eric-haibin-lin commented on issue #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319#issuecomment-397789580
@marcoabreu My work also depends on this PR.
https://github.com/apache/incubator-mxnet/pull/11306
-
piiswrong opened a new pull request #11320: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11320
This reverts commit 9b8eb567273d9521c639ea783e1117fe09cb3d64.
## Description ##
(Brief description on what this PR is about)
## Checklist ##
chenbiaolong commented on issue #8474: How to train variable input shape CNN?
URL:
https://github.com/apache/incubator-mxnet/issues/8474#issuecomment-397787011
@lanking520 @marcoabreu We found MXNet is convenient in c++
deployment,especially when we want to improve our throughputs. The C+
tqchen commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397786640
Just putting some comment about our existing effort here so that we don't
waste effort one things already tried. If there is really a solution out of
this
TaoLv commented on issue #11262: [MXNET-542] Fix mkldnn performance regression
+ improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#issuecomment-397785653
I verifed the performance of commit 83ce4f0 with inference of ResNet50,
VGG16 and Inception-v3. It looks go
haojin2 commented on issue #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319#issuecomment-397786323
@marcoabreu Can we bring this commit back? I've got some work (#11179) that
depends of part of this PR.
haojin2 commented on issue #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319#issuecomment-397786323
@marcoabreu Can we bring this commit back? I've got some ongoing work
(#11179) that depends of part of this PR.
chenbiaolong commented on issue #8474: How to train variable input shape CNN?
URL:
https://github.com/apache/incubator-mxnet/issues/8474#issuecomment-397514469
@lanking520 the main framework we use to train models is Tensorflow now.In
last year we use tensorflow train variable shape imag
TaoLv commented on issue #11262: [MXNET-542] Fix mkldnn performance regression
+ improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#issuecomment-397785653
I verifed the performance of commit 83ce4f0. It looks good to me, compared
with commit 9514a1e.
--
yjcn commented on issue #11242: Same model with c_predict_api gets an incorrect
result but it is right in python.
URL:
https://github.com/apache/incubator-mxnet/issues/11242#issuecomment-397784515
@kalyc
this is python ssd 's log.
![default](https://user-images.githubusercontent.co
larroy commented on issue #11302: [MXNET-553] Restructure dockcross dockerfiles
to fix caching
URL: https://github.com/apache/incubator-mxnet/pull/11302#issuecomment-397782922
LGTM I can try what I asked myself as I'm not sure my question was well
understood.
-
tqchen commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397782646
just to be fair, I dismissed the your comment with specific comment that
addresses your issue
https://github.com/apache/incubator-mxnet/pull/11313
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195892008
##
File path: src/operator/nn/mkldnn/mkldnn_base-inl.h
##
@@ -327,7 +327
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891771
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891794
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891904
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891782
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891956
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
TaoLv commented on a change in pull request #11262: [MXNET-542] Fix mkldnn
performance regression + improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#discussion_r195891865
##
File path: src/operator/nn/mkldnn/mkldnn_base.cc
##
@@ -77,26 +77,35
marcoabreu commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397780181
I was not the one who closed this PR, it was @tqchen. He was the one who
dismissed my review and blocked other people from reviewing it. So why
TaoLv commented on issue #11262: [MXNET-542] Fix mkldnn performance regression
+ improve test logging
URL: https://github.com/apache/incubator-mxnet/pull/11262#issuecomment-397780096
@azai91 Yes, please see the comments and description in
https://github.com/apache/incubator-mxnet/pull/1031
zheng-da commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397779855
I don't understand what the problem is. Because you veto, the PR should
never be merged?
We should open the PR and get proper reviews and
marcoabreu commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397779766
Exactly, that's all I'm asking for. Give the community enough time to review
the PR and address their concerns before merging it.
---
zheng-da commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397779690
@marcoabreu I think you should at least give people who understand the PR
the opportunity to review it.
---
zheng-da commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397779615
@marcoabreu I have been intensively using CachedOp with static memory
allocation this week. I also read the code this week. Other than t
marcoabreu commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397779609
Hi Da, thanks for your comments. I'm sure that CachedOp is behaving as
expected. The problems are described at
https://github.com/apache/incub
zheng-da commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397779383
@marcoabreu I have been intensively using CachedOp with static memory
allocation this week. I also read the code in the past week. Other than the
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch piiswrong-patch-3
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
was 29faa9f fix
The revisions that were on this branch are still contained in
other references; the
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch piiswrong-patch-2
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
was fcdca35 Update cached_op.cc
The revisions that were on this branch are still contained in
other
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch revert-11318-revert-11313-static
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
was c8a9374 Revert "Revert "Static alloc for hybridblock (#11313)"
(#11318)"
The re
tqchen closed pull request #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitH
tqchen opened a new pull request #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319
Bring back the changes of hybrid-blocks. The test coverage issues has been
addressed.
This is an
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397778667
This PR is already fixing the problems and added test coverages. If we just
block this, how can we move forward?
ThomasDelteil commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397778642
I think I have raised specific concerns (twice) that have gone unanswered,
just in case they got lost in the flow, let me write the
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397778667
This PR is already fixing the problems and added test coverages. If we just
block this, how can we move forward?
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397778646
This PR is already fixing the problems and added test coverages. If we just
block this, how can we move forward?
marcoabreu closed pull request #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319
This is an automated message from the Apache Git Service.
To respond to the message, please log on
marcoabreu commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-397778459
https://github.com/apache/incubator-mxnet/issues/11171 has not been addressed
Pedros review in https://github.com/apache/incubat
tqchen opened a new pull request #11319: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11319
Bring back the changes of hybrid-blocks. The test coverage issues has been
addressed.
This is an
This is an automated email from the ASF dual-hosted git repository.
zhasheng pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new 0070e1c Bump the publish ti
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-39458
If you have specific concerns, raise them. Your concern of lack of test
coverage has been addressed, and you are abusing your right as com
marcoabreu commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-39312
My as well as Thomas' concerns have not been addressed.
This is an a
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-39148
it seems you have been using veto too often lately, please @marcoabreu we
are not being unreasonable here, your concern about testcases h
marcoabreu closed pull request #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318
This is an automated message from the Apache Git Service.
To respond to the message, pleas
tqchen closed pull request #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318
This is an automated message from the Apache Git Service.
To respond to the message, please lo
marcoabreu opened a new pull request #11318: Revert "Static alloc for
hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318
Reverts apache/incubator-mxnet#11313
My veto has been ignored and thus I'm allowed to revert this PR. See
https://www.apache.org/foundation/vo
tqchen commented on issue #11318: Revert "Static alloc for hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318#issuecomment-39148
it seems p
This is an automated message from the Apache Git Service.
To r
marcoabreu opened a new pull request #11318: Revert "Static alloc for
hybridblock"
URL: https://github.com/apache/incubator-mxnet/pull/11318
Reverts apache/incubator-mxnet#11313
My veto has been ignored and thus I'm allowed to revert this PR. See
https://www.apache.org/foundation/vo
This is an automated email from the ASF dual-hosted git repository.
marcoabreu pushed a change to branch revert-11313-static
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
at 24f14f7 Revert "Static alloc for hybridblock (#11313)"
This branch includes the following
This is an automated email from the ASF dual-hosted git repository.
marcoabreu pushed a commit to branch revert-11313-static
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
commit 24f14f7fdbef427aab25738608267c57de645987
Author: Marco de Abreu
AuthorDate: Fri Jun 15 18:27:3
This is an automated email from the ASF dual-hosted git repository.
tqchen 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 5431e12 Static alloc for hybridblock (
tqchen commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397774432
I think @yzhliu knows more details, the original design of scala was done in
an implicit way by doing disposable at finalizer (like what you suggested).
tqchen closed pull request #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:
As this is a foreign pull r
tqchen commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397774432
I think @yzhliu knows more details, the original design of scala was done in
an implicit way by doing disposable at finalizer.
The problem is that
tqchen commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397774432
I think @yzhliu knows more details, the original design of scala was done in
an implicit way by doing disposable at finalizer (like what you suggested).
nswamy commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397773700
We could explore using WeakReferences like how ND4J does or use ObjectPoo.
Scala Implicits helps to remove the ugliness of boiler plate code. I'll write
marcoabreu commented on a change in pull request #11294: [MXNET-543][WIP]Scala
Spark test
URL: https://github.com/apache/incubator-mxnet/pull/11294#discussion_r195887039
##
File path:
scala-package/spark/src/test/scala/org/apache/mxnet/spark/MXNetGeneralSuite.scala
##
@@
tqchen commented on issue #11307: Memory Management
URL:
https://github.com/apache/incubator-mxnet/issues/11307#issuecomment-397770425
The fact is you can't. Unfortunately, we have to blame JVM in this case,
which does not have ARC(accurate ref counting). The gc works well for the
memorie
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a commit to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/v1.2.0 by this push:
new daa957c Fix discard_stderr (#1130
anirudh2290 closed pull request #11309: Backport fix discard_stderr
URL: https://github.com/apache/incubator-mxnet/pull/11309
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:
As this is a foreign pu
safrooze commented on issue #9881: Inconsistent weight decay logics in multiple
optimizers
URL:
https://github.com/apache/incubator-mxnet/issues/9881#issuecomment-397769492
Also there is this paper that discusses how weight_decay must be applied
independently from optimizer: https://arxiv
This is an automated email from the ASF dual-hosted git repository.
nswamy 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 02e8a71 [MXNET-539] Allow Scala users
nswamy closed pull request #11256: [MXNET-539] Allow Scala users to specify
data/label names for NDArrayIter
URL: https://github.com/apache/incubator-mxnet/pull/11256
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sa
lanking520 commented on a change in pull request #11294: [MXNET-543][WIP]Scala
Spark test
URL: https://github.com/apache/incubator-mxnet/pull/11294#discussion_r195884144
##
File path:
scala-package/spark/src/test/scala/org/apache/mxnet/spark/MXNetGeneralSuite.scala
##
@@
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a commit to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
commit 332056a3413d8065adf7ede5d6a789954b3a8326
Author: Thomas Delteil
AuthorDate: Thu Jun 14 14:10:11 2018 -0700
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a change to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 00ec21c Scala inference memory leak fix #11204 (#11216)
new f6a8cb7 Add NEWS and README
new
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a commit to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
commit 719beeaab96be2b282761ddbf479670226a77f83
Author: Eric Junyuan Xie
AuthorDate: Wed Jun 13 22:01:40 2018 -07
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a commit to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
commit f6a8cb76b7e182e86c4c9a382421baba56bd6cb4
Author: Anirudh Subramanian
AuthorDate: Fri Jun 15 00:15:06 2018
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a commit to branch v1.2.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
commit dfc17bb6ec2627e4ee7bb65c89cc16154c9fa2f9
Author: Eric Junyuan Xie
AuthorDate: Thu Jun 14 14:02:48 2018 -07
lanking520 commented on issue #11239: [MXNET-319] Javadoc fix
URL: https://github.com/apache/incubator-mxnet/pull/11239#issuecomment-397765994
Currently, I created a file called `FILEHASH` lives in the core used to
check if there is a change on generated base files. Add a require as well to
apeforest opened a new pull request #11317: Fix the inaccruate instruction on
installing nose
URL: https://github.com/apache/incubator-mxnet/pull/11317
## Description ##
The instruction for installing nose package should add the 'sudo' in the
command. Otherwise, the Mac OS bash would no
marcoabreu opened a new issue #11249: Flaky Scala test:
IllegalArgumentException: requirement failed: Failed to start ps scheduler
URL: https://github.com/apache/incubator-mxnet/issues/11249
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/incubator-mxnet/detail/PR-11246/1/
nswamy commented on issue #11249: Flaky Scala test: IllegalArgumentException:
requirement failed: Failed to start ps scheduler
URL:
https://github.com/apache/incubator-mxnet/issues/11249#issuecomment-397764278
I meant this is not blocking the pipeline, i'll keep it reopen and edit the
Iss
marcoabreu commented on issue #11265: Flaky test: Python 3: CPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11265#issuecomment-397763495
I don't think there is much we can do - especially considering our Windows
slaves are effectively a black box. Closing for now
---
marcoabreu closed issue #11265: Flaky test: Python 3: CPU Win
URL: https://github.com/apache/incubator-mxnet/issues/11265
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitH
lanking520 commented on issue #11292: [MXNET-531] CNN Examples for Scala new APi
URL: https://github.com/apache/incubator-mxnet/pull/11292#issuecomment-397763241
Currently facing the CI problem as shown below:
```
- Example CI - CNN Example *** FAILED ***
java.nio.charset.Malf
lanking520 commented on issue #11292: [MXNET-531] CNN Examples for Scala new APi
URL: https://github.com/apache/incubator-mxnet/pull/11292#issuecomment-397763241
Currently facing the CI problem as shown below:
```
- Example CI - CNN Example *** FAILED ***
java.nio.charset.Malf
asmushetzel commented on issue #11271: nd.topk regression with nan values
URL:
https://github.com/apache/incubator-mxnet/issues/11271#issuecomment-397763186
@piiswrong
I searched a bit through the documentation and the only place where I could
find something about NaN handling was in t
nswamy commented on issue #11249: Flaky Scala test: IllegalArgumentException:
requirement failed: Failed to start ps scheduler
URL:
https://github.com/apache/incubator-mxnet/issues/11249#issuecomment-397762978
This is resolved lets open new if there is an issue
---
nswamy closed issue #11249: Flaky Scala test: IllegalArgumentException:
requirement failed: Failed to start ps scheduler
URL: https://github.com/apache/incubator-mxnet/issues/11249
This is an automated message from the Apach
marcoabreu commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397762309
Sorry if I mixed it up. I tried to read the description, but it was not very
helpful.
tqchen commented on issue #11311: Revert "Static memory allocation for
cached_op (#10817)"
URL: https://github.com/apache/incubator-mxnet/pull/11311#issuecomment-397759632
with great power comes great responsibility. I would never abused that power
to any changes just to make a point and b
piiswrong commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397758785
This is a different PR. If you have any specific concerns on the code please
utter them.
If you don't understand the code please let othe
piiswrong commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397758785
This is a different PR. If you have any legitimate concerns on the code
please specify them.
If you don't understand the code please let
tqchen commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397758362
regression tests are already introduced in
https://github.com/apache/incubator-mxnet/pull/11313/files#diff-962bd5bb7248659d7eb3be37ee8a4c6bR1288
-
tqchen commented on issue #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313#issuecomment-397758362
regression tests are already introduced in
https://github.com/apache/incubator-mxnet/pull/11313/files#diff-962bd5bb7248659d7eb3be37ee8a4c6bR1288
-
eric-haibin-lin commented on issue #11316: MXNET_FORCE_ADDTAKEGRAD to disable
AddTakeGradLargeBatchCaller
URL: https://github.com/apache/incubator-mxnet/pull/11316#issuecomment-397758053
Excellent finding!
Worse case we can modify the sparse kernel a bit (just remove the lookup
table)
mklissa commented on issue #11296: Adding 2 tutorials on Learning Rate Schedules
URL: https://github.com/apache/incubator-mxnet/pull/11296#issuecomment-397757892
Very informative and well-written! Good job!
This is an automate
marcoabreu commented on issue #11311: Revert "Static memory allocation for
cached_op (#10817)"
URL: https://github.com/apache/incubator-mxnet/pull/11311#issuecomment-397757505
I'd like to introduce you to the Apache guidelines:
https://www.apache.org/foundation/voting.html
"2. Code
This is an automated email from the ASF dual-hosted git repository.
zhasheng pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new c84786c Bump the publish ti
leezu opened a new pull request #11316: MXNET_FORCE_ADDTAKEGRAD to disable
AddTakeGradLargeBatchCaller
URL: https://github.com/apache/incubator-mxnet/pull/11316
Adds an environment variable that allows working around #11314 until a
proper fix is found.
nswamy opened a new issue #11315: Support multiple input types in Scala
URL: https://github.com/apache/incubator-mxnet/issues/11315
Currently MXNet-Scala APIs support only float32, this would cause accuracy
issues when customers port models between developed on Python to Scala. This
needs
leezu opened a new issue #11314: Embedding Backward
(AddTakeGradLargeBatchCaller) non-deterministic nan values
URL: https://github.com/apache/incubator-mxnet/issues/11314
## Description
The `AddTakeGradLargeBatchCaller` operator called during backward of
Embedding is broken and resu
tqchen commented on issue #11310: Update cached_op.cc
URL: https://github.com/apache/incubator-mxnet/pull/11310#issuecomment-397755433
replace by #10817
This is an automated message from the Apache Git Service.
To respond to
tqchen closed pull request #11310: Update cached_op.cc
URL: https://github.com/apache/incubator-mxnet/pull/11310
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and us
lanking520 commented on a change in pull request #11294: [MXNET-543][WIP]Scala
Spark test
URL: https://github.com/apache/incubator-mxnet/pull/11294#discussion_r195873690
##
File path:
scala-package/spark/src/test/scala/org/apache/mxnet/spark/MXNetGeneralSuite.scala
##
@@
lanking520 commented on a change in pull request #11294: [MXNET-543][WIP]Scala
Spark test
URL: https://github.com/apache/incubator-mxnet/pull/11294#discussion_r195873656
##
File path:
scala-package/spark/src/test/scala/org/apache/mxnet/spark/MXNetGeneralSuite.scala
##
@@
piiswrong opened a new pull request #11313: Static alloc for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/11313
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items f
piiswrong opened a new pull request #11310: Update cached_op.cc
URL: https://github.com/apache/incubator-mxnet/pull/11310
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items for your P
tqchen commented on issue #11310: Update cached_op.cc
URL: https://github.com/apache/incubator-mxnet/pull/11310#issuecomment-397755065
please bring in the new changes, and rebase against the master
This is an automated message
1 - 100 of 236 matches
Mail list logo