0] Artifact URL Adjustment

[GitBox]

connorgoggins commented on pull request #18340:
URL: https://github.com/apache/incubator-mxnet/pull/18340#issuecomment-629583487


   @mxnet-label-bot add [Website]



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

[GitHub] [incubator-mxnet] connorgoggins commented on pull request #18340: [Website 2.0] Artifact URL Adjustment

[GitBox]

connorgoggins commented on pull request #18340:
URL: https://github.com/apache/incubator-mxnet/pull/18340#issuecomment-629583448


   @mxnet-label-bot add [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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18340: [Website 2.0] Artifact URL Adjustment

[GitBox]

mxnet-bot commented on pull request #18340:
URL: https://github.com/apache/incubator-mxnet/pull/18340#issuecomment-629583305


   Hey @connorgoggins , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [windows-cpu, unix-gpu, sanity, centos-cpu, 
centos-gpu, unix-cpu, website, windows-gpu, clang, miscellaneous, edge]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] connorgoggins opened a new pull request #18340: [Website 2.0] Artifact URL Adjustment

[GitBox]

connorgoggins opened a new pull request #18340:
URL: https://github.com/apache/incubator-mxnet/pull/18340


   ## Description ##
   This PR officially transfers ownership of code version control and 
compressed artifact storage for the static artifacts for the MXNet website from 
Connor Goggins to the MXNet Website team. The new S3 bucket (owned by the AWS 
account for ai-mxnet-engineer...@amazon.com) gives the team the ability to 
update the artifacts as needed. The specific purpose of this PR is to change 
the artifact download link in the Jenkinsfile to point to the compressed 
artifacts in the new S3 bucket instead of the previous bucket.
   
   These changes will increase the accessibility of the MXNet Website 2.0 
static artifacts to the team in case of future modifications. For additional 
details and supporting documentation, please read [this 
doc](https://quip-amazon.com/4i2GAb8huyKC/MXNet-Website-20-Static-Artifact-Mangement).
   
   ## Checklist ##
   ### Essentials ###
   - [x] Changes are complete (i.e. I finished coding on this PR)
   - [x] Code is well-documented
   - [x] To the best of my knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   
   ### Changes ###
   - M docs/static_site/Makefile
   
   ## Comments ##
   These changes have been tested with a complete website build on Jenkins dev 
and local hosting. All functionality remains the same.
   
   @sandeep-krishnamurthy @aaronmarkham @ys2843 



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

[GitHub] [incubator-mxnet] leezu commented on pull request #18315: Fix coredumps

[GitBox]

leezu commented on pull request #18315:
URL: https://github.com/apache/incubator-mxnet/pull/18315#issuecomment-629573080


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18315: Fix coredumps

[GitBox]

mxnet-bot commented on pull request #18315:
URL: https://github.com/apache/incubator-mxnet/pull/18315#issuecomment-629573096


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] szha commented on issue #18244: unix-cpu MKL/MKL-DNN Test Time

[GitBox]

szha commented on issue #18244:
URL: 
https://github.com/apache/incubator-mxnet/issues/18244#issuecomment-629566327


   Any update on 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

[GitHub] [incubator-mxnet] rondogency commented on pull request #18331: Remove test metric perf

[GitBox]

rondogency commented on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629566211


   @leezu the performance benchmark we are running is focusing on model level 
performance on end to end training/inference, this one is more related to 
opperf to test components



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18326: [R] Fix incorrect copyto usage & incorrect website title for Symbol API in R

[GitBox]

ChaiBapchya commented on pull request #18326:
URL: https://github.com/apache/incubator-mxnet/pull/18326#issuecomment-629564086


   @leezu @access2rohit plz review/merge



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

[GitHub] [incubator-mxnet] waytrue17 commented on pull request #18339: update dockerfile for jetson

[GitBox]

waytrue17 commented on pull request #18339:
URL: https://github.com/apache/incubator-mxnet/pull/18339#issuecomment-629563544


   @ciyongch 



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

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

[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 4bdada6  Bump the publish timestamp.
4bdada6 is described below

commit 4bdada6e752b8d2aaeb528748b3ebaf74035dc61
Author: mxnet-ci 
AuthorDate: Sat May 16 00:48:10 2020 +

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..b98f4b3
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Sat May 16 00:48:10 UTC 2020

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18339: update dockerfile for jetson

[GitBox]

mxnet-bot commented on pull request #18339:
URL: https://github.com/apache/incubator-mxnet/pull/18339#issuecomment-629562168


   Hey @waytrue17 , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [centos-gpu, miscellaneous, edge, unix-gpu, sanity, 
website, unix-cpu, centos-cpu, windows-gpu, clang, windows-cpu]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] waytrue17 opened a new pull request #18339: update dockerfile for jetson

[GitBox]

waytrue17 opened a new pull request #18339:
URL: https://github.com/apache/incubator-mxnet/pull/18339


   ## Description ##
   Trying to fix #18311 on v1.7.x by copying Dockerfile.build.jetson from 
master to v1.7.x
   
   ### Essentials ###
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   
   
   



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18288: Website global search feature

[GitBox]

mxnet-bot commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629557848


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18288: Website global search feature

[GitBox]

ys2843 commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629557829


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18338: [DO NOT MERGE] Use mkl cmake flag to force DNNL to delegate FC op to MKL

[GitBox]

ChaiBapchya commented on pull request #18338:
URL: https://github.com/apache/incubator-mxnet/pull/18338#issuecomment-629557360


   Related issue : https://github.com/apache/incubator-mxnet/issues/17980



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18338: [DO NOT MERGE] Use mkl cmake flag to force DNNL to delegate FC op to MKL

[GitBox]

mxnet-bot commented on pull request #18338:
URL: https://github.com/apache/incubator-mxnet/pull/18338#issuecomment-629557180


   Hey @ChaiBapchya , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [sanity, centos-gpu, windows-gpu, unix-cpu, unix-gpu, 
windows-cpu, centos-cpu, website, edge, miscellaneous, clang]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] ChaiBapchya opened a new pull request #18338: [DO NOT MERGE] Use mkl cmake flag to force DNNL to delegate FC op to MKL

[GitBox]

ChaiBapchya opened a new pull request #18338:
URL: https://github.com/apache/incubator-mxnet/pull/18338


   ## Description ##
   Currently, for GEMM ops [DNNL delegates that op to MKL]
   However, for FC DNNL doesn't.
   Workaround : export USE_MKL & pass location of MKL installation.
   This forces DNNL to delegate ops to MKL
   This PR tries to verify correctness of the workaround [if it breaks any 
unit-tests]
   
   @leezu @PatricZhao @TaoLv @kpuatamazon @kpu
   
   This PR is not meant to be merged. But to serve as a proxy for internal 
releases.
   If it passes the CI without issues, PR will be closed.
   Pipelines to be watched for : windows-cpu, unix-cpu & centos-cpu
   Specifically mkldnn/mkl builds.
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   
   



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

[GitHub] [incubator-mxnet] szha commented on pull request #18335: Skip test_metric_performance on MKL builds

[GitBox]

szha commented on pull request #18335:
URL: https://github.com/apache/incubator-mxnet/pull/18335#issuecomment-629552229


   I'm moving this to benchmark scripts in #18252



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

[GitHub] [incubator-mxnet] szha commented on pull request #18331: Remove test metric perf

[GitBox]

szha commented on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629552076


   I'm moving it to benchmark scripts in #18252 



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

[GitHub] [incubator-mxnet] szha commented on pull request #18252: [CI] run operator tests with naive engine

[GitBox]

szha commented on pull request #18252:
URL: https://github.com/apache/incubator-mxnet/pull/18252#issuecomment-629545644


   @marcoabreu all but the operator unit tests still run on threaded engine.



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

[GitHub] [incubator-mxnet] aaraujom commented on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

aaraujom commented on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629542494


   Hi thanks for checking. Let me try to reproduce this in my end and fix 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

[incubator-mxnet] branch master updated (3e676fc -> 37280e4)

[lausen]

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

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


from 3e676fc  Fix memory leaks in Gluon (#18328)
 add 37280e4  Fix deferred compute mode for operators using new FFI (#18284)

No new revisions were added by this update.

Summary of changes:
 src/api/operator/utils.cc  |  6 +-
 src/api/operator/utils.h   |  3 ++-
 tests/python/unittest/test_deferred_compute.py | 12 
 3 files changed, 19 insertions(+), 2 deletions(-)

[incubator-mxnet] branch master updated: Fix deferred compute mode for operators using new FFI (#18284)

[lausen]

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

lausen 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 37280e4  Fix deferred compute mode for operators using new FFI (#18284)
37280e4 is described below

commit 37280e4ddf00cacdac50c1e798fd2a14da38ae8d
Author: Leonard Lausen 
AuthorDate: Fri May 15 15:36:55 2020 -0700

Fix deferred compute mode for operators using new FFI (#18284)
---
 src/api/operator/utils.cc  |  6 +-
 src/api/operator/utils.h   |  3 ++-
 tests/python/unittest/test_deferred_compute.py | 12 
 3 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/src/api/operator/utils.cc b/src/api/operator/utils.cc
index 307bb29..6cfbd27 100644
--- a/src/api/operator/utils.cc
+++ b/src/api/operator/utils.cc
@@ -30,6 +30,10 @@ bool is_recording() {
   return Imperative::Get()->is_recording();
 }
 
+bool is_deferred_compute() {
+  return Imperative::Get()->is_deferred_compute();
+}
+
 void SetInOut(std::vector* ndinputs,
   std::vector* ndoutputs,
   int num_inputs,
@@ -94,7 +98,7 @@ std::vector Invoke(const nnvm::Op* op,
   Imperative::DCInfo::Compute(*input);
 }
 auto state = Imperative::Get()->Invoke(Context::CPU(), *attrs, ndinputs, 
ndoutputs);
-if (Imperative::Get()->is_recording()) {
+if (is_recording()) {
   Imperative::Get()->RecordOp(std::move(*attrs), ndinputs, ndoutputs, 
state);
 }
   }
diff --git a/src/api/operator/utils.h b/src/api/operator/utils.h
index 8943e80..014ff15 100644
--- a/src/api/operator/utils.h
+++ b/src/api/operator/utils.h
@@ -48,10 +48,11 @@ std::vector Invoke(const nnvm::Op* op,
  NDArray** outputs);
 
 bool is_recording();
+bool is_deferred_compute();
 
 template
 void SetAttrDict(nnvm::NodeAttrs* attrs) {
-  if (is_recording()) {
+  if (is_recording() || is_deferred_compute()) {
 ::dmlc::get(attrs->parsed).SetAttrDict(&(attrs->dict));
   }
 }
diff --git a/tests/python/unittest/test_deferred_compute.py 
b/tests/python/unittest/test_deferred_compute.py
index e68373f..d93e67f 100644
--- a/tests/python/unittest/test_deferred_compute.py
+++ b/tests/python/unittest/test_deferred_compute.py
@@ -162,6 +162,18 @@ def test_dc_no_inputs_subset_of_output():
 _all_assert_dc(_dc_empty_setup, f)
 
 
+def test_dc_numpy_tril():
+def f(a, *, nd):
+assert nd is mx.np
+a = nd.ones((2, 2))
+b = nd.tril(a, 1)
+c = nd.tril(a, -1)
+return [b, c]
+
+for mode in ('all', 'symbolic', 'imperative', 
'imperativewithnondccompute'):
+_assert_dc(_dc_simple_setup, f, mode=mode)
+
+
 ###
 # Test cases with inputs
 ###

[GitHub] [incubator-mxnet] leezu merged pull request #18284: Fix parse operator attributes in new FFI

[GitBox]

leezu merged pull request #18284:
URL: https://github.com/apache/incubator-mxnet/pull/18284


   



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

[GitHub] [incubator-mxnet] leezu closed issue #18004: Wrong result when using new numpy ffi in deferred compute

[GitBox]

leezu closed issue #18004:
URL: https://github.com/apache/incubator-mxnet/issues/18004


   



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18315: Fix coredumps

[GitBox]

mxnet-bot commented on pull request #18315:
URL: https://github.com/apache/incubator-mxnet/pull/18315#issuecomment-629534560


   Jenkins CI successfully triggered : [centos-cpu, unix-cpu]



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

[GitHub] [incubator-mxnet] leezu commented on pull request #18315: Fix coredumps

[GitBox]

leezu commented on pull request #18315:
URL: https://github.com/apache/incubator-mxnet/pull/18315#issuecomment-629534528


   @mxnet-bot run ci [unix-cpu, centos-cpu]



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18337: fix native cd builds

[GitBox]

mxnet-bot commented on pull request #18337:
URL: https://github.com/apache/incubator-mxnet/pull/18337#issuecomment-629530814


   Hey @mseth10 , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [windows-cpu, centos-cpu, edge, clang, unix-gpu, 
centos-gpu, miscellaneous, windows-gpu, unix-cpu, website, sanity]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] mseth10 opened a new pull request #18337: fix native cd builds

[GitBox]

mseth10 opened a new pull request #18337:
URL: https://github.com/apache/incubator-mxnet/pull/18337


   ## Description ##
   Fix CD builds for 'native' flavor



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya commented on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629526785


   Can confirm that this issue is specific to AVX512 kernels.
   Tried this on c5.xl 
   $ lscpu
   ```
   Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz
   Flags:   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge 
mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm 
constant_tsc arch_perfmon rep_good nopl xtopology nonstop_tsc cpuid aperfmperf 
tsc_known_freq pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic 
movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm 
abm 3dnowprefetch invpcid_single pti fsgsbase tsc_adjust bmi1 hle avx2 smep 
bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap clflushopt clwb 
avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves ida arat pku ospke
   ```
   ## Results
   ### Default [slower]
   ```
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.133789
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.132812
   ```
   ```
   [{'FullyConnected': [
   {'inputs': {'data': (4, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.10202302001744101, 
'p50_time_FullyConnected': 0.10086749989568489, 'p90_time_FullyConnected': 
0.10658760029400582, 'p99_time_FullyConnected': 0.13521948004836298}, 
   {'inputs': {'data': (5, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.10642346004715364, 
'p50_time_FullyConnected': 0.09991750016524747, 'p90_time_FullyConnected': 
0.10565369971118344, 'p99_time_FullyConnected': 0.2586996700802042}, 
   {'inputs': {'data': (5, 512), 'weight': (1536, 512), 'no_bias': True, 
'num_hidden': 1536}, 'avg_time_FullyConnected': 0.16890607999812346, 
'p50_time_FullyConnected': 0.16431500012004108, 'p90_time_FullyConnected': 
0.1781331999154645, 'p99_time_FullyConnected': 0.2831235897247094}, 
   {'inputs': {'data': (5, 512), 'weight': (2048, 512), 'no_bias': True, 
'num_hidden': 2048}, 'avg_time_FullyConnected': 0.20140223995440465, 
'p50_time_FullyConnected': 0.19778950013460417, 'p90_time_FullyConnected': 
0.20401089991537447, 'p99_time_FullyConnected': 0.3063294199228036}, 
   {'inputs': {'data': (5, 2048), 'weight': (512, 2048), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.21596427998701984, 
'p50_time_FullyConnected': 0.209670038254, 'p90_time_FullyConnected': 
0.21819640001012885, 'p99_time_FullyConnected': 0.3412436299549877}]}]
   ```
   ### MKL Workaround [Faster]
   ```
   MKL_VERBOSE 
SGEMM(T,N,512,5,2048,0x7f9bcf6fac28,0x7f9bc22f4040,2048,0x7f9b1400ce80,2048,0x7f9bcf6fac30,0x7f9b1405e840,512)
 21.25us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:18
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.0378418
   MKL_VERBOSE 
SGEMM(T,N,512,5,2048,0x7f9bcf6fac28,0x7f9bc22f4040,2048,0x7f9b1400ce80,2048,0x7f9bcf6fac30,0x7f9b14061c00,512)
 20.94us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:18
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.0371094
   ```
   ```
   [{'FullyConnected': [
   {'inputs': {'data': (4, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.11772135999308375, 
'p50_time_FullyConnected': 0.1149684999290912, 'p90_time_FullyConnected': 
0.1244978000613628, 'p99_time_FullyConnected': 0.14825501980340045}, 
   {'inputs': {'data': (5, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.120828840035756, 
'p50_time_FullyConnected': 0.11370450010872446, 'p90_time_FullyConnected': 
0.12752780021401122, 'p99_time_FullyConnected': 0.2412066401620902}, 
   {'inputs': {'data': (5, 512), 'weight': (1536, 512), 'no_bias': True, 
'num_hidden': 1536}, 'avg_time_FullyConnected': 0.13385597998421872, 
'p50_time_FullyConnected': 0.12600750005731243, 'p90_time_FullyConnected': 
0.14806160011175962, 'p99_time_FullyConnected': 0.2509373301927551}, 
   {'inputs': {'data': (5, 512), 'weight': (2048, 512), 'no_bias': True, 
'num_hidden': 2048}, 'avg_time_FullyConnected': 0.14175208003507578, 
'p50_time_FullyConnected': 0.1372545000322134, 'p90_time_FullyConnected': 
0.14401020002878798, 'p99_time_FullyConnected': 0.2423993399725075}, 
   {'inputs': {'data': (5, 2048), 'weight': (512, 2048), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.143890859962994, 
'p50_time_FullyConnected': 0.139797637755, 'p90_time_FullyConnected': 

[GitHub] [incubator-mxnet] bill10 opened a new issue #18336: topk operation is very memory intensive

[GitBox]

bill10 opened a new issue #18336:
URL: https://github.com/apache/incubator-mxnet/issues/18336


   ## Description
   I am not sure if this is a bug, but the topk operation has a surprisingly 
high memory cost compared to, e.g., pytorch.
   
   ### Error Message
   This resulted in out-of-memory error for large tensors. 
   
   ## To Reproduce
   For mxnet 1.6.0, python 3.6
   ```
   %memit X = mx.nd.random.uniform(shape=(32, 200))
   #peak memory: 419.61 MiB, increment: 249.80 MiB
   ```
   ```
   %memit y=X.topk(k=100)
   #peak memory: 1396.98 MiB, increment: 977.97 MiB
   ```
   For pytorch 1.4:
   ```
   %memit X=th.rand(size=(32, 200))
   # peak memory: 417.15 MiB, increment: 244.70 MiB
   ```
   ```
   %memit y=X.topk(k=100)
   # peak memory: 906.36 MiB, increment: 489.05 MiB
   ```
   
   If I read correctly, mxnet uses twice as much the memory as pytorch uses 
(977 vs 489). 
   
   ## Environment
   mxnet 1.6.0, python 3.6
   



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

[GitHub] [incubator-mxnet] zhreshold commented on pull request #18325: Change SGD with momentum to include momentum correction by default

[GitBox]

zhreshold commented on pull request #18325:
URL: https://github.com/apache/incubator-mxnet/pull/18325#issuecomment-629504587


   @szhengac Can you provide feedback on 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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18335: Skip test_metric_performance on MKL builds

[GitBox]

mxnet-bot commented on pull request #18335:
URL: https://github.com/apache/incubator-mxnet/pull/18335#issuecomment-629504113


   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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [website, windows-gpu, miscellaneous, windows-cpu, 
clang, centos-cpu, centos-gpu, sanity, edge, unix-cpu, unix-gpu]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] leezu opened a new pull request #18335: Skip test_metric_performance on MKL builds

[GitBox]

leezu opened a new pull request #18335:
URL: https://github.com/apache/incubator-mxnet/pull/18335


   https://github.com/apache/incubator-mxnet/issues/18244
   
   metrics now use mxnet operators
   
   Fixes https://github.com/apache/incubator-mxnet/issues/18330 



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

[GitHub] [incubator-mxnet] leezu commented on pull request #18331: Remove test metric perf

[GitBox]

leezu commented on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629477586


   I think we should integrate it in the performance benchmark. Currently there 
is not mechanism to monitor the results



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya commented on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629477511


   Yes logs agree with these statements. But the perf difference isn't visible 
via opperf.
   My bad, lhs, rhs wrongly interpreted. But that's for dot,batch_dot. FC 
values for data, weight are same as yours.
   
   CPU : Intel(R) Xeon(R) CPU E5-2686 v4 @ 2.30GHz
   Instance p3.8xl
   
   This instance doesn't have avx512
   ```
   $ lscpu | grep Flags | grep avx
   Flags:   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge 
mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx pdpe1gb rdtscp lm 
constant_tsc rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq 
ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes 
xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch cpuid_fault 
invpcid_single pti fsgsbase bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx 
xsaveopt
   ``` 
   
   OMP_Threading i used was default. 
   
   However trying FC with OMP_NUM_THREADS=4, still shows that workaround makes 
FC slower [despite it using MKL] than default[which doesn't use MKL]
   
   ### Workaround [slow]
   ```
   MKL_VERBOSE 
SGEMM(T,N,512,5,2048,0x7f0b79ff8c28,0x7f0b62cfb040,2048,0x7f0b6800b740,2048,0x7f0b79ff8c30,0x7f0b68074e80,512)
 119.93us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:4
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.142822
   ```
   ```
   {'inputs': {'data': (4, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.0925}, 
   {'inputs': {'data': (5, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.0921}, 
   {'inputs': {'data': (5, 512), 'weight': (1536, 512), 'no_bias': True, 
'num_hidden': 1536}, 'avg_time_forward_FullyConnected': 0.1441}, 
   {'inputs': {'data': (5, 512), 'weight': (2048, 512), 'no_bias': True, 
'num_hidden': 2048}, 'avg_time_forward_FullyConnected': 0.1688}, 
   {'inputs': {'data': (5, 2048), 'weight': (512, 2048), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.1773}]}]
   ```
   ### Default [faster]
   ```
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.11499
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.115967
   ```
   ```
   {'inputs': {'data': (4, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.0691}, 
   {'inputs': {'data': (5, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.0669}, 
   {'inputs': {'data': (5, 512), 'weight': (1536, 512), 'no_bias': True, 
'num_hidden': 1536}, 'avg_time_forward_FullyConnected': 0.1166}, 
   {'inputs': {'data': (5, 512), 'weight': (2048, 512), 'no_bias': True, 
'num_hidden': 2048},  'avg_time_forward_FullyConnected': 0.1438}, 
   {'inputs': {'data': (5, 2048), 'weight': (512, 2048), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_forward_FullyConnected': 0.1509}]}]
   ```
   
   ## Python's time-it function in OpPerf
   Also, instead of using MXNet's native [built-in] profiler, if i use python 
time-it function still
   ### Workaround [slow]
   ```
   MKL_VERBOSE 
SGEMM(T,N,512,5,2048,0x7f86067f9c28,0x7f85e34fd040,2048,0x7f85f000b740,2048,0x7f86067f9c30,0x7f85f005fc80,512)
 120.56us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:4
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.143066
   ```
   ```
   [{'FullyConnected': [
   {'inputs': {'data': (4, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.18504947423934937, 
'p50_time_FullyConnected': 0.1817569718696177, 'p90_time_FullyConnected': 
0.19327133195474744, 'p99_time_FullyConnected': 0.21545797004364425}, 
   {'inputs': {'data': (5, 512), 'weight': (512, 512), 'no_bias': True, 
'num_hidden': 512}, 'avg_time_FullyConnected': 0.1922117848880589, 
'p50_time_FullyConnected': 0.18283800454810262, 'p90_time_FullyConnected': 
0.19954713061451912, 'p99_time_FullyConnected': 0.36291924072429527}, 
   {'inputs': {'data': (5, 512), 'weight': (1536, 512), 'no_bias': True, 
'num_hidden': 1536}, 'avg_time_FullyConnected': 0.24743830785155296, 
'p50_time_FullyConnected': 0.23927190341055393, 'p90_time_FullyConnected': 
0.25965459644794464, 'p99_time_FullyConnected': 0.3613424429204313}, 
   {'inputs': {'data': (5, 512), 'weight': (2048, 512), 'no_bias': True, 
'num_hidden': 2048}, 'avg_time_FullyConnected': 0.2728361077606678, 

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18288: Website global search feature

[GitBox]

ys2843 edited a comment on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-628990327


   > Looks good.
   > 
   > * Should we use offsite js files or host locally?
   > * Should the API key be in a js file are can it be in a config file so we 
keep this stuff in someplace central?
   > * Wonder if we could have the list of versions also in a config file and 
not in .js, so it is more obvious and easier to maintain... then other parts of 
the site like the install selector or other things that need the list of 
versions can share?
   > 
   > Just some thoughts... otherwise this looks good to me.
   > Good work!
   
   Should we use offsite js files or host locally?
   + Based on the research, the CDN works pretty well in China. I will keep an 
eye on it, if there is problem loading this file in any country, I will switch 
to host the file locally.
   
   Should the API key be in a js file are can it be in a config file so we keep 
this stuff in someplace central?
   + Didn't move API key to Jekyll at last, because Jekyll does not work very 
well with files such as `assets/*.js`, the way to get variable from Jekyll to 
JS is not straight forward ( use `_include`), please correct me if I am wrong. 
And after this version is archived, it is also easier to maintain if the key is 
kept in 1 JS file, but not spread to every page by Jekyll include 

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18288: Website global search feature

[GitBox]

ys2843 edited a comment on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-628990327


   > Looks good.
   > 
   > * Should we use offsite js files or host locally?
   > * Should the API key be in a js file are can it be in a config file so we 
keep this stuff in someplace central?
   > * Wonder if we could have the list of versions also in a config file and 
not in .js, so it is more obvious and easier to maintain... then other parts of 
the site like the install selector or other things that need the list of 
versions can share?
   > 
   > Just some thoughts... otherwise this looks good to me.
   > Good work!
   
   Should we use offsite js files or host locally?
   + Based on the research, the CDN works pretty well in China. I will keep an 
eye on it, if there is problem loading this file in any country, I will switch 
to host the file locally.
   
   Should the API key be in a js file are can it be in a config file so we keep 
this stuff in someplace central?
   + Didn't move API key to Jekyll at last, because Jekyll does not work with 
files such as `assets/*.js` very well, the way to get variable from Jekyll to 
JS is not straight forward ( use `_include`), please correct me if I am wrong. 
And after this version is archived, it is also easier to maintain if the key is 
kept in 1 JS file, but not spread to every page by Jekyll include 

[GitHub] [incubator-mxnet-ci] ChaiBapchya commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

ChaiBapchya commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r426024021



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   Addressed.
   
https://github.com/apache/incubator-mxnet-ci/pull/25/commits/2da87ac0424dd68b799cd24ded81e27ebc3a0969





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

[GitHub] [incubator-mxnet-ci] mseth10 commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

mseth10 commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r425988017



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   As per offline discussion, we'll trigger the lambda at 2 am UTC and 
check for CD pipelines triggered in the last 8 hours. That will cover all 
builds for the day.





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

[GitHub] [incubator-mxnet-ci] mseth10 commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

mseth10 commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r425988017



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   As per offline discussion, we'll trigger the lambda at 2 am UTC and 
check for CD pipelines triggered in the last 8 hours.





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

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on issue #18330: test_metric_performance

[GitBox]

ChaiBapchya edited a comment on issue #18330:
URL: 
https://github.com/apache/incubator-mxnet/issues/18330#issuecomment-629044532


   
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Funix-cpu/detail/PR-18326/1/pipeline
   
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/rest/organizations/jenkins/pipelines/mxnet-validation/pipelines/unix-cpu/branches/PR-18327/runs/1/nodes/365/log/?start=0
   
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Funix-cpu/detail/PR-18324/3/pipeline
   #18324 #18326 #18327 unrelated PRs



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18324: [OpPerf] Add example of using opperf with internal op locally

[GitBox]

ChaiBapchya commented on pull request #18324:
URL: https://github.com/apache/incubator-mxnet/pull/18324#issuecomment-629445905


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18324: [OpPerf] Add example of using opperf with internal op locally

[GitBox]

mxnet-bot commented on pull request #18324:
URL: https://github.com/apache/incubator-mxnet/pull/18324#issuecomment-629445934


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on pull request #18331: Remove test metric perf

[GitBox]

ChaiBapchya edited a comment on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629444500


   Quick history check reveals this file was added by @safrooze in #9705 
   Can we move this to nightly since that was planned? or do we get rid of it 
altogether? 



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18331: Remove test metric perf

[GitBox]

ChaiBapchya commented on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629444500


   Quick history check reveals this file was added by @safrooze in #9705 



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

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on pull request #18331: Remove test metric perf

[GitBox]

ChaiBapchya edited a comment on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629444500


   Quick history check reveals this file was added by @safrooze in #9705 
   Can we move this to nightly since that was planned? or do we get rid of it 
altogether? The test errors out 



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

[GitHub] [incubator-mxnet] ys2843 edited a comment on pull request #18288: Website global search feature

[GitBox]

ys2843 edited a comment on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629441115


   > Very nice. Thank you @ys2843 This will be very helpful for the community.
   > Please update the release manager checklist for changing the defaults.
   > 
   > Should we change anything with algolia (like default version) whenever 
there is a new release?
   > Also, for updating the search results, what is the process? What do you 
mean by improving the search result.
   > 
   > Overall, this is really nice improvement for easily finding the content. 
Thank you.
   
   Thank you for reviewing. Will update the release manager checklist.
   > Should we change anything with algolia (like default version) whenever 
there is a new release?
   
   No, we don't need any change from algolia. We only need to update website 
code to set the default version. I will document this part.
   
   > Also, for updating the search results, what is the process? What do you 
mean by improving the search result.
   
   To make the search result more relevant and accurate, the process is to 
update the [configuration of DocSearch's web 
crawler](https://github.com/algolia/docsearch-configs/blob/master/configs/apache_mxnet.json).
 
   Currently the config is not ideal, and the best way to update and test the 
config is to run the web crawler locally. I have set up the crawler on an ec2 
and run it with new config, the data is sent to my test account for review, 
then I can make adjustment to the config accordingly.



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18326: [R] Fix incorrect copyto usage & incorrect website title for Symbol API in R

[GitBox]

ChaiBapchya commented on pull request #18326:
URL: https://github.com/apache/incubator-mxnet/pull/18326#issuecomment-629442039


   @mxnet-label-bot add [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

[GitHub] [incubator-mxnet] ChaiBapchya commented on pull request #18130: Set cache_intermediate to True

[GitBox]

ChaiBapchya commented on pull request #18130:
URL: https://github.com/apache/incubator-mxnet/pull/18130#issuecomment-629442396


   Closing since this one is useful for local development & not on 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

[GitHub] [incubator-mxnet] ChaiBapchya closed pull request #18130: Set cache_intermediate to True

[GitBox]

ChaiBapchya closed pull request #18130:
URL: https://github.com/apache/incubator-mxnet/pull/18130


   



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

[GitHub] [incubator-mxnet] ys2843 commented on pull request #18288: Website global search feature

[GitBox]

ys2843 commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629441115


   > Very nice. Thank you @ys2843 This will be very helpful for the community.
   > Please update the release manager checklist for changing the defaults.
   > 
   > Should we change anything with algolia (like default version) whenever 
there is a new release?
   > Also, for updating the search results, what is the process? What do you 
mean by improving the search result.
   > 
   > Overall, this is really nice improvement for easily finding the content. 
Thank you.
   
   Thank you for reviewing. Will update the release manager checklist.
   > Should we change anything with algolia (like default version) whenever 
there is a new release?
   
   No, we don't need any change from algolia. We only need to update website 
code to set the default version. I will document this part.
   
   > Also, for updating the search results, what is the process? What do you 
mean by improving the search result.
   
   To make the search result more relevant and accurate, the process is to 
update the [configuration of DocSearch's web 
crawler](https://github.com/algolia/docsearch-configs/blob/master/configs/apache_mxnet.json).
 
   Currently the config is not ideal, and the best way to update and test the 
config is to run the web crawler locally. I have set up the crawler on an ec2 
and run it with new config, and then make adjustment according to the result 
from the crawler.



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18284: Fix parse operator attributes in new FFI

[GitBox]

mxnet-bot commented on pull request #18284:
URL: https://github.com/apache/incubator-mxnet/pull/18284#issuecomment-629428729


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] leezu commented on pull request #18284: Fix parse operator attributes in new FFI

[GitBox]

leezu commented on pull request #18284:
URL: https://github.com/apache/incubator-mxnet/pull/18284#issuecomment-629428687


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] sandeep-krishnamurthy commented on pull request #18288: Website global search feature

[GitBox]

sandeep-krishnamurthy commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629426892


   @szha / @leezu - FYI.



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

[GitHub] [incubator-mxnet] sandeep-krishnamurthy commented on pull request #18288: Website global search feature

[GitBox]

sandeep-krishnamurthy commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629425112


   @mxnet-bot run ci [unix-cpu]



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18288: Website global search feature

[GitBox]

mxnet-bot commented on pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#issuecomment-629425161


   Jenkins CI successfully triggered : [unix-cpu]



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

[GitHub] [incubator-mxnet] sandeep-krishnamurthy commented on a change in pull request #18288: Website global search feature

[GitBox]

sandeep-krishnamurthy commented on a change in pull request #18288:
URL: https://github.com/apache/incubator-mxnet/pull/18288#discussion_r425990851



##
File path: docs/static_site/src/assets/js/globalSearch.js
##
@@ -0,0 +1,152 @@
+/*!
+ * 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.
+ */
+
+/* Installation page display functions for install selector.
+   This utility allows direct links to specific install instructions.
+*/
+
+$(document).ready(function () {
+  const DEFAULT_CURRENT_VERSION = "1.6";
+  const VERSIONS = [
+"master",

Review comment:
   +1





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

[GitHub] [incubator-mxnet-ci] mseth10 commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

mseth10 commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r425988017



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   As per offline discussion, checking for last 6 hours would suffice.





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

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

[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 e3cceb2  Bump the publish timestamp.
e3cceb2 is described below

commit e3cceb2ba28e168473b10ae3427fa87a50e47ddb
Author: mxnet-ci 
AuthorDate: Fri May 15 18:48:13 2020 +

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..dc047eb
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Fri May 15 18:48:13 UTC 2020

[GitHub] [incubator-mxnet-ci] ChaiBapchya commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

ChaiBapchya commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r425978610



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   So we are May 7 11:59pm UTC [lambda is triggered]
   
   Latest build as of May 7 11:59pm UTC is _Build 1088_ [May 7 9:54:39pm]
   
   
http://jenkins.mxnet-ci.amazon-ml.com/job/restricted-mxnet-cd/job/mxnet-cd-release-job/1088/
   
    If we relax the same day check & only keep the 24hour check.
   So all builds after May 6 9:55:39 pm will be accepted.
   1082 to 1088
   
   
   Here 1082 isn't supposed to be selected as it was part of the previous day 
trigger
   _Build 1082_ May 6, 2020, 9:57:08 PM [which is not supposed to be picked]
   
http://jenkins.mxnet-ci.amazon-ml.com/job/restricted-mxnet-cd/job/mxnet-cd-release-job/1082/
   
    Hence we need both checks





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

[GitHub] [incubator-mxnet-ci] ChaiBapchya commented on a change in pull request #25: Serverless implementation for jenkins pipeline monitor lambda

[GitBox]

ChaiBapchya commented on a change in pull request #25:
URL: https://github.com/apache/incubator-mxnet-ci/pull/25#discussion_r425978610



##
File path: services/jenkins-pipeline-monitor/handler.py
##
@@ -0,0 +1,134 @@
+import os
+import boto3
+import json
+import logging
+import secret_manager
+
+from jenkinsapi.jenkins import Jenkins
+
+logging.getLogger().setLevel(logging.INFO)
+logging.getLogger('boto3').setLevel(logging.CRITICAL)
+logging.getLogger('botocore').setLevel(logging.CRITICAL)
+
+
+def get_jenkins_obj(secret):
+"""
+This method returns an object of Jenkins instantiated using username, 
password
+"""
+jenkins_url, jenkins_username, jenkins_password = 
os.environ["JENKINS_URL"], secret["jenkins_username"], 
secret["jenkins_password"]
+return Jenkins(jenkins_url, username=jenkins_username, 
password=jenkins_password)
+
+
+def get_secret():
+"""
+This method is to get secret value from Secrets Manager
+"""
+secret = json.loads(secret_manager.get_secret())
+return secret
+
+
+def get_pipeline_job(jenkinsObj):
+job = jenkinsObj["restricted-mxnet-cd/mxnet-cd-release-job"]
+return job
+
+
+def get_latest_build_number(job):
+return job.get_last_build().get_number()
+
+
+def get_build_from_build_number(job, build_number):
+return job.get_build(build_number)
+
+
+def get_build_timestamp(build):
+return build.get_timestamp()
+
+
+def get_build_date(timestamp):
+return timestamp.date()
+
+
+def is_latest_day_build(current_build, latest_build):
+current_build_timestamp = get_build_timestamp(current_build)
+latest_build_timestamp = get_build_timestamp(latest_build)
+# if 2 builds are within 24 hours and on the same day

Review comment:
   So we are May 9 11:59pm UTC [lambda is triggered]
   
   Latest build as of May 9 11:59pm UTC is _Build 1088_ [May 7 9:54:39pm]
   
   
http://jenkins.mxnet-ci.amazon-ml.com/job/restricted-mxnet-cd/job/mxnet-cd-release-job/1088/
   
    If we relax the same day check & only keep the 24hour check.
   So all builds after May 6 9:55:39 pm will be accepted.
   1082 to 1088
   
   
   Here 1082 isn't supposed to be selected as it was part of the previous day 
trigger
   _Build 1082_ May 6, 2020, 9:57:08 PM [which is not supposed to be picked]
   
http://jenkins.mxnet-ci.amazon-ml.com/job/restricted-mxnet-cd/job/mxnet-cd-release-job/1082/
   
    Hence we need both checks





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

[GitHub] [incubator-mxnet] marcoabreu commented on pull request #18252: [CI] run operator tests with naive engine

[GitBox]

marcoabreu commented on pull request #18252:
URL: https://github.com/apache/incubator-mxnet/pull/18252#issuecomment-629411739


   What's our strategy to test multi threaded engine?



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

[GitHub] [incubator-mxnet] ptrendx commented on a change in pull request #18325: Change SGD with momentum to include momentum correction by default

[GitBox]

ptrendx commented on a change in pull request #18325:
URL: https://github.com/apache/incubator-mxnet/pull/18325#discussion_r425969856



##
File path: python/mxnet/optimizer/lars.py
##
@@ -252,6 +253,13 @@ def fused_step(self, indices, weights, grads, states):
 wd = wds[i]
 lr = lrs[i]
 lr *= self._get_lars(index, weight, grad, wd)
+# normal SGD picks up momentum correction by default,
+# so need to modify the momentum to undo that.
+# The correction term is previous_lr / current_lr.
+kwargs['momentum'] = (self.momentum * (self.last_lr / lr)) \

Review comment:
   Thanks, done





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

[GitHub] [incubator-mxnet] JonTanS closed issue #16822: Trouble building Mxnet 1.6.x

[GitBox]

JonTanS closed issue #16822:
URL: https://github.com/apache/incubator-mxnet/issues/16822


   



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

[GitHub] [incubator-mxnet] kpu commented on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

kpu commented on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629400657


   `dot` hasn't been taken over by MKLDNN = DNNL = oneAPI yet, so you shouldn't 
expect to see a difference there because it never calls MKLDNN's GEMM anyway.  
FullyConnected has been taken over by MKLDNN = DNNL = oneAPI.  This 
inconsistency is part of the bug report.  Your logs agree with these 
statements.  
   
   Do you really mean LHS (4, 512, 512)?  I'm talking about LHS (4,512) RHS 
(512, 512).  
   
   What CPU is this on?  Note that mine were on a Skylake Xeon `c9.9xlarge`.  
I'd expect the AVX512 kernels to be pretty different from the AVX2 kernels.  
   
   What OMP threading were you using?  I'd expect for matrices this small that 
thread scaling is terrible for anything but a tiny number of threads.  (I 
mentioned using 4).  



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

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya edited a comment on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629385677


   > Tested with MXNet 
[cfb474b](https://github.com/apache/incubator-mxnet/commit/cfb474ba743d5ea85161bf19875488f4cb409d3c).
 Compiled with mostly-default cmake settings:
   > 
   > ```shell
   > cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release ..
   > ```
   > 
   > Then when I run
   > 
   > ```
   > export MKL_VERBOSE=1
   > export MKLDNN_VERBOSE=1
   > python3
   > Python 3.6.9 (default, Nov  7 2019, 10:44:02) 
   > [GCC 8.3.0] on linux
   > Type "help", "copyright", "credits" or "license" for more information.
   > >>> import mxnet as mx
   > Numpy + Intel(R) MKL: THREADING LAYER: (null)
   > Numpy + Intel(R) MKL: setting Intel(R) MKL to use INTEL OpenMP runtime
   > Numpy + Intel(R) MKL: preloading libiomp5.so runtime
   > ```
   
   @kpuatamazon @kpu
   Running on Ubuntu 18.04 [which doesn't have MKL installed by default] with 
default cmake config doesn't use MKL as blas.
   Hence we can't get the above exports.
   
   Thus for Ubuntu 18.04 base AMI, one has to install MKL in /opt/intel & 
update the cmake command to
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   This I found uses mkl as blas & export MKL_VERBOSE=1 confirms it.
   
   With this addition to both [default & workaround] I reran the opperf & I 
didn't see much perf differences.
   
   ## Commands
   Default
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   Workaround
   ```
   export CXXFLAGS="${CXXFLAGS} -DUSE_MKL -I/opt/intel/mkl/include"
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   ## Logs
    Default
   Batch_dot
   ```
   MKL_VERBOSE 
SGEMM_BATCH(T,N,0x7fdafe19ecec,0x7fdafe19ecf0,0x7fdafe19ecf4,0x7fdafe19ed04,0x7fda6001dfd0,0x7fdafe19ecf8,0x7fda6001e490,0x7fdafe19ecfc,0x7fdafe19ed08,0x3fd7ec0,0x7fdafe19ed00,0x7fdafe19ec28,0x7fdafe
 28.71ms CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   MKL_VERBOSE 
SGEMM_BATCH(T,N,0x7fdafe19ecec,0x7fdafe19ecf0,0x7fdafe19ecf4,0x7fdafe19ed04,0x7fda6001dfd0,0x7fdafe19ecf8,0x7fda6001e490,0x7fdafe19ecfc,0x7fdafe19ed08,0x3fd7ec0,0x7fdafe19ed00,0x7fdafe19ec28,0x7fdafe
 28.53ms CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   ```
   FC
   ```
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.0551758
   
dnnl_verbose,exec,cpu,inner_product,gemm:jit,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.0559082
   ```
   
    Workaround
   Batch_dot [same as default]
   ```
   MKL_VERBOSE 
SGEMM_BATCH(T,N,0x7f985b78acec,0x7f985b78acf0,0x7f985b78acf4,0x7f985b78ad04,0x7f97b4016cd0,0x7f985b78acf8,0x7f97b401e550,0x7f985b78acfc,0x7f985b78ad08,0x26f2890,0x7f985b78ad00,0x7f985b78ac28,0x7f985b
 28.72ms CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   MKL_VERBOSE 
SGEMM_BATCH(T,N,0x7f985b78acec,0x7f985b78acf0,0x7f985b78acf4,0x7f985b78ad04,0x7f97b4016cd0,0x7f985b78acf8,0x7f97b401e550,0x7f985b78acfc,0x7f985b78ad08,0x26f2890,0x7f985b78ad00,0x7f985b78ac28,0x7f985b
 28.77ms CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   ```
   FC [additional MKL_VERBOSE before DNNL_VERBOSE]
   ```
   MKL_VERBOSE 
SGEMM(T,N,512,4,512,0x7f985b789c28,0x7f97b5e52e80,512,0x7f97b401e600,512,0x7f985b789c30,0x7f976e89dd80,512)
 39.68us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb4ic512oc512,0.0769043
   MKL_VERBOSE 
SGEMM(T,N,512,5,2048,0x7f985b789c28,0x7f976c400100,2048,0x7f976e887100,2048,0x7f985b789c30,0x7f976e8da000,512)
 79.41us CNR:OFF Dyn:1 FastMM:1 TID:0  NThr:16
   
dnnl_verbose,exec,cpu,inner_product,gemm:blas,forward_inference,src_f32::blocked:ab:f0
 wei_f32::blocked:ab:f0 bia_undef::undef::f0 
dst_f32::blocked:ab:f0,,,mb5ic2048oc512,0.11377
   ```
   ## Results
Operator| LHS   | RHS   | MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | Dot| (4, 512, 512) | (4, 512, 512) | 4.1112
| 4.8241|
   || (5, 512, 512) | (5, 512, 512) | 6.4421
| 7.607 |
   || (5, 512, 1536 | (5, 512, 1536)| 20.3648   
| 19.2217   |
   || (5, 512, 2048)| (5, 512, 2048)| 23.3236   
| 23.2849   |
   || (5, 2048, 512)| (5, 2048, 512)| 123.1235  
| 123.9806  |
   ||   |   |   
|   |
   | Batch_dot  | (4, 

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya edited a comment on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629385677


   > Tested with MXNet 
[cfb474b](https://github.com/apache/incubator-mxnet/commit/cfb474ba743d5ea85161bf19875488f4cb409d3c).
 Compiled with mostly-default cmake settings:
   > 
   > ```shell
   > cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release ..
   > ```
   > 
   > Then when I run
   > 
   > ```
   > export MKL_VERBOSE=1
   > export MKLDNN_VERBOSE=1
   > python3
   > Python 3.6.9 (default, Nov  7 2019, 10:44:02) 
   > [GCC 8.3.0] on linux
   > Type "help", "copyright", "credits" or "license" for more information.
   > >>> import mxnet as mx
   > Numpy + Intel(R) MKL: THREADING LAYER: (null)
   > Numpy + Intel(R) MKL: setting Intel(R) MKL to use INTEL OpenMP runtime
   > Numpy + Intel(R) MKL: preloading libiomp5.so runtime
   > ```
   
   @kpuatamazon @kpu
   Running on Ubuntu 18.04 [which doesn't have MKL installed by default] with 
default cmake config doesn't use MKL as blas.
   Hence we can't get the above exports.
   
   Thus for Ubuntu 18.04 base AMI, one has to install MKL in /opt/intel & 
update the cmake command to
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   This I found uses mkl as blas & export MKL_VERBOSE=1 confirms it.
   
   With this addition to both [default & workaround] I reran the opperf & I 
didn't see much perf differences.
   
   Default
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   Workaround
   ```
   export CXXFLAGS="${CXXFLAGS} -DUSE_MKL -I/opt/intel/mkl/include"
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   Results
Operator| LHS   | RHS   | MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | Dot| (4, 512, 512) | (4, 512, 512) | 4.1112
| 4.8241|
   || (5, 512, 512) | (5, 512, 512) | 6.4421
| 7.607 |
   || (5, 512, 1536 | (5, 512, 1536)| 20.3648   
| 19.2217   |
   || (5, 512, 2048)| (5, 512, 2048)| 23.3236   
| 23.2849   |
   || (5, 2048, 512)| (5, 2048, 512)| 123.1235  
| 123.9806  |
   ||   |   |   
|   |
   | Batch_dot  | (4, 512, 512) | (4, 512, 512) | 1.4105
| 1.407 |
   || (5, 512, 512) | (5, 512, 512) | 1.7558
| 1.7511|
   || (5, 512, 1536)| (5, 512, 1536)| 6.5931
| 6.5585|
   || (5, 512, 2048)| (5, 512, 2048)| 9.1452
| 9.1031|
   || (5, 2048, 512)| (5, 2048, 512)| 29.0192   
| 28.9236   |
   
Operator| Data  | Weight| MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | FC | (4, 512)  | (512, 512)| 0.057 
| 0.0685|
   || (5, 512)  | (512, 512)| 0.0591
| 0.0698|
   || (5, 512)  | (1536, 512)   | 0.0823
| 0.0939|
   || (5, 512)  | (2048, 512)   | 0.0916
| 0.1026|
   || (5, 2048) | (512, 2048)   | 0.1146
| 0.1267|



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

[GitHub] [incubator-mxnet] szha closed issue #18334: test_numpy_ndarray.py::test_np_ndarray_boolean_indexing

[GitBox]

szha closed issue #18334:
URL: https://github.com/apache/incubator-mxnet/issues/18334


   



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

[GitHub] [incubator-mxnet] ChaiBapchya commented on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya commented on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-629385677


   > Tested with MXNet 
[cfb474b](https://github.com/apache/incubator-mxnet/commit/cfb474ba743d5ea85161bf19875488f4cb409d3c).
 Compiled with mostly-default cmake settings:
   > 
   > ```shell
   > cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release ..
   > ```
   > 
   > Then when I run
   > 
   > ```
   > export MKL_VERBOSE=1
   > export MKLDNN_VERBOSE=1
   > python3
   > Python 3.6.9 (default, Nov  7 2019, 10:44:02) 
   > [GCC 8.3.0] on linux
   > Type "help", "copyright", "credits" or "license" for more information.
   > >>> import mxnet as mx
   > Numpy + Intel(R) MKL: THREADING LAYER: (null)
   > Numpy + Intel(R) MKL: setting Intel(R) MKL to use INTEL OpenMP runtime
   > Numpy + Intel(R) MKL: preloading libiomp5.so runtime
   > ```
   
   Running on Ubuntu 18.04 [which doesn't have MKL installed by default] with 
default cmake config doesn't use MKL as blas.
   Hence we can't get the above exports.
   
   Thus for Ubuntu 18.04 base AMI, one has to install MKL in /opt/intel & 
update the cmake command to
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   This I found uses mkl as blas & export MKL_VERBOSE=1 confirms it.
   
   With this addition to both [default & workaround] I reran the opperf & I 
didn't see much perf differences.
   
   Default
   ```
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   Workaround
   ```
   export CXXFLAGS="${CXXFLAGS} -DUSE_MKL -I/opt/intel/mkl/include"
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DUSE_BLAS=mkl ..
   ```
   
   Results
Operator| LHS   | RHS   | MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | Dot| (4, 512, 512) | (4, 512, 512) | 4.1112
| 4.8241|
   || (5, 512, 512) | (5, 512, 512) | 6.4421
| 7.607 |
   || (5, 512, 1536 | (5, 512, 1536)| 20.3648   
| 19.2217   |
   || (5, 512, 2048)| (5, 512, 2048)| 23.3236   
| 23.2849   |
   || (5, 2048, 512)| (5, 2048, 512)| 123.1235  
| 123.9806  |
   ||   |   |   
|   |
   | Batch_dot  | (4, 512, 512) | (4, 512, 512) | 1.4105
| 1.407 |
   || (5, 512, 512) | (5, 512, 512) | 1.7558
| 1.7511|
   || (5, 512, 1536)| (5, 512, 1536)| 6.5931
| 6.5585|
   || (5, 512, 2048)| (5, 512, 2048)| 9.1452
| 9.1031|
   || (5, 2048, 512)| (5, 2048, 512)| 29.0192   
| 28.9236   |
   
Operator| Data  | Weight| MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | FC | (4, 512)  | (512, 512)| 0.057 
| 0.0685|
   || (5, 512)  | (512, 512)| 0.0591
| 0.0698|
   || (5, 512)  | (1536, 512)   | 0.0823
| 0.0939|
   || (5, 512)  | (2048, 512)   | 0.0916
| 0.1026|
   || (5, 2048) | (512, 2048)   | 0.1146
| 0.1267|



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

[GitHub] [incubator-mxnet] ChaiBapchya edited a comment on issue #17980: When compiled with MKL, fully_connected calls DNNL while dot and batch_dot call MKL

[GitBox]

ChaiBapchya edited a comment on issue #17980:
URL: 
https://github.com/apache/incubator-mxnet/issues/17980#issuecomment-623803913


   > In case somebody finds this issue and wants their optimized build, here is 
a different workaround that removes the need for `LD_PRELOAD`. Just do this 
before running cmake the first time:
   > 
   > ```shell
   > export CXXFLAGS="${CXXFLAGS} -DUSE_MKL -I/opt/intel/mkl/include"
   > ```
   > 
   > Then `cmake` can be run normally:
   > 
   > ```shell
   > cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release ..
   > ```
   > 
   > and the compiled MXNet can be run normally without any special environment 
variables.
   
   @kpuatamazon Hi I was trying to benchmark using opperf for mkl [default] vs 
workaround
   And despite ensuring mkl is installed & using export CXXFlags followed by 
usual cmake command, build failed with 
   ```
   gemm.cpp:(.text+0xb45): undefined reference to `cblas_gemm_s8u8s32'
   ```
   
   I tried the undocumented abominable kludge option you mentioned and that 
worked smoothly.
   ```
   export LD_PRELOAD=/opt/intel/mkl/lib/intel64/libmkl_rt.so
   rm -rf build/
   mkdir -p build && cd build
   cmake -GNinja -DUSE_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -D_DNNL_USE_MKL=FULL 
-DMKLINC=/opt/intel/mkl/include ..
   cmake --build . --parallel 1024
   ```
   
   Script for OpPerf : 
https://gist.github.com/ChaiBapchya/5f2342f75ddeb1e21f14acac665c76ad
   
   Results
   | Operator   | LHS   | RHS   | MKL Default   
| MKL Workaround|
   |--- |   |   |-  
|   |
   | Dot| (4, 512, 512) | (4, 512, 512) | 15.1122   
| 4.1254|
   || (5, 512, 512) | (5, 512, 512) | 38.1678   
| 7.5323|
   || (5, 512, 1536 | (5, 512, 1536)| 21.6601   
| 19.2503   |
   || (5, 512, 2048)| (5, 512, 2048)| 29.0369   
| 23.7432   |
   || (5, 2048, 512)| (5, 2048, 512)| 167.5528  
| 129.9957  |
   ||   |   |   
|   |
   | Batch_dot  | (4, 512, 512) | (4, 512, 512) | 1.7898
| 1.5445|
   || (5, 512, 512) | (5, 512, 512) | 2.2457
| 1.9361|
   || (5, 512, 1536)| (5, 512, 1536)| 6.1453
| 5.4034|
   || (5, 512, 2048)| (5, 512, 2048)| 8.246 
| 8.0442|
   || (5, 2048, 512)| (5, 2048, 512)| 160.6243  
| 29.0772   |
   ||   |   |   
|   |
   ||  **Data** |  **Weight**   |   
|   |
   | FC | (4, 512)  | (512, 512)| 0.0609
| 0.068 |
   || (5, 512)  | (512, 512)| 0.0633
| 0.0731|
   || (5, 512)  | (1536, 512)   | 0.0916
| 0.0996|
   || (5, 512)  | (2048, 512)   | 0.1081
| 0.1084|
   ||   |   |   
|   |
   
   However @kpuatamazon when I try to test out with default [i.e. default -> 
workaround -> default] by unsetting the environment variable LD_PRELOAD, it 
failed to build default with `gemm.cpp:(.text+0xe6b): undefined reference to 
`cblas_gemm_s8u8s32'`



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

[incubator-mxnet] branch master updated (fec534a -> 3e676fc)

[lausen]

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

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


from fec534a  Fix missing MKLDNN headers (#18310)
 add 3e676fc  Fix memory leaks in Gluon (#18328)

No new revisions were added by this update.

Summary of changes:
 python/mxnet/gluon/block.py| 25 
 tests/python/unittest/test_gluon.py| 38 ++
 tests/python/unittest/test_thread_local.py |  5 ++--
 3 files changed, 56 insertions(+), 12 deletions(-)

[incubator-mxnet] branch master updated (fec534a -> 3e676fc)

[lausen]

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

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


from fec534a  Fix missing MKLDNN headers (#18310)
 add 3e676fc  Fix memory leaks in Gluon (#18328)

No new revisions were added by this update.

Summary of changes:
 python/mxnet/gluon/block.py| 25 
 tests/python/unittest/test_gluon.py| 38 ++
 tests/python/unittest/test_thread_local.py |  5 ++--
 3 files changed, 56 insertions(+), 12 deletions(-)

[GitHub] [incubator-mxnet] leezu commented on pull request #18328: Fix memory leaks in Gluon

[GitBox]

leezu commented on pull request #18328:
URL: https://github.com/apache/incubator-mxnet/pull/18328#issuecomment-629372543


   @ciyongch should we backport this to 1.7?



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

[GitHub] [incubator-mxnet] leezu merged pull request #18328: Fix memory leaks in Gluon

[GitBox]

leezu merged pull request #18328:
URL: https://github.com/apache/incubator-mxnet/pull/18328


   



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

[GitHub] [incubator-mxnet] szha opened a new issue #18334: test_numpy_ndarray.py::test_np_ndarray_boolean_indexing

[GitBox]

szha opened a new issue #18334:
URL: https://github.com/apache/incubator-mxnet/issues/18334


   ## Description
   test_numpy_ndarray.py::test_np_ndarray_boolean_indexing
   
   ## Occurrences
   
http://jenkins.mxnet-ci.amazon-ml.com/blue/rest/organizations/jenkins/pipelines/mxnet-validation/pipelines/centos-cpu/branches/PR-18252/runs/9/nodes/236/steps/284/log/?start=0
   
   ## What have you tried to solve it?
   
   1. rewrite and break down the test.
   2. mark as flaky



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

[GitHub] [incubator-mxnet] John1231983 opened a new issue #18333: How to implement cross entropy for binary segmentation using symbol only?

[GitBox]

John1231983 opened a new issue #18333:
URL: https://github.com/apache/incubator-mxnet/issues/18333


   Given a softmax classified feature ff (`Bx2xHxW`). And a target label size 
of `Bx1xHxW` . I want to implement cross entropy loss using symbol only. This 
is my implementation
   ```
   # target size of Bx1xHxW
   target_squeeze = mx.symbol.squeeze(target, axis=1) #size of BxHxW
   target_squeeze = mx.sym.one_hot(target_squeeze, depth = 2, on_value = -1.0, 
off_value = 0.0) 
   # Transpose from BxHxWx2 to Bx2xHxW
target_squeeze = mx.symbol.transpose(target_squeeze, axes=(0,3,1,2))
   # Get log of feature f
   f_log  = mx.sym.log(f)
   batch_size =32
   f_sum = mx.symbol.sum(target_squeeze * f_log)/batch_size
   f_sum = mx.symbol.MakeLoss(f_sum, name = 'loss_ce')
   ```
   Is my implementation correct? If not, please correct it for me. Thanks in 
advantage



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

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

[GitBox]

D-Roberts commented on a change in pull request #18197:
URL: https://github.com/apache/incubator-mxnet/pull/18197#discussion_r425903506



##
File path: src/operator/numpy/linalg/np_qr-inl.h
##
@@ -542,36 +591,119 @@ void QrBackwardImpl(const TBlob& grad_a,
 const nnvm::NodeAttrs& attrs) {
   Stream *s = ctx.get_stream();
   const mxnet::TShape& a_shape = a.shape_;
+  const mxnet::TShape& q_shape = q.shape_;
   const mxnet::TShape& r_shape = r.shape_;
   const int a_ndim = a_shape.ndim();
+  const int m = a.size(a_ndim - 2);
   const int n = a.size(a_ndim - 1);
 
   if (kNullOp == req[0]) { return; }
 
   if (0U == a_shape.Size()) { return; }
 
   MSHADOW_SGL_DBL_TYPE_SWITCH(grad_a.type_flag_, DType, {
-// case m >= n; Q of same shape with A and R is (n, n)
-DType *m_ptr = reinterpret_cast(workspace.dptr_);
-DType *grad_a_ptr = m_ptr + r_shape.Size();
-TBlob temp_m(m_ptr, r_shape, xpu::kDevMask);
+// common for all shapes (m, n)
+DType *grad_a_ptr = reinterpret_cast(workspace.dptr_);
 TBlob grad_a_data(grad_a_ptr, a_shape, xpu::kDevMask);
-// dR_T
-mxnet_op::Kernel::Launch(
-  s, r_shape.Size(), grad_r.dptr(), m_ptr, n, n, n * n);
-
-qr_backward::op(grad_a_data.FlatToKD(s),
-grad_q.FlatToKD(s),
-grad_r.FlatToKD(s),
-a.FlatToKD(s),
-q.FlatToKD(s),
-r.FlatToKD(s),
-temp_m.FlatToKD(s),
-ctx, attrs);
-
+if (m >= n) {
+  // Q of same shape with A (m, n) and R is (n, n)
+  DType *m_ptr = grad_a_ptr + a_shape.Size();
+  TBlob temp_m(m_ptr, r_shape, xpu::kDevMask);
+  // dR_T
+  mxnet_op::Kernel::Launch(
+s, r_shape.Size(), grad_r.dptr(), m_ptr, n, n, n * n);
+  qr_backward::op(grad_a_data.FlatToKD(s),
+  grad_q.FlatToKD(s),
+  grad_r.FlatToKD(s),
+  q.FlatToKD(s),
+  r.FlatToKD(s),
+  temp_m.FlatToKD(s),
+  ctx, attrs);
+} else {
+  // R is same shape with A (m, n) and Q is (m, m)
+  // Partition A = (X | Y); R = (U | V)
+  // X and U are (m, m); Y and V are (m, n - m)
+  mxnet::TShape v_shape(q_shape);
+  v_shape[a_ndim - 1] = n - m;
+
+  DType *m_ptr = grad_a_ptr + a_shape.Size();
+  DType *u_ptr = m_ptr + q_shape.Size();
+  DType *dq_prime_ptr = u_ptr + q_shape.Size();
+  DType *dv_ptr = dq_prime_ptr + q_shape.Size();
+  DType *y_ptr = dv_ptr + v_shape.Size();
+  DType *du_ptr = y_ptr + v_shape.Size();
+  DType *dx_ptr = du_ptr + q_shape.Size();
+  DType *dy_ptr = dx_ptr + q_shape.Size();
+
+  TBlob temp_m(m_ptr, q_shape, xpu::kDevMask);
+  TBlob u_data(u_ptr, q_shape, xpu::kDevMask);
+  TBlob dq_prime_data(dq_prime_ptr, q_shape, xpu::kDevMask);
+  TBlob dv_data(dv_ptr, v_shape, xpu::kDevMask);
+  TBlob y_data(y_ptr, v_shape, xpu::kDevMask);
+  TBlob du_data(du_ptr, q_shape, xpu::kDevMask);
+  TBlob dx_data(dx_ptr, q_shape, xpu::kDevMask);
+  TBlob dy_data(dy_ptr, v_shape, xpu::kDevMask);
+
+  Tensor R = r.FlatToKD(s);
+  Tensor dR = grad_r.FlatToKD(s);
+  Tensor Q = q.FlatToKD(s);
+  Tensor dQ = grad_q.FlatToKD(s);
+  Tensor dQ_prime = dq_prime_data.FlatToKD(s);
+  Tensor A = a.FlatToKD(s);
+  Tensor dA = grad_a_data.FlatToKD(s);
+  Tensor U = u_data.FlatToKD(s);
+  Tensor dU = du_data.FlatToKD(s);
+  Tensor dV = dv_data.FlatToKD(s);
+  Tensor Y = y_data.FlatToKD(s);
+  Tensor dX = dx_data.FlatToKD(s);
+  Tensor dY = dy_data.FlatToKD(s);
+  Tensor M = temp_m.FlatToKD(s);
+
+  // U
+  for (index_t i = 0; i < R.size(0); ++i) {
+const Tensor& Ri = R[i];
+const Tensor& Ui = U[i];
+Tensor Um(Ri.dptr_, Shape2(m, m), Ri.stride_, s);
+Copy(Ui, Um, s);
+  }
+  // dU
+  for (index_t i = 0; i < dR.size(0); ++i) {
+const Tensor& dRi = dR[i];
+const Tensor& dUi = dU[i];
+Tensor dUm(dRi.dptr_, Shape2(m, m), dRi.stride_, s);
+Copy(dUi, dUm, s);
+  }
+  // Y
+  mxnet_op::Kernel::Launch(
+s, A.size(0), m, n, A.dptr_, A.stride_, Y.dptr_, Y.stride_);
+  // dV
+  mxnet_op::Kernel::Launch(
+s, dR.size(0), m, n, dR.dptr_, dR.stride_, dV.dptr_, dV.stride_);
+  // store dU_T in M
+  mxnet_op::Kernel::Launch(
+s, q_shape.Size(), dU.dptr_, m_ptr, m, m, m * m);
+  // dq_prime = dQ
+  Copy(dQ_prime, dQ, s);
+  // dq_prime = dQ+Y@dV.T
+  gemm::op(Y, dV, dQ_prime, DType(1.0), DType(1.0), false, true, s);
+  // dX = op call
+  qr_backward::op(dX,
+  dQ_prime,
+  dU,
+  Q,
+  U,
+  M,
+  ctx, attrs);
+  // dY = Q@dV
+  gemm::op(Q, dV, dY, 

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

[GitBox]

hzfan commented on a change in pull request #18197:
URL: https://github.com/apache/incubator-mxnet/pull/18197#discussion_r425852127



##
File path: src/operator/numpy/linalg/np_qr-inl.h
##
@@ -542,36 +591,119 @@ void QrBackwardImpl(const TBlob& grad_a,
 const nnvm::NodeAttrs& attrs) {
   Stream *s = ctx.get_stream();
   const mxnet::TShape& a_shape = a.shape_;
+  const mxnet::TShape& q_shape = q.shape_;
   const mxnet::TShape& r_shape = r.shape_;
   const int a_ndim = a_shape.ndim();
+  const int m = a.size(a_ndim - 2);
   const int n = a.size(a_ndim - 1);
 
   if (kNullOp == req[0]) { return; }
 
   if (0U == a_shape.Size()) { return; }
 
   MSHADOW_SGL_DBL_TYPE_SWITCH(grad_a.type_flag_, DType, {
-// case m >= n; Q of same shape with A and R is (n, n)
-DType *m_ptr = reinterpret_cast(workspace.dptr_);
-DType *grad_a_ptr = m_ptr + r_shape.Size();
-TBlob temp_m(m_ptr, r_shape, xpu::kDevMask);
+// common for all shapes (m, n)
+DType *grad_a_ptr = reinterpret_cast(workspace.dptr_);
 TBlob grad_a_data(grad_a_ptr, a_shape, xpu::kDevMask);
-// dR_T
-mxnet_op::Kernel::Launch(
-  s, r_shape.Size(), grad_r.dptr(), m_ptr, n, n, n * n);
-
-qr_backward::op(grad_a_data.FlatToKD(s),
-grad_q.FlatToKD(s),
-grad_r.FlatToKD(s),
-a.FlatToKD(s),
-q.FlatToKD(s),
-r.FlatToKD(s),
-temp_m.FlatToKD(s),
-ctx, attrs);
-
+if (m >= n) {
+  // Q of same shape with A (m, n) and R is (n, n)
+  DType *m_ptr = grad_a_ptr + a_shape.Size();
+  TBlob temp_m(m_ptr, r_shape, xpu::kDevMask);
+  // dR_T
+  mxnet_op::Kernel::Launch(
+s, r_shape.Size(), grad_r.dptr(), m_ptr, n, n, n * n);
+  qr_backward::op(grad_a_data.FlatToKD(s),
+  grad_q.FlatToKD(s),
+  grad_r.FlatToKD(s),
+  q.FlatToKD(s),
+  r.FlatToKD(s),
+  temp_m.FlatToKD(s),
+  ctx, attrs);
+} else {
+  // R is same shape with A (m, n) and Q is (m, m)
+  // Partition A = (X | Y); R = (U | V)
+  // X and U are (m, m); Y and V are (m, n - m)
+  mxnet::TShape v_shape(q_shape);
+  v_shape[a_ndim - 1] = n - m;
+
+  DType *m_ptr = grad_a_ptr + a_shape.Size();
+  DType *u_ptr = m_ptr + q_shape.Size();
+  DType *dq_prime_ptr = u_ptr + q_shape.Size();
+  DType *dv_ptr = dq_prime_ptr + q_shape.Size();
+  DType *y_ptr = dv_ptr + v_shape.Size();
+  DType *du_ptr = y_ptr + v_shape.Size();
+  DType *dx_ptr = du_ptr + q_shape.Size();
+  DType *dy_ptr = dx_ptr + q_shape.Size();
+
+  TBlob temp_m(m_ptr, q_shape, xpu::kDevMask);
+  TBlob u_data(u_ptr, q_shape, xpu::kDevMask);
+  TBlob dq_prime_data(dq_prime_ptr, q_shape, xpu::kDevMask);
+  TBlob dv_data(dv_ptr, v_shape, xpu::kDevMask);
+  TBlob y_data(y_ptr, v_shape, xpu::kDevMask);
+  TBlob du_data(du_ptr, q_shape, xpu::kDevMask);
+  TBlob dx_data(dx_ptr, q_shape, xpu::kDevMask);
+  TBlob dy_data(dy_ptr, v_shape, xpu::kDevMask);
+
+  Tensor R = r.FlatToKD(s);
+  Tensor dR = grad_r.FlatToKD(s);
+  Tensor Q = q.FlatToKD(s);
+  Tensor dQ = grad_q.FlatToKD(s);
+  Tensor dQ_prime = dq_prime_data.FlatToKD(s);
+  Tensor A = a.FlatToKD(s);
+  Tensor dA = grad_a_data.FlatToKD(s);
+  Tensor U = u_data.FlatToKD(s);
+  Tensor dU = du_data.FlatToKD(s);
+  Tensor dV = dv_data.FlatToKD(s);
+  Tensor Y = y_data.FlatToKD(s);
+  Tensor dX = dx_data.FlatToKD(s);
+  Tensor dY = dy_data.FlatToKD(s);
+  Tensor M = temp_m.FlatToKD(s);
+
+  // U
+  for (index_t i = 0; i < R.size(0); ++i) {
+const Tensor& Ri = R[i];
+const Tensor& Ui = U[i];
+Tensor Um(Ri.dptr_, Shape2(m, m), Ri.stride_, s);
+Copy(Ui, Um, s);
+  }
+  // dU
+  for (index_t i = 0; i < dR.size(0); ++i) {
+const Tensor& dRi = dR[i];
+const Tensor& dUi = dU[i];
+Tensor dUm(dRi.dptr_, Shape2(m, m), dRi.stride_, s);
+Copy(dUi, dUm, s);
+  }
+  // Y
+  mxnet_op::Kernel::Launch(
+s, A.size(0), m, n, A.dptr_, A.stride_, Y.dptr_, Y.stride_);
+  // dV
+  mxnet_op::Kernel::Launch(
+s, dR.size(0), m, n, dR.dptr_, dR.stride_, dV.dptr_, dV.stride_);
+  // store dU_T in M
+  mxnet_op::Kernel::Launch(
+s, q_shape.Size(), dU.dptr_, m_ptr, m, m, m * m);
+  // dq_prime = dQ
+  Copy(dQ_prime, dQ, s);
+  // dq_prime = dQ+Y@dV.T
+  gemm::op(Y, dV, dQ_prime, DType(1.0), DType(1.0), false, true, s);
+  // dX = op call
+  qr_backward::op(dX,
+  dQ_prime,
+  dU,
+  Q,
+  U,
+  M,
+  ctx, attrs);
+  // dY = Q@dV
+  gemm::op(Q, dV, dY, 

[GitHub] [incubator-mxnet] hzfan commented on a change in pull request #17749: Fix races in block scope

[GitBox]

hzfan commented on a change in pull request #17749:
URL: https://github.com/apache/incubator-mxnet/pull/17749#discussion_r425799124



##
File path: python/mxnet/gluon/block.py
##
@@ -91,22 +92,22 @@ def create(prefix, params, hint):
 def __enter__(self):
 if self._block._empty_prefix:
 return self
-self._old_scope = getattr(_BlockScope._current, "value", None)
+self._local._old_scope = getattr(_BlockScope._current, "value", None)
 _BlockScope._current.value = self

Review comment:
   Seems that it has been thread-local ?
   
https://github.com/apache/incubator-mxnet/blob/f3e8dc14de881a5b8ccc3ff5985167f8d6fa2fb8/python/mxnet/gluon/block.py#L47
   





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

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

[GitBox]

mxnet-bot commented on pull request #18197:
URL: https://github.com/apache/incubator-mxnet/pull/18197#issuecomment-629227942


   Jenkins CI successfully triggered : [centos-cpu, unix-cpu]



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

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

[GitBox]

D-Roberts commented on pull request #18197:
URL: https://github.com/apache/incubator-mxnet/pull/18197#issuecomment-629227877


   @mxnet-bot run ci [centos-cpu, unix-cpu]



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

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #18325: Change SGD with momentum to include momentum correction by default

[GitBox]

wkcn commented on a change in pull request #18325:
URL: https://github.com/apache/incubator-mxnet/pull/18325#discussion_r425780192



##
File path: python/mxnet/optimizer/lars.py
##
@@ -252,6 +253,13 @@ def fused_step(self, indices, weights, grads, states):
 wd = wds[i]
 lr = lrs[i]
 lr *= self._get_lars(index, weight, grad, wd)
+# normal SGD picks up momentum correction by default,
+# so need to modify the momentum to undo that.
+# The correction term is previous_lr / current_lr.
+kwargs['momentum'] = (self.momentum * (self.last_lr / lr)) \

Review comment:
   Thank you for the fix!
   
   It will reduce the times of getting an item of `self.last_lr`.
   ```python
   kwargs['momentum'] = (self.momentum * (self.last_lr.get(index, lr) / lr)) \
if lr != 0 else \
self.momentum
   ```





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

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

[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 4327a7f  Bump the publish timestamp.
4327a7f is described below

commit 4327a7fc85f8b43c91c8a1782e83820e5b633187
Author: mxnet-ci 
AuthorDate: Fri May 15 12:48:04 2020 +

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..1fd79c6
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Fri May 15 12:48:04 UTC 2020

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18320: Improve log_softmax op performance by using DNNL support

[GitBox]

mxnet-bot commented on pull request #18320:
URL: https://github.com/apache/incubator-mxnet/pull/18320#issuecomment-629132145


   Jenkins CI successfully triggered : [unix-cpu, centos-cpu]



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

[GitHub] [incubator-mxnet] bgawrych commented on pull request #18320: Improve log_softmax op performance by using DNNL support

[GitBox]

bgawrych commented on pull request #18320:
URL: https://github.com/apache/incubator-mxnet/pull/18320#issuecomment-629132092


   @mxnet-bot run ci [centos-cpu, unix-cpu]



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

[GitHub] [incubator-mxnet] Kh4L opened a new pull request #18332: Fix FInferType of ops to support partial type inference

[GitBox]

Kh4L opened a new pull request #18332:
URL: https://github.com/apache/incubator-mxnet/pull/18332


   ## Description ##
   As described in #16757 , a list of operators `FInferType` functions were 
throwing an exception if the input type is not defined (dtype=`-1`).
   
   This was preventing us from running a partial type inference where some not 
(yet) defined input `dtype`s would trigger an exception.
   
   The expected behavior is that `FInferType` returns false when the type 
cannot be inferred correctly, a boolean value that the executor inference pass 
uses to wet the `Graph` attribute `dtype_num_unknown_nodes`.
   This attribute is then used by the executor or any graph modules (e.g. 
Partition API) to determine the correctness, or know if a new inference pass is 
required.
   
   This PR fixes it by making the `FInferType` functions return `false` when 
the input dtype is `-1`.
   
   



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18332: Fix FInferType of ops to support partial type inference

[GitBox]

mxnet-bot commented on pull request #18332:
URL: https://github.com/apache/incubator-mxnet/pull/18332#issuecomment-629131254


   Hey @Kh4L , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [clang, unix-gpu, edge, website, centos-gpu, 
unix-cpu, centos-cpu, miscellaneous, windows-cpu, sanity, windows-gpu]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] BenjaminCHEN2016 commented on pull request #18313: [Fix][Numpy][WIP] fix mix numpy scalar and MXNet numpy ndarray

[GitBox]

BenjaminCHEN2016 commented on pull request #18313:
URL: https://github.com/apache/incubator-mxnet/pull/18313#issuecomment-629091482


   Conversion rules:
   | Expression | a type | b type | out type|
   | --- | --- | --- | --- |
   | `a += b` | onp | mx_np | onp |
   | `a += b` | mx_np | onp | mx_np |
   | `c = a + b` | onp | mx_np | mx_np |
   | `c = a + b` | mx_np | onp | mx_np |



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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18160: [1.x] Add BatchNormWithReLU bf16 into AMP list

[GitBox]

mxnet-bot commented on pull request #18160:
URL: https://github.com/apache/incubator-mxnet/pull/18160#issuecomment-629089260


   Jenkins CI successfully triggered : [sanity]



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

[GitHub] [incubator-mxnet] xinyu-intel commented on pull request #18160: [1.x] Add BatchNormWithReLU bf16 into AMP list

[GitBox]

xinyu-intel commented on pull request #18160:
URL: https://github.com/apache/incubator-mxnet/pull/18160#issuecomment-629089225


   @mxnet-bot run ci [sanity]



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

[GitHub] [incubator-mxnet] xidulu commented on issue #18068: [Numpy] Failed for Boolean type

[GitBox]

xidulu commented on issue #18068:
URL: 
https://github.com/apache/incubator-mxnet/issues/18068#issuecomment-629089392


   Thanks for your issue,
   This is indeed a very interesting use case.
   I noticed that this problem also occurred in PyTorch:
   ```python
   >>> inputs = torch.tensor([1,1,1])
   >>> fake_data = torch.tensor([1,1,0])
   >>> 1 - (inputs == fake_data)
   Traceback (most recent call last):
 File "", line 1, in 
 File "/home/ubuntu/anaconda3/lib/python3.7/site-packages/torch/tensor.py", 
line 394, in __rsub__
   return _C._VariableFunctions.rsub(self, other)
   RuntimeError: Subtraction, the `-` operator, with a bool tensor is not 
supported. If you are trying to invert a mask, use the `~` or `logical_not()` 
operator instead.
   >>>
   ```
   
   If we are capable of providing a solution to this case, that would be a huge 
advantage over PyTorch.



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

[GitHub] [incubator-mxnet] acphile opened a new pull request #18331: No test metric perf

[GitBox]

acphile opened a new pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331


   related to https://github.com/apache/incubator-mxnet/issues/18330
   We think that "test_metric_perf" does not enforce anything and currently 
there are speed issues related to using mxnet.numpy. We should simple delete 
the test because it only prints the elapsed time.
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to 
the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) 
created (except PRs with tiny changes)
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [ ] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [ ] To the best of my 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

[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18331: No test metric perf

[GitBox]

mxnet-bot commented on pull request #18331:
URL: https://github.com/apache/incubator-mxnet/pull/18331#issuecomment-629086999


   Hey @acphile , 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 commands: 
   - To trigger all jobs: @mxnet-bot run ci [all] 
   - To trigger specific jobs: @mxnet-bot run ci [job1, job2] 
   *** 
   **CI supported jobs**: [windows-cpu, centos-cpu, windows-gpu, unix-cpu, 
centos-gpu, unix-gpu, clang, website, sanity, miscellaneous, edge]
   *** 
   _Note_: 
Only following 3 categories can trigger CI :PR Author, MXNet Committer, 
Jenkins Admin. 
   All CI tests must pass before the PR can be merged. 
   



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

[GitHub] [incubator-mxnet] sxjscience commented on pull request #18328: Fix memory leaks in Gluon

[GitBox]

sxjscience commented on pull request #18328:
URL: https://github.com/apache/incubator-mxnet/pull/18328#issuecomment-629079777


   @Jerryzcn I think this is also related to your previous benchmark.



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

[GitHub] [incubator-mxnet] leezu commented on issue #18330: test_metric_performance

[GitBox]

leezu commented on issue #18330:
URL: 
https://github.com/apache/incubator-mxnet/issues/18330#issuecomment-629077891


   The test doesn't actually enforce anything. It's output is not monitored 
(AFAIK) so it may be best to simply delete this 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

[GitHub] [incubator-mxnet] leezu commented on pull request #9705: [MXNET-91] Added unittest for benchmarking metric performance

[GitBox]

leezu commented on pull request #9705:
URL: https://github.com/apache/incubator-mxnet/pull/9705#issuecomment-629076060


   Why does this test only print numbers but doesn't actually enforce anything? 



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