This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a commit to branch v1.6.x
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/v1.6.x by this push:
new e147c94 [Backport][v1.6.x] Fix the wr
This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a commit to branch v1.6.x
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/v1.6.x by this push:
new e147c94 [Backport][v1.6.x] Fix the wr
This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a commit to branch v1.6.x
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/v1.6.x by this push:
new 8b71f61 Initial checkin (#16856) (#16
ptrendx merged pull request #16872: Backport #16856 to 1.6
URL: https://github.com/apache/incubator-mxnet/pull/16872
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
ptrendx merged pull request #16884: [Backport][v1.6.x] Fix the wrong result of
sum, mean, argmin, argmax when inputs contain inf or nan
URL: https://github.com/apache/incubator-mxnet/pull/16884
This is an automated message f
This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a change to branch v1.6.x
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from e73c186 Backport #16798, #16836 and #16838 to 1.6 (#16874)
add 7784ac4 Add support for boolean input
ptrendx merged pull request #16892: Backport #16796 to 1.6
URL: https://github.com/apache/incubator-mxnet/pull/16892
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
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 dd65a23 Bump the publis
eric-haibin-lin commented on a change in pull request #16885: [WIP]Multi
Precision Lamb Update operator
URL: https://github.com/apache/incubator-mxnet/pull/16885#discussion_r349858159
##
File path: python/mxnet/optimizer/optimizer.py
##
@@ -1244,6 +1245,89 @@ def update(se
TaoLv commented on issue #16891: Upgrading MKLDNN to 1.0 causes performance
regression.
URL:
https://github.com/apache/incubator-mxnet/issues/16891#issuecomment-557760466
@leleamol How did you install the mxnet package, from source code or the
nightly build? If build from source code, cou
TaoLv commented on issue #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890#issuecomment-557759661
@ptrendx No, it should not affect 1.6.0 release.
This is an automated message fr
rongzha1 commented on issue #16845: MXNet 1.6.0 performance regression
URL:
https://github.com/apache/incubator-mxnet/issues/16845#issuecomment-557757080
seems no regression issue on mkl dnn CPU platform.
test platform: skx-8180
mxnet version: v1.5.x and v1.6.x
v1.5.x build cmd: m
MoisesHer opened a new pull request #16893: Add multi-tensor lamb Op
URL: https://github.com/apache/incubator-mxnet/pull/16893
Multi-tensor LAMB Optimizer OP **(in development / debugging)**
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items for you
ptrendx opened a new pull request #16892: Backport #16796 to 1.6
URL: https://github.com/apache/incubator-mxnet/pull/16892
## Description ##
Backport #16796 to 1.6
This is an automated message from the Apache Git Servic
This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 4da14a2 add op bitwise_or [numpy] (#16801)
add e3a7141 Add support for boolean inputs to FusedOp (#1
ptrendx closed issue #16723: [Bug] fused_op does not support boolean type
URL: https://github.com/apache/incubator-mxnet/issues/16723
This is an automated message from the Apache Git Service.
To respond to the message, please
ptrendx merged pull request #16796: Add support for boolean inputs to FusedOp
URL: https://github.com/apache/incubator-mxnet/pull/16796
This is an automated message from the Apache Git Service.
To respond to the message, plea
samskalicky commented on issue #16891: Upgrading MKLDNN to 1.0 causes
performance regression.
URL:
https://github.com/apache/incubator-mxnet/issues/16891#issuecomment-557742154
@mxnet-label-bot add [R1.6.0]
This is an automa
wkcn commented on issue #16884: [Backport][v1.6.x] Fix the wrong result of sum,
mean, argmin, argmax when inputs contain inf or nan
URL: https://github.com/apache/incubator-mxnet/pull/16884#issuecomment-557741981
Hi @ptrendx , I have deleted these modifications : )
wkcn edited a comment on issue #16881: Add TypeFlag=>string macro
URL: https://github.com/apache/incubator-mxnet/pull/16881#issuecomment-557741215
Hi @Kh4L,
mshadow is not archived. It has been a part of MXNet, and we can modify it
in MXNet repository.
https://github.com/apache/incuba
wkcn commented on issue #16881: Add TypeFlag=>string macro
URL: https://github.com/apache/incubator-mxnet/pull/16881#issuecomment-557741215
Hi @Kh4L,
mshadow is not archived. It has been a part of MXNet, and we can modify it.
https://github.com/apache/incubator-mxnet/pull/15600
--
This is an automated email from the ASF dual-hosted git repository.
apeforest pushed a change to branch benchmark
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 169ed69 Speed fused_op compilation by caching ptx and jit-compiled
functions (#16783)
add a47b5
This is an automated email from the ASF dual-hosted git repository.
apeforest pushed a change to branch benchmark
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 169ed69 Speed fused_op compilation by caching ptx and jit-compiled
functions (#16783)
add a47b5
ptrendx commented on issue #16875: Flaky: test_fusion.test_fusion
URL:
https://github.com/apache/incubator-mxnet/issues/16875#issuecomment-557722661
I ran that part of this test 1 times after the change without problem.
Closing.
ptrendx closed issue #16875: Flaky: test_fusion.test_fusion
URL: https://github.com/apache/incubator-mxnet/issues/16875
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Git
ptrendx commented on issue #16872: Backport #16856 to 1.6
URL: https://github.com/apache/incubator-mxnet/pull/16872#issuecomment-557719170
There was an issue with the CI, I restarted the stuck pipelines.
This is an automated m
leleamol commented on issue #16891: Upgrading MKLDNN to 1.0 causes performance
regression.
URL:
https://github.com/apache/incubator-mxnet/issues/16891#issuecomment-557716942
#16555
@TaoLv @pengzhao-intel @zixuanweeei @samskalicky
--
nikudyshko edited a comment on issue #16741: Error detecting C++11 and C++14
URL:
https://github.com/apache/incubator-mxnet/issues/16741#issuecomment-557715315
@larroy For Windows we still have MinGW, clang, etc. And it's better to
leave testing for them as it is. As for VS - hard to tell.
nikudyshko commented on issue #16741: Error detecting C++11 and C++14
URL:
https://github.com/apache/incubator-mxnet/issues/16741#issuecomment-557715315
For Windows we still have MinGW, clang, etc. And it's better to leave
testing for them as it is. As for VS - hard to tell. Versions befor
haojin2 commented on issue #16238: [Flaky] test_convolution_multiple_streams
URL:
https://github.com/apache/incubator-mxnet/issues/16238#issuecomment-557707378
Happening again:
http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Funix-gpu/detail/PR-16883/1/pi
sxjscience edited a comment on issue #16887: [Numpy] Bug of basic indexing
URL:
https://github.com/apache/incubator-mxnet/issues/16887#issuecomment-557701936
Another case fails after I changed
https://github.com/apache/incubator-mxnet/blob/4da14a22385622c35e9a5c9c3e8a17c07f718cad/python/mx
sxjscience commented on issue #16887: [Numpy] Bug of basic indexing
URL:
https://github.com/apache/incubator-mxnet/issues/16887#issuecomment-557701936
Another case:
```python
import mxnet as mx
from mxnet import gluon
mx.npx.set_np()
a = mx.np.random.normal(0, 1, (5, 5, 5
sxjscience commented on issue #16102: Usability degradation
URL:
https://github.com/apache/incubator-mxnet/issues/16102#issuecomment-557697428
@anirudh2290 Thanks... I'm considering to try it on my personal mac first to
understand how many tests will fail...
--
reminisce commented on issue #16889: Segmentation Fault in module mxnet.np with
missmatch dimensions
URL:
https://github.com/apache/incubator-mxnet/issues/16889#issuecomment-557697497
@GPoloVera D2L will depend on MXNet v1.6.0 eventually. I encourage you to
always use the latest nightly w
anirudh2290 edited a comment on issue #16102: Usability degradation
URL:
https://github.com/apache/incubator-mxnet/issues/16102#issuecomment-557695542
@sxjscience I have never built and run MXNet on macOS ! Its a good thought
to add Macos to CI, though I think its a project in itself since
anirudh2290 commented on issue #16102: Usability degradation
URL:
https://github.com/apache/incubator-mxnet/issues/16102#issuecomment-557695542
@sxjscience I have never run MXNet on macOS ! Its a good thought to add
Macos to CI, though I think its a project in itself since there may be oth
haojin2 edited a comment on issue #16875: Flaky: test_fusion.test_fusion
URL:
https://github.com/apache/incubator-mxnet/issues/16875#issuecomment-557695353
@ptrendx If you have verified that this change could make this test pass
more than 500 trials then I'm for it.
Note: A test that su
haojin2 commented on issue #16875: Flaky: test_fusion.test_fusion
URL:
https://github.com/apache/incubator-mxnet/issues/16875#issuecomment-557695353
@ptrendx If you have verified that this change could make this test pass
more than 500 trials then I'm for it.
Note: A test that only succ
sxjscience commented on issue #16796: Add support for boolean inputs to FusedOp
URL: https://github.com/apache/incubator-mxnet/pull/16796#issuecomment-557693123
@ptrendx Thanks for the fix. I've confirmed that it solves my previous
failure.
-
sxjscience commented on issue #16102: Usability degradation
URL:
https://github.com/apache/incubator-mxnet/issues/16102#issuecomment-557691997
@anirudh2290 In terms of the mac OS CI, do you have any idea about that?
@zhanghang1989 has observed some other problems in MXNet in macOS. I guess
leleamol opened a new issue #16891: Upgrading MKLDNN to 1.0 causes performance
regression.
URL: https://github.com/apache/incubator-mxnet/issues/16891
## Description
The change that upgraded MKLDNN to 1.0 caused performance (images/sec) to
drop by 200 points.
### Error Message
stu1130 commented on issue #16872: Backport #16856 to 1.6
URL: https://github.com/apache/incubator-mxnet/pull/16872#issuecomment-557664219
@ptrendx done
This is an automated message from the Apache Git Service.
To respond to t
ptrendx commented on issue #16824: Enable unit tests for TVM ops for all cuda
compute capabilities
URL: https://github.com/apache/incubator-mxnet/pull/16824#issuecomment-557659818
Ok, if it is expected onmaster and does not affect 1.6 release then that's
fine.
access2rohit commented on issue #16715: Lamb optimizer update
URL: https://github.com/apache/incubator-mxnet/pull/16715#issuecomment-557651851
> @access2rohit could you answer Sam's comments?
Done
This is an automated m
reminisce commented on issue #16824: Enable unit tests for TVM ops for all cuda
compute capabilities
URL: https://github.com/apache/incubator-mxnet/pull/16824#issuecomment-557651409
@ptrendx It fails because of the following line where the `<` operator
targeted for `sm_70` is invoked on a
access2rohit commented on a change in pull request #16715: Lamb optimizer update
URL: https://github.com/apache/incubator-mxnet/pull/16715#discussion_r349745221
##
File path: python/mxnet/optimizer/optimizer.py
##
@@ -1244,6 +1244,54 @@ def update(self, index, weight, grad,
access2rohit commented on a change in pull request #16715: Lamb optimizer update
URL: https://github.com/apache/incubator-mxnet/pull/16715#discussion_r349744815
##
File path: src/operator/optimizer_op-inl.h
##
@@ -1563,6 +1563,192 @@ inline void AdamUpdateEx(const nnvm::Nod
access2rohit commented on a change in pull request #16715: Lamb optimizer update
URL: https://github.com/apache/incubator-mxnet/pull/16715#discussion_r349744737
##
File path: src/operator/optimizer_op-inl.h
##
@@ -1563,6 +1563,192 @@ inline void AdamUpdateEx(const nnvm::Nod
ptrendx commented on issue #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890#issuecomment-557644707
@TaoLv @PatricZhao Does this impact 1.6?
This is an automated message from the
ptrendx commented on issue #16848: Flaky test_np_mixed_precision_binary_funcs
URL:
https://github.com/apache/incubator-mxnet/issues/16848#issuecomment-557640633
The PR was merged and backported to v1.6.x branch, can we close it @leezu
@haojin2?
ptrendx commented on issue #16875: Flaky: test_fusion.test_fusion
URL:
https://github.com/apache/incubator-mxnet/issues/16875#issuecomment-557640003
@haojin2 Do you agree that this (changing `a/b` to `a/(b+1)`) fixes the
flakiness and we can close this issue?
-
sxjscience commented on issue #16796: Add support for boolean inputs to FusedOp
URL: https://github.com/apache/incubator-mxnet/pull/16796#issuecomment-557636893
@ptrendx I'm testing it locally
This is an automated message from
ptrendx commented on issue #16796: Add support for boolean inputs to FusedOp
URL: https://github.com/apache/incubator-mxnet/pull/16796#issuecomment-557634856
@sxjscience Could you validate the fix in your original code?
This i
ptrendx commented on issue #16884: [Backport][v1.6.x] Fix the wrong result of
sum, mean, argmin, argmax when inputs contain inf or nan
URL: https://github.com/apache/incubator-mxnet/pull/16884#issuecomment-557633922
@wkcn is that class part of 1.6? If not, then I agree you should delete
th
szha closed issue #15749: [RFC] RFC Issue Mirroring to d...@mxnet.apache.org
URL: https://github.com/apache/incubator-mxnet/issues/15749
This is an automated message from the Apache Git Service.
To respond to the message, ple
szha commented on issue #15749: [RFC] RFC Issue Mirroring to
d...@mxnet.apache.org
URL:
https://github.com/apache/incubator-mxnet/issues/15749#issuecomment-557633070
RFC issue mirroring has been up and running. The mirroring on the email side
requires replying all to the thread on dev@.
ptrendx commented on issue #16824: Enable unit tests for TVM ops for all cuda
compute capabilities
URL: https://github.com/apache/incubator-mxnet/pull/16824#issuecomment-557617785
Hmmm, so it looks like a slightly different error with similar result - I
did not do anything with fp16 - this
Kh4L commented on a change in pull request #16881: Add TypeFlag=>string macro
URL: https://github.com/apache/incubator-mxnet/pull/16881#discussion_r349531359
##
File path: include/mxnet/base.h
##
@@ -85,6 +85,18 @@
*/
#define PROFILER_MESSAGE_FUNCNAME (__FUNCTION__)
+/
Kh4L commented on issue #16881: Add TypeFlag=>string macro
URL: https://github.com/apache/incubator-mxnet/pull/16881#issuecomment-557469070
> I prefer to add the type name in DataType class, and get the type name
from ‘mshadow::DataType::kName’.
>
https://github.com/apache/incubator-mxn
TaoLv commented on issue #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890#issuecomment-557458688
`FullyConnectedGradComputeExCPU` and `BackwardFCStorageType` should be
changed accordingly.
TaoLv commented on issue #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890#issuecomment-557458265
To be more accurate, MKL-DNN FC backward path was not enabled before, so
there is no bug of it in previously MXNet versions. This PR tries to ena
xinyu-intel commented on issue #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890#issuecomment-557457865
enable mkldnn inner product backward or this may not take effect.
This is
GPoloVera commented on issue #16889: Segmentation Fault in module mxnet.np with
missmatch dimensions
URL:
https://github.com/apache/incubator-mxnet/issues/16889#issuecomment-557448909
@reminisce, that solved the problem, thank you very much!!
The thing is the version is still the 1.
GPoloVera closed issue #16889: Segmentation Fault in module mxnet.np with
missmatch dimensions
URL: https://github.com/apache/incubator-mxnet/issues/16889
This is an automated message from the Apache Git Service.
To respond
reminisce commented on issue #16889: Segmentation Fault in module mxnet.np with
missmatch dimensions
URL:
https://github.com/apache/incubator-mxnet/issues/16889#issuecomment-557445450
Can you try `pip install mxnet --pre` to install the latest wheel? It should
have been fixed after 9/15.
leezu commented on a change in pull request #16888: Add evaluation_loss to the
estimator base class.
URL: https://github.com/apache/incubator-mxnet/pull/16888#discussion_r349483497
##
File path: python/mxnet/gluon/contrib/estimator/estimator.py
##
@@ -64,13 +64,17 @@ def _
rongzha1 opened a new pull request #16890: Mkldnn fullyConnect bwd bug fix
URL: https://github.com/apache/incubator-mxnet/pull/16890
## Description ##
fix mkldnn fc bwd bug due to data inplace
when data inplace, if do bwd_data will change src data which will lead
bwd_weight calc error
reminisce edited a comment on issue #16824: Enable unit tests for TVM ops for
all cuda compute capabilities
URL: https://github.com/apache/incubator-mxnet/pull/16824#issuecomment-557414420
@ptrendx This does not affect 1.6. We plan not to release TVM powered
operators in 1.6. If you see th
reminisce commented on issue #16887: [Numpy] Bug of basic indexing
URL:
https://github.com/apache/incubator-mxnet/issues/16887#issuecomment-557442363
`NDArray._basic_indexing_contiguous_flat_begin_end` has a bug. I will send
in a fix.
--
sxjscience commented on issue #16889: Segmentation Fault in module mxnet.np
with missmatch dimensions
URL:
https://github.com/apache/incubator-mxnet/issues/16889#issuecomment-557438331
@GPoloVera I cannot reproduce it with the master version. Actually, I've
checked the exception here in t
hgt312 commented on issue #16870: [numpy] add logical op
URL: https://github.com/apache/incubator-mxnet/pull/16870#issuecomment-557437272
Are the dispatch tests migrated from official numpy?
This is an automated message from t
GPoloVera opened a new issue #16889: Segmentation Fault in module mxnet.np with
missmatch dimensions
URL: https://github.com/apache/incubator-mxnet/issues/16889
## Description
When using the **module np from mxnet**, any operation that in the numpy
package raises an exception "Value Err
72 matches
Mail list logo
