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 a43fb91 Bump the publis
vexilligera commented on a change in pull request #16667: [NumPy] Add NumPy
support for Cholesky decomposition
URL: https://github.com/apache/incubator-mxnet/pull/16667#discussion_r340447515
##
File path: python/mxnet/symbol/numpy/linalg.py
##
@@ -128,3 +128,65 @@ def svd(
ZhennanQin commented on a change in pull request #16672: [mkldnn-v1.0]fallback
mkldnn fc bwd in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672#discussion_r340435359
##
File path: src/operator/nn/fully_connected.cc
##
@@ -147,7 +147,8 @@ void Full
rongzha1 commented on a change in pull request #16672: [mkldnn-v1.0]fallback
mkldnn fc bwd in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672#discussion_r340435205
##
File path: src/operator/nn/fully_connected.cc
##
@@ -147,7 +147,8 @@ void FullyC
pengzhao-intel commented on a change in pull request #16672:
[mkldnn-v1.0]fallback mkldnn fc bwd in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672#discussion_r340434779
##
File path: src/operator/nn/fully_connected.cc
##
@@ -147,7 +147,8 @@ void
pengzhao-intel commented on issue #16555: Upgrade MKL-DNN dependency to v1.0
URL: https://github.com/apache/incubator-mxnet/pull/16555#issuecomment-547742586
The only known issue is fixed by #16672
This is an automated me
ZhennanQin commented on a change in pull request #16672: [mkldnn-v1.0]fallback
mkldnn fc bwd in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672#discussion_r340433577
##
File path: src/operator/nn/fully_connected.cc
##
@@ -147,7 +147,8 @@ void Full
This is an automated email from the ASF dual-hosted git repository.
patriczhao pushed a change to branch mkldnn-v1.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from b5d4f87 Use memcopy instead of set_handle when num_layer=0,
direction=1 (#16663)
add e663ce1
pengzhao-intel merged pull request #16672: [mkldnn-v1.0]fallback mkldnn fc bwd
in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672
This is an automated message from the Apache Git Service.
To respond
hgt312 commented on issue #16660: [WIP] [Numpy] TVM implementation for binary
ops
URL: https://github.com/apache/incubator-mxnet/pull/16660#issuecomment-547738144
How about move op_maps to another file?
This is an automated m
xidulu commented on a change in pull request #16667: [NumPy] Add NumPy support
for Cholesky decomposition
URL: https://github.com/apache/incubator-mxnet/pull/16667#discussion_r340425892
##
File path: python/mxnet/symbol/numpy/linalg.py
##
@@ -128,3 +128,65 @@ def svd(a):
pengzhao-intel commented on a change in pull request #16672:
[mkldnn-v1.0]fallback mkldnn fc bwd in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672#discussion_r340425718
##
File path: src/operator/nn/fully_connected.cc
##
@@ -147,7 +147,8 @@ void
wkcn edited a comment on issue #16601: Where is this op.h file ?
URL:
https://github.com/apache/incubator-mxnet/issues/16601#issuecomment-547727342
@CoinCheung Yes. You can use `python OpWrapperGenerator.py ` to generate the op.h.
e.g. `python OpWrapperGenerator.py ../../build/libmxnet.
wkcn commented on issue #16601: Where is this op.h file ?
URL:
https://github.com/apache/incubator-mxnet/issues/16601#issuecomment-547727342
@CoinCheung Yes. You can use `python OpWrapperGenerator.py ` to generate the op.h.
T
rongzha1 opened a new pull request #16672: [mkldnn-v1.0]fallback mkldnn fc bwd
in imperative mode
URL: https://github.com/apache/incubator-mxnet/pull/16672
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to r
CoinCheung commented on issue #16601: Where is this op.h file ?
URL:
https://github.com/apache/incubator-mxnet/issues/16601#issuecomment-547715380
Thanks for replying.
Did I configure the building process correctly if the `libmxnet.so` is
generated in the `build/` folder ?
---
pengzhao-intel commented on issue #16555: Upgrade MKL-DNN dependency to v1.0
URL: https://github.com/apache/incubator-mxnet/pull/16555#issuecomment-547695006
@all, only one issue is left at now and we will fix it soon. I prefer to
merge this PR in one or two days so that we have buffer time
This is an automated email from the ASF dual-hosted git repository.
marcoabreu pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 5aa74e0 Move ops which don't support FP16 dtype to FP32 list (#16668)
add 8e50fd9 no such method
This is an automated email from the ASF dual-hosted git repository.
marcoabreu pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 5aa74e0 Move ops which don't support FP16 dtype to FP32 list (#16668)
add 8e50fd9 no such method
marcoabreu merged pull request #16610: no such method => modified function args
URL: https://github.com/apache/incubator-mxnet/pull/16610
This is an automated message from the Apache Git Service.
To respond to the message, pl
marcoabreu commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340390820
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,15 +1993,17 @@ def array(object, d
marcoabreu commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340390770
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,15 +1993,17 @@ def array(object, d
marcoabreu commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340390575
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,15 +1993,17 @@ def array(object, d
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 bd24c78 Bump the publis
ChaiBapchya closed issue #16352: [CI] Julia Windows CPU timeout
URL: https://github.com/apache/incubator-mxnet/issues/16352
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
ChaiBapchya commented on issue #16352: [CI] Julia Windows CPU timeout
URL:
https://github.com/apache/incubator-mxnet/issues/16352#issuecomment-547685853
Alright. Will close it then.
This is an automated message from the Apach
ChaiBapchya commented on issue #16603: Significant slowdown in some DGL models
URL:
https://github.com/apache/incubator-mxnet/issues/16603#issuecomment-547685567
Tried reproducing
Got a segfault
Build flags
```
[✔ CUDA, ✔ CUDNN, ✔ NCCL, ✔ CUDA_RTC, ✖ TENSORRT, ✔ CPU_SSE, ✔
anirudh2290 commented on a change in pull request #16654: [WIP] Multithreaded
Inference Support
URL: https://github.com/apache/incubator-mxnet/pull/16654#discussion_r340370719
##
File path: include/mxnet/c_api.h
##
@@ -1274,10 +1274,28 @@ MXNET_DLL int MXCreateCachedOpEx(S
zachgk commented on a change in pull request #16654: [WIP] Multithreaded
Inference Support
URL: https://github.com/apache/incubator-mxnet/pull/16654#discussion_r340365798
##
File path: include/mxnet/c_api.h
##
@@ -1274,10 +1274,28 @@ MXNET_DLL int MXCreateCachedOpEx(Symbol
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340365102
##
File path: tests/python/unittest/test_numpy_ndarray.py
##
@@ -483,24 +484,29 @@ def __init__(
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340361343
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -577,7 +577,7 @@ def __setitem__(self, key,
Laurawly commented on a change in pull request #16660: [WIP] [Numpy] TVM
implementation for binary ops
URL: https://github.com/apache/incubator-mxnet/pull/16660#discussion_r340226828
##
File path: contrib/tvmop/core/umath.py
##
@@ -120,3 +121,327 @@ def _compute_binary_sca
Laurawly commented on a change in pull request #16660: [WIP] [Numpy] TVM
implementation for binary ops
URL: https://github.com/apache/incubator-mxnet/pull/16660#discussion_r340226115
##
File path: contrib/tvmop/core/umath.py
##
@@ -120,3 +121,327 @@ def _compute_binary_sca
stu1130 commented on issue #16670: cuDNN RNN dtype_with_fallback_ calc needs
update
URL:
https://github.com/apache/incubator-mxnet/issues/16670#issuecomment-547627773
Thanks for the suggestion @DickJC123
I created a quick fix #16671
sshearn commented on issue #16620: Incompatible input shape
URL:
https://github.com/apache/incubator-mxnet/issues/16620#issuecomment-547627263
@lanking520 I tried using the C++ API as well and got:
`Check failed: MXSymbolInferShapeEx(GetHandle(), keys.size(), keys.data(),
arg_ind_pt
stu1130 opened a new pull request #16671: Fix cuDNN RNN dtype_with_fallback_ bug
URL: https://github.com/apache/incubator-mxnet/pull/16671
## Description ##
Fix #16670
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items for your PR.
- [ ] The P
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 60d74bc Showing proper error message when an attempt is made to
create large tensor but MXNet is not
anirudh2290 merged pull request #16668: Move ops which don't support FP16 dtype
to FP32 list
URL: https://github.com/apache/incubator-mxnet/pull/16668
This is an automated message from the Apache Git Service.
To respond to t
anirudh2290 commented on issue #16603: Significant slowdown in some DGL models
URL:
https://github.com/apache/incubator-mxnet/issues/16603#issuecomment-547614675
@ChaiBapchya it would be worth checking if this PR :
https://github.com/apache/incubator-mxnet/pull/16526 caused some slowdown.
ChaiBapchya commented on issue #16603: Significant slowdown in some DGL models
URL:
https://github.com/apache/incubator-mxnet/issues/16603#issuecomment-547613652
Waiting on @zheng-da for some details
@mxnet-label-bot add [Pending requester info]
-
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340286560
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dtyp
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340285907
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dtyp
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 00fb7ab Bump the publis
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340248087
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -451,7 +451,7 @@ void CastCompute(co
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340241395
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -577,7 +577,7 @@ def __setitem__(self, key,
haojin2 commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340241352
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dtyp
anirudh2290 commented on issue #16651: I'm sorry I've triggered an error in
mxnet source code, how can I debug it? It seems like a check failure on custom
operators, how can I find more details?
URL:
https://github.com/apache/incubator-mxnet/issues/16651#issuecomment-547554835
> Can y
DickJC123 opened a new issue #16670: cuDNN RNN dtype_with_fallback_ calc needs
update
URL: https://github.com/apache/incubator-mxnet/issues/16670
## Description
In src/operator/rnn-inl.h, the calculation of the mathPrec passed to
cudnnSetRNNDescriptor_v6() needs review (the variable is
reminisce commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340231639
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dt
reminisce commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340230096
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dt
reminisce commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340231639
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, dt
reminisce commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340235216
##
File path: src/operator/tensor/elemwise_unary_op.h
##
@@ -451,7 +451,7 @@ void CastCompute(
reminisce commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340228358
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -577,7 +577,7 @@ def __setitem__(self, ke
anirudh2290 commented on issue #16603: Significant slowdown in some DGL models
URL:
https://github.com/apache/incubator-mxnet/issues/16603#issuecomment-547550022
@access2rohit , @ChaiBapchya are you guys looking at this ?
Thi
anirudh2290 opened a new issue #16669: FP16 dtype support for
_contrib_MultiBox* ops
URL: https://github.com/apache/incubator-mxnet/issues/16669
## Description
Currently, these ops don't support FP16 dtype. Please see related PR :
https://github.com/apache/incubator-mxnet/pull/16668/fil
anirudh2290 edited a comment on issue #16668: Move ops which don't support FP16
dtype to FP32 list
URL: https://github.com/apache/incubator-mxnet/pull/16668#issuecomment-547544508
> LGTM. Why did you remove the BN v1 and Polling v1? Do they support fp16?
it was a mistake. removed it
anirudh2290 commented on issue #16668: Move ops which don't support FP16 dtype
to FP32 list
URL: https://github.com/apache/incubator-mxnet/pull/16668#issuecomment-547544508
> LGTM. Why did you remove the BN v1 and Polling v1? Do they support fp16?
it was a mistake. removed it now.
-
sxjscience commented on a change in pull request #16664: Miscellaneous fix for
several numpy issues
URL: https://github.com/apache/incubator-mxnet/pull/16664#discussion_r340225307
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1993,10 +1993,12 @@ def array(object, d
anirudh2290 commented on issue #16668: Move ops which don't support FP16 dtype
to FP32 list
URL: https://github.com/apache/incubator-mxnet/pull/16668#issuecomment-547539365
@ptrendx can you please take a look.
This is an auto
anirudh2290 opened a new pull request #16668: Move ops which don't support FP16
dtype to FP32 list
URL: https://github.com/apache/incubator-mxnet/pull/16668
## Description ##
Some legacy ops call this InferType function here:
https://github.com/apache/incubator-mxnet/blob/master/include
vexilligera opened a new pull request #16667: Add NumPy support for Cholesky
decomposition
URL: https://github.com/apache/incubator-mxnet/pull/16667
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to remove i
wkcn commented on issue #16557: Where is the place that mxnet call cblas_gemm
if I use openblas?
URL:
https://github.com/apache/incubator-mxnet/issues/16557#issuecomment-547426565
https://github.com/apache/incubator-mxnet/blob/60d74bc948869588c2f143fd3d55231859dc979f/src/operator/linalg_i
wkcn edited a comment on issue #16557: Where is the place that mxnet call
cblas_gemm if I use openblas?
URL:
https://github.com/apache/incubator-mxnet/issues/16557#issuecomment-547426565
Hi @knjwhn,
https://github.com/apache/incubator-mxnet/blob/60d74bc948869588c2f143fd3d55231859dc
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 c621d02 Bump the publis
knjwhn commented on issue #16557: Where is the place that mxnet call cblas_gemm
if I use openblas?
URL:
https://github.com/apache/incubator-mxnet/issues/16557#issuecomment-547397484
Thanks a lot , I modified the makefile and compile it successsfully. and I
run the example/image-classficat
pengzhao-intel merged pull request #16663: [mkldnn-v1.0] Use memcpy instead of
set_handle with unidirectional 1-layer RNN
URL: https://github.com/apache/incubator-mxnet/pull/16663
This is an automated message from the Apache
deHsien opened a new issue #1: Disable python logging verbose from C++
implementation
URL: https://github.com/apache/incubator-mxnet/issues/1
I don’t want to show the logging information, e.g.
```bash
src/engine/engine.cc:55: MXNet start using engine: NaiveEngine
```
Ho
zixuanweeei commented on issue #16663: [mkldnn-v1.0] Use memcpy instead of
set_handle with unidirectional 1-layer RNN
URL: https://github.com/apache/incubator-mxnet/pull/16663#issuecomment-547342969
`IMPERATIVE.FullyConnectedOp` failed again. @pengzhao-intel
--
liuzhi136 opened a new issue #16665: MapPlanKernel ErrStr:an illegal memory
access was encountered
URL: https://github.com/apache/incubator-mxnet/issues/16665
I got the same problem with
https://github.com/apache/incubator-mxnet/issues/11914, when I trying to do the
NMT task on my own mod
haojin2 opened a new pull request #16664: Miscellaneous fix for several numpy
issues
URL: https://github.com/apache/incubator-mxnet/pull/16664
## Description ##
#16658 #16650 and inconsistent type when an array is created from an
official numpy array:
```Python
from mxnet import n
ZHAIXINGZHAIYUE closed issue #16531: Just use fp16 is slower than
mixed_precision! Why?
URL: https://github.com/apache/incubator-mxnet/issues/16531
This is an automated message from the Apache Git Service.
To respond to the
ZHAIXINGZHAIYUE commented on issue #16531: Just use fp16 is slower than
mixed_precision! Why?
URL:
https://github.com/apache/incubator-mxnet/issues/16531#issuecomment-547326899
it is influenced by kvstore(local or device).
T
xiezhq-hermann commented on issue #16209: Complex data type support and numpy
operator fft
URL: https://github.com/apache/incubator-mxnet/pull/16209#issuecomment-547314236
doc of `numpy.fft` needed in `_numpy_op_doc`, the op `fft` will be moved to
`npi` scope later
--
reminisce commented on issue #16653: [Bug][Numpy] The data type of power is not
correct
URL:
https://github.com/apache/incubator-mxnet/issues/16653#issuecomment-547301386
This is a known issue of many operators from which floats are expected with
integral inputs. Instead of tweaking many
reminisce commented on issue #16658: [Numpy] Boolean array bug
URL:
https://github.com/apache/incubator-mxnet/issues/16658#issuecomment-547299345
In imperative mode, `astype` used `copy` operator which has fully supported
boolean ndarrays, while in symbolic mode, `astype` calls `cast`, whi
Heermosi edited a comment on issue #16651: I'm sorry I've triggered an error in
mxnet source code, how can I debug it? It seems like a check failure on custom
operators, how can I find more details?
URL:
https://github.com/apache/incubator-mxnet/issues/16651#issuecomment-547282626
> This
76 matches
Mail list logo