ashokei opened a new pull request #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670
## Description ##
mkl-dnn operators only get triggered for certain tensor shapes, this PR adds
mkldnn specific unittest, and enables jenkins for it. This
ashokei closed pull request #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670
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
ashokei commented on issue #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670#issuecomment-383826767
@marcoabreu you already accepted these changes under #10021 , i just moved
unittest and jenkins specific stuff into separate PR, as that PR
ashokei commented on issue #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670#issuecomment-383826633
@marcoabreu you already accepted these changes under
https://github.com/apache/incubator-mxnet/pull/10021 , i just moved unittest
and jenki
ashokei opened a new pull request #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670
## Description ##
mkl-dnn operators only get triggered for certain tensor shapes, this PR adds
mkldnn specific unittest, and enables jenkins for it. This
ashokei commented on issue #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670#issuecomment-383826633
@marcoabreu you already accepted these changes under
https://github.com/apache/incubator-mxnet/pull/10021 , i just moved unittest
and jenki
anirudh2290 commented on a change in pull request #10666: [MXNET-359] fix
checks on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183628689
##
File path: src/operator/nn/mkldnn/mkldnn_deconvolution.cc
##
@@ -32,6 +3
anirudh2290 commented on a change in pull request #10666: [MXNET-359] fix
checks on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183628039
##
File path: src/operator/nn/mkldnn/mkldnn_convolution.cc
##
@@ -123,16 +1
anirudh2290 commented on a change in pull request #10666: [MXNET-359] fix
checks on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183628192
##
File path: src/operator/nn/mkldnn/mkldnn_deconvolution.cc
##
@@ -67,31 +
anirudh2290 commented on a change in pull request #10666: [MXNET-359] fix
checks on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183627992
##
File path: src/operator/nn/mkldnn/mkldnn_convolution.cc
##
@@ -88,26 +91
BUG1989 commented on issue #9552: [MXNET-133] Model Quantization with
Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#issuecomment-383833595
Thank you very much for sharing the Int8 quantize implement.Surely choose
the current calibration dataset is very important,in o
zheng-da commented on a change in pull request #10666: [MXNET-359] fix checks
on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183631914
##
File path: src/operator/nn/mkldnn/mkldnn_convolution.cc
##
@@ -88,26 +91,23
zheng-da commented on a change in pull request #10666: [MXNET-359] fix checks
on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183632343
##
File path: src/operator/nn/mkldnn/mkldnn_deconvolution.cc
##
@@ -67,31 +73,
zheng-da commented on a change in pull request #10666: [MXNET-359] fix checks
on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183632594
##
File path: src/operator/nn/mkldnn/mkldnn_deconvolution.cc
##
@@ -32,6 +32,1
zheng-da commented on issue #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670#issuecomment-383836613
@ashokei maybe you don't need to move the unit test, which is specific to
your fix.
I was thinking of your modification in Jenkins. If y
rasefon opened a new pull request #10671: attrs key should be 'attrs' not 'attr'
URL: https://github.com/apache/incubator-mxnet/pull/10671
## Description ##
Visualization in scala package doesn't work correctly because of the key of
attrs is wrong, it should be 'attrs'.
## Checkli
anirudh2290 commented on issue #9920: Asynchronous Issue on CustomOP and
mxnet.image.ImageDetIter
URL:
https://github.com/apache/incubator-mxnet/issues/9920#issuecomment-383856962
@IrvingShu I am currently working on fixing this. Please see:
https://github.com/anirudh2290/mxnet/blob/fix_g
xylcbd commented on issue #7551: How to get the symbol of gluon's RNN model by
hybridizing
URL:
https://github.com/apache/incubator-mxnet/issues/7551#issuecomment-383862174
same here.
This is an automated message from the A
stupidZZ opened a new issue #10672: Bug at L304: Gluon/block.py, v1.2.0
URL: https://github.com/apache/incubator-mxnet/issues/10672
`def _collect_params_with_prefix(self, prefix=''):
if prefix:
prefix += '.'
ret = {prefix + key : val for key, val in s
zfs1212 commented on issue #3739: Can't use mxnet after Quick Installation
URL:
https://github.com/apache/incubator-mxnet/issues/3739#issuecomment-383902694
Why is there no make folder in my mxnet directory?
This is an automa
zfs1212 commented on issue #3739: Can't use mxnet after Quick Installation
URL:
https://github.com/apache/incubator-mxnet/issues/3739#issuecomment-383902694
I have this problem too.But why is there no make folder in my mxnet
directory?
-
zfs1212 commented on issue #3739: Can't use mxnet after Quick Installation
URL:
https://github.com/apache/incubator-mxnet/issues/3739#issuecomment-383902694
I have this problem too.But why is there no make folder in my mxnet
directory?
-
ZiyueHuang commented on a change in pull request #10649: [WIP] add
FResourceRequestEx interface
URL: https://github.com/apache/incubator-mxnet/pull/10649#discussion_r183704642
##
File path: src/executor/attach_op_resource_pass.cc
##
@@ -47,8 +50,13 @@ Graph AttachOpResourc
juliusshufan commented on issue #10646: Correctly set the device context to CPU
at symbolic mode when --gpus not set
URL: https://github.com/apache/incubator-mxnet/pull/10646#issuecomment-383924001
Hi, looks like the failed case breaking the pre-ci is irrelevant to my
commit.
Thanks
jaheba opened a new issue #10673: Segmentation Fault when using large
mx.nd.array
URL: https://github.com/apache/incubator-mxnet/issues/10673
## Description
A segmentation fault can occur on some platforms when using large
`mx.nd.array`s. The same code runs fine using `numpy.array`.
reminisce commented on issue #10673: Segmentation Fault when using large
mx.nd.array
URL:
https://github.com/apache/incubator-mxnet/issues/10673#issuecomment-383945045
If I remember correctly, advanced indexing does not support negative
indices, but should have printed clearer error messa
jaheba commented on issue #10673: Segmentation Fault when using large
mx.nd.array
URL:
https://github.com/apache/incubator-mxnet/issues/10673#issuecomment-383952596
Removing the negative index does indeed prevent the segfault.
Do you know why it is not failing on mac os, though?
jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r183753288
##
File path: src/operator/quantization/quantize_graph_pass.cc
##
@@ -198,7 +198,7 @@
reminisce commented on issue #10673: Segmentation Fault when using large
mx.nd.array
URL:
https://github.com/apache/incubator-mxnet/issues/10673#issuecomment-383955099
segfault happens when memory access violates some os rules. This can be
platform and program dependent.
Your worka
PKurnikov opened a new issue #10674: How to resize mx.NDArray 2d tensor with
interp = 'Nearest'
URL: https://github.com/apache/incubator-mxnet/issues/10674
I need function like mx.nd.SpatialTransformer or BilinearResize2D, but
interp parametr is {'bilinear'}, required.
---
jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r183758965
##
File path: include/mxnet/c_api.h
##
@@ -1423,13 +1423,15 @@ MXNET_DLL int MXSymbol
jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r183759204
##
File path: src/operator/quantization/dequantize.cc
##
@@ -23,11 +23,32 @@
* \bri
jaheba commented on issue #10673: Segmentation Fault when using large
mx.nd.array
URL:
https://github.com/apache/incubator-mxnet/issues/10673#issuecomment-383962530
Thanks for your help. :)
I am not sure I can follow your efficiency argument however. Is `xs[(a,b)]`
less efficient
jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r183241662
##
File path: src/operator/quantization/quantize_graph_pass.cc
##
@@ -129,7 +129,7 @@
jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r183241106
##
File path: src/operator/quantization/mkldnn/mkldnn_quantize-inl.h
##
@@ -0,0 +1,10
piiswrong closed pull request #10671: attrs key should be 'attrs' not 'attr'
URL: https://github.com/apache/incubator-mxnet/pull/10671
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 f
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new b847c6f attrs key should be 'attrs' not
piiswrong closed pull request #10657: Fix typo in __setitem__ doc
URL: https://github.com/apache/incubator-mxnet/pull/10657
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
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new a8e1912 Fix typo in __setitem__ doc (#10
anirudh2290 commented on a change in pull request #10666: [MXNET-359] fix
checks on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r183821568
##
File path: src/operator/nn/mkldnn/mkldnn_convolution.cc
##
@@ -88,26 +91
piiswrong commented on issue #10354: Expose the number of GPUs.
URL: https://github.com/apache/incubator-mxnet/pull/10354#issuecomment-384019960
they are run with both. gpu/* tests are run with only gpu
This is an automated me
piiswrong closed pull request #10607: New tutorial on how to create a new
custom layer in Gluon
URL: https://github.com/apache/incubator-mxnet/pull/10607
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 provena
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new 8011f3b New tutorial on how to create a
piiswrong closed pull request #10021: [MXNET-33] SSD example not working with
mkl-dnn
URL: https://github.com/apache/incubator-mxnet/pull/10021
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 t
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new fd55a2e [MXNET-33] SSD example not worki
piiswrong commented on a change in pull request #10461: allow user to define
unknown token symbol
URL: https://github.com/apache/incubator-mxnet/pull/10461#discussion_r183823838
##
File path: python/mxnet/rnn/io.py
##
@@ -58,6 +62,8 @@ def encode_sentences(sentences, vocab
ashokei closed pull request #10670: enable jenkins for mkldnn unittest
URL: https://github.com/apache/incubator-mxnet/pull/10670
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
piiswrong commented on a change in pull request #10633: [MXNET-346] Hard
Sigmoid Operator
URL: https://github.com/apache/incubator-mxnet/pull/10633#discussion_r183824195
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -388,6 +388,94 @@ void CastCompute(const nn
haojin2 commented on a change in pull request #10633: [MXNET-346] Hard Sigmoid
Operator
URL: https://github.com/apache/incubator-mxnet/pull/10633#discussion_r183825064
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -388,6 +388,94 @@ void CastCompute(const nnvm
piiswrong commented on a change in pull request #10633: [MXNET-346] Hard
Sigmoid Operator
URL: https://github.com/apache/incubator-mxnet/pull/10633#discussion_r183826137
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -388,6 +388,94 @@ void CastCompute(const nn
cells2numbers opened a new issue #10675: R Installation on MacOS fails (file
not found)
URL: https://github.com/apache/incubator-mxnet/issues/10675
The installation of mxnet fails on MacOS using RStudio because the file
mxnet_0.10.1.tar.gz is not found. Probably the file is outdated. Could
anirudh2290 commented on issue #10661: mark MKLDNN experimantal.
URL: https://github.com/apache/incubator-mxnet/pull/10661#issuecomment-384027516
@zheng-da can you also add the same line at the end of ONNX description. cc
@anirudhacharya
---
anirudhacharya opened a new pull request #10676: fix for asymmetric padding
URL: https://github.com/apache/incubator-mxnet/pull/10676
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable item
anirudh2290 commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183832404
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,10 +214,14 @@ def conv(attrs,
anirudh2290 commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183832591
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,10 +214,14 @@ def conv(attrs,
haojin2 commented on a change in pull request #10633: [MXNET-346] Hard Sigmoid
Operator
URL: https://github.com/apache/incubator-mxnet/pull/10633#discussion_r183846508
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -388,6 +388,94 @@ void CastCompute(const nnvm
zheng-da commented on issue #10666: [MXNET-359] fix checks on convolution
parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#issuecomment-384045260
@piiswrong @eric-haibin-lin @ashokei @pengzhao-intel @TaoLv
Could you help review the code and merge it, so it
spidyDev commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183849902
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,38 @@ def conv(attrs, inpu
spidyDev commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183849902
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,38 @@ def conv(attrs, inpu
spidyDev commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183850165
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -229,9 +258,38 @@ def deconv(attrs, in
spidyDev commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183850413
##
File path: tests/python-pytest/onnx/onnx_test.py
##
@@ -254,4 +255,4 @@ def test_bvlc_rcnn_ilsvrc13():
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183850592
##
File path: tests/python-pytest/onnx/onnx_test.py
##
@@ -254,4 +255,4 @@ def test_bvlc_rcnn_ilsvrc13():
szha commented on issue #10630: make the slice of sequential a sequential
URL: https://github.com/apache/incubator-mxnet/pull/10630#issuecomment-384050775
@piiswrong
This is an automated message from the Apache Git Service.
T
anirudhacharya opened a new pull request #10677: Mark ONNX-MXNet experimental
URL: https://github.com/apache/incubator-mxnet/pull/10677
## Description ##
mark onnx-mxnet as experimental in release notes.
## Checklist ##
### Essentials ###
Please feel free to remove inapplicab
haojin2 commented on a change in pull request #10633: [MXNET-346] Hard Sigmoid
Operator
URL: https://github.com/apache/incubator-mxnet/pull/10633#discussion_r183846508
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -388,6 +388,94 @@ void CastCompute(const nnvm
anirudhacharya closed pull request #10662: [MXNET-347] Logical AND operator
URL: https://github.com/apache/incubator-mxnet/pull/10662
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 fo
zheng-da commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183858705
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,25 @@ def conv(attrs, inpu
zheng-da commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183860494
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -229,9 +245,25 @@ def deconv(attrs, in
zheng-da commented on issue #10676: fix for asymmetric padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#issuecomment-384060380
Is there a unit test to cover this?
This is an automated message from the Apache G
anirudh2290 commented on issue #10661: mark MKLDNN experimantal.
URL: https://github.com/apache/incubator-mxnet/pull/10661#issuecomment-384067467
ONNX in release notes is updated as part of #10677
This is an automated message
anirudh2290 commented on a change in pull request #10677: Mark ONNX-MXNet
experimental
URL: https://github.com/apache/incubator-mxnet/pull/10677#discussion_r183868344
##
File path: NEWS.md
##
@@ -5,7 +5,7 @@ MXNet Change Log
- Implemented new [Scala Inference
APIs](https
zheng-da opened a new pull request #10678: invalidate output arrays for
imperative.
URL: https://github.com/apache/incubator-mxnet/pull/10678
## Description ##
we need to invalidate the MKLDNN arrays when they are reused as output
arrays. We have done it in the symbolic executor. We nee
anirudh2290 commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183877365
##
File path: tests/python-pytest/onnx/onnx_test.py
##
@@ -124,12 +124,13 @@ def test_super_resolution_examp
anirudh2290 commented on issue #10566: [MXNET-324] Use default nvcc when
configured nvcc not present.
URL: https://github.com/apache/incubator-mxnet/pull/10566#issuecomment-384080239
@piiswrong @marcoabreu can this be merged ?
---
anirudh2290 commented on issue #10569: [MXNET-324] Use default nvcc when
configured nvcc not present.
URL: https://github.com/apache/incubator-mxnet/pull/10569#issuecomment-384080306
@piiswrong @marcoabreu can this be merged ?
---
eric-haibin-lin commented on issue #10666: [MXNET-359] fix checks on
convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#issuecomment-384083892
Do we still want this change if onnx correctly handles padding?
-
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183883862
##
File path: tests/python-pytest/onnx/onnx_test.py
##
@@ -124,12 +124,13 @@ def test_super_resolution_ex
anirudh2290 commented on issue #10666: [MXNET-359] fix checks on convolution
parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#issuecomment-384085337
@eric-haibin-lin yes we still need this change to make the behavior
consistent (with or without MKLDNN) enable
yajiedesign commented on issue #10629: [MXNET-343]fix Mkldnn with msvc
URL: https://github.com/apache/incubator-mxnet/pull/10629#issuecomment-384090424
can we merge?
This is an automated message from the Apache Git Service.
To
ThomasDelteil commented on a change in pull request #10485: [MXNET-304][RFC]
Jenkins docs build
URL: https://github.com/apache/incubator-mxnet/pull/10485#discussion_r183890449
##
File path: ci/docker/install/ubuntu_scala.sh
##
@@ -1,4 +1,4 @@
-#!/bin/bash
+ #!/bin/bash
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183891749
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,25 @@ def conv(attrs
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183891749
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,25 @@ def conv(attrs
piiswrong closed pull request #10632: move parameters to backend for hybridblock
URL: https://github.com/apache/incubator-mxnet/pull/10632
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
This is an automated email from the ASF dual-hosted git repository.
jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new faffe56 move parameters to backend for h
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183895796
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -229,9 +245,25 @@ def deconv(att
anirudh2290 commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183895852
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,25 @@ def conv(attrs, i
anirudhacharya commented on issue #10676: fix for asymmetric padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#issuecomment-384096721
The CI failed in one of the flaky tests while trying to fetch test files -
ProtocolError: ("Connection broken: ConnectionResetError(104, 'C
anirudhacharya commented on a change in pull request #10676: fix for asymmetric
padding
URL: https://github.com/apache/incubator-mxnet/pull/10676#discussion_r183897755
##
File path: python/mxnet/contrib/onnx/_import/op_translations.py
##
@@ -214,9 +214,25 @@ def conv(attrs
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183910073
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -136,20 +9
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183910178
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -136,20 +9
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183909390
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -136,20 +9
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183909723
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -174,6 +18
haojin2 commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns
and dot(dns, csr.T) = dns on GPU
URL: https://github.com/apache/incubator-mxnet/pull/10371#issuecomment-384114091
Benchmark results: (density(%) speedup rate)
- dot(dense, csr):
- without fallback time:
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183908931
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -48,67 +49
yzhliu commented on a change in pull request #10660: [MXNET-357][WIP] New Scala
API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183910476
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -136,20 +9
haojin2 commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns
and dot(dns, csr.T) = dns on GPU
URL: https://github.com/apache/incubator-mxnet/pull/10371#issuecomment-384114091
Benchmark results: [density of rhs csr(%)] [speedup rate]
- dot(dense, csr):
- without fa
haojin2 commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns
and dot(dns, csr.T) = dns on GPU
URL: https://github.com/apache/incubator-mxnet/pull/10371#issuecomment-384114344
benchmark script:
`import mxnet as mx
import numpy as np
from mxnet.test_utils import ass
lanking520 commented on a change in pull request #10660: [MXNET-357][WIP] New
Scala API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183911000
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -136,2
haojin2 commented on issue #10371: [MXNET-263] Support for dot(dns, csr) = dns
and dot(dns, csr.T) = dns on GPU
URL: https://github.com/apache/incubator-mxnet/pull/10371#issuecomment-384114344
benchmark script:
`import mxnet as mx
import numpy as np
from mxnet.test_utils import ass
lanking520 commented on a change in pull request #10660: [MXNET-357][WIP] New
Scala API Design
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r183911074
##
File path:
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
##
@@ -48,67
1 - 100 of 139 matches
Mail list logo