eric-haibin-lin commented on issue #8974: MXNET compatibility with MKL
libraries bundled in Microsoft R Open
URL:
https://github.com/apache/incubator-mxnet/issues/8974#issuecomment-350362565
@ykim362 @ashokei Can you provide some help on this issue?
cjolivier01 commented on a change in pull request #9002: Misc performance
changes
URL: https://github.com/apache/incubator-mxnet/pull/9002#discussion_r155867137
##
File path: src/imperative/imperative_utils.h
##
@@ -453,7 +453,7 @@ inline void PushOperator(const
eric-haibin-lin commented on issue #8669: module.Module and CSVIter
URL:
https://github.com/apache/incubator-mxnet/issues/8669#issuecomment-350362191
I see. MXNet should definitely update the api to allow custom data/label
names.
As a temporary walk-around, you can change the
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155863610
##
File path: src/kvstore/kvstore_dist.h
##
@@ -32,11 +32,6 @@
#include "mxnet/engine.h"
#include
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155863468
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software
This is an automated email from the ASF dual-hosted git repository.
zhreshold 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 075ef68 Fix 'from_layers' of
zhreshold closed pull request #8998: Fix names of extract-layers of resnet101
URL: https://github.com/apache/incubator-mxnet/pull/8998
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
eric-haibin-lin opened a new pull request #9003: Update PULL_REQUEST_TEMPLATE.md
URL: https://github.com/apache/incubator-mxnet/pull/9003
Added more items to guide pull request reviews.
This is an automated message from the
szha commented on a change in pull request #8763: [WIP] Add text apis
URL: https://github.com/apache/incubator-mxnet/pull/8763#discussion_r155858164
##
File path: python/mxnet/text/__init__.py
##
@@ -0,0 +1,25 @@
+# Licensed to the Apache Software Foundation (ASF) under
This is an automated email from the ASF dual-hosted git repository.
haibin pushed a commit to branch v1.0.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/v1.0.0 by this push:
new 0f05c65 Fix race condition in engine
piiswrong closed pull request #8987: Fix multi task
URL: https://github.com/apache/incubator-mxnet/pull/8987
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 request (from
piiswrong commented on a change in pull request #8990: Fix Example
example/dec/dec.py
URL: https://github.com/apache/incubator-mxnet/pull/8990#discussion_r155855492
##
File path: example/deep-embedded-clustering/dec.py
##
@@ -82,8 +82,8 @@ def list_arguments(self):
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 b652a20 Fix multi task (#8987)
b652a20
piiswrong closed pull request #8991: Some fixes for
example/reinforcement-learning/parallel_actor_critic
URL: https://github.com/apache/incubator-mxnet/pull/8991
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake
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 f154101 Some fixes for
piiswrong commented on a change in pull request #8993: Fix compatibility of
examples with py3
URL: https://github.com/apache/incubator-mxnet/pull/8993#discussion_r155855108
##
File path: example/rcnn/rcnn/core/tester.py
##
@@ -15,13 +15,14 @@
# specific language
piiswrong closed pull request #8996: Implement assignment operator on Timing
'Info' struct
URL: https://github.com/apache/incubator-mxnet/pull/8996
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:
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 c94b104 Implement assignment operator
piiswrong commented on a change in pull request #9002: Misc performance changes
URL: https://github.com/apache/incubator-mxnet/pull/9002#discussion_r155854440
##
File path: src/imperative/imperative_utils.h
##
@@ -453,7 +453,7 @@ inline void PushOperator(const OpStatePtr&
sxjscience commented on issue #5861: [WIP] Add norm_clipping and summary to
module + Update lstm_ptb example
URL: https://github.com/apache/incubator-mxnet/pull/5861#issuecomment-350348702
@eric-haibin-lin
```log
# Small Configuration
python lstm_ptb.py --num-hidden 200 \
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155851393
##
File path: src/kvstore/kvstore_local.h
##
@@ -256,7 +256,13 @@ class KVStoreLocal : public KVStore {
sxjscience commented on issue #2612: Slow speed of the AddTakeGradKernel in
embedding layer
URL:
https://github.com/apache/incubator-mxnet/issues/2612#issuecomment-350345689
@eric-haibin-lin The backward pass of the embedding layer has been
accelerated. However, we still need to compute
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155850266
##
File path: src/executor/graph_executor.cc
##
@@ -54,6 +54,14 @@ GraphExecutor::~GraphExecutor() {
}
}
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155850103
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software Foundation
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155850103
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software Foundation
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155849981
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software Foundation
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155847361
##
File path: prepare_mkldnn.sh
##
@@ -0,0 +1,121 @@
+#!/bin/bash
+
+# Licensed to the Apache Software
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155847115
##
File path: include/mxnet/ndarray.h
##
@@ -820,20 +798,14 @@ class NDArray {
// storage shape is also
larroy commented on issue #8988: [CI] Fix indentation of Jenkinsfile, no code
changes
URL: https://github.com/apache/incubator-mxnet/pull/8988#issuecomment-350342430
@cjolivier01 @piiswrong please merge
This is an automated
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155846518
##
File path: include/mxnet/ndarray.h
##
@@ -35,12 +35,12 @@
#include
#include
#include
+#if
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155846846
##
File path: include/mxnet/ndarray.h
##
@@ -608,6 +549,37 @@ class NDArray {
<<
zheng-da commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155845973
##
File path: Makefile
##
@@ -40,11 +40,11 @@ endif
# use customized config file
include $(config)
-ifeq
szha commented on a change in pull request #8957: Adding the steps to build
MXNet with NCCL
URL: https://github.com/apache/incubator-mxnet/pull/8957#discussion_r155841925
##
File path: docs/install/build_from_source.md
##
@@ -317,6 +317,40 @@ Use
This is an automated email from the ASF dual-hosted git repository.
zhasheng 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 5c11ecd Fix race condition in
szha closed pull request #8995: Fix race condition in engine start/stop
URL: https://github.com/apache/incubator-mxnet/pull/8995
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
leleamol commented on a change in pull request #8957: Adding the steps to build
MXNet with NCCL
URL: https://github.com/apache/incubator-mxnet/pull/8957#discussion_r155840307
##
File path: docs/install/build_from_source.md
##
@@ -317,6 +317,40 @@ Use
szha commented on issue #8995: Fix race condition in engine start/stop
URL: https://github.com/apache/incubator-mxnet/pull/8995#issuecomment-350333477
Given that it?s urgently needed, I?m merging. I agree that tests should be
added for catching this if feasible.
eric-haibin-lin commented on issue #2612: Slow speed of the AddTakeGradKernel
in embedding layer
URL:
https://github.com/apache/incubator-mxnet/issues/2612#issuecomment-350332561
Is this issue resolved?
This is an automated
marcoabreu commented on issue #9000: Flaky test OOM: test_optimizers:test_sgd
URL:
https://github.com/apache/incubator-mxnet/issues/9000#issuecomment-350330053
@ptrendx On GPU-instances we are only having 1 executor, meaning that
nosetests-3.4 is only triggered once and running in
marcoabreu commented on issue #9000: Flaky test OOM: test_optimizers:test_sgd
URL:
https://github.com/apache/incubator-mxnet/issues/9000#issuecomment-350330053
@ptrendx On GPU-instances we are only having 1 executor, meaning that
nosetests-3.4 is only triggered once and running in
eric-haibin-lin commented on issue #8941: MXNET backend in Keras 2
URL:
https://github.com/apache/incubator-mxnet/issues/8941#issuecomment-350329414
Yes! it's on our roadmap and under development.
@jiajiechen is there a target date that we're shooting?
marcoabreu commented on a change in pull request #8957: Adding the steps to
build MXNet with NCCL
URL: https://github.com/apache/incubator-mxnet/pull/8957#discussion_r155835284
##
File path: docs/install/build_from_source.md
##
@@ -317,6 +317,40 @@ Use
eric-haibin-lin commented on issue #8977: v1.0 with NCCL
URL:
https://github.com/apache/incubator-mxnet/issues/8977#issuecomment-350328153
It's on our roadmap. For now you can build it from source following the
instructions here: https://github.com/apache/incubator-mxnet/pull/8957/files
ptrendx commented on issue #9000: Flaky test OOM: test_optimizers:test_sgd
URL:
https://github.com/apache/incubator-mxnet/issues/9000#issuecomment-350326544
@marcoabreu Are you sure you are not running multiple tests in parallel on
the same machine that combined use that much memory?
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155833196
##
File path: src/ndarray/ndarray.cc
##
@@ -31,10 +31,12 @@
#include
#include
#include
+#include
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155829152
##
File path: src/executor/graph_executor.cc
##
@@ -54,6 +54,14 @@ GraphExecutor::~GraphExecutor() {
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155828922
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155828689
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155833088
##
File path: src/kvstore/kvstore_local.h
##
@@ -256,7 +256,13 @@ class KVStoreLocal : public KVStore {
eric-haibin-lin commented on a change in pull request #8302: Refactor operators
& MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155832508
##
File path: src/executor/infer_graph_attr_pass.cc
##
@@ -416,11 +416,6 @@ nnvm::Graph
ptrendx commented on a change in pull request #8957: Adding the steps to build
MXNet with NCCL
URL: https://github.com/apache/incubator-mxnet/pull/8957#discussion_r155833160
##
File path: docs/install/build_from_source.md
##
@@ -317,6 +317,40 @@ Use
ptrendx commented on a change in pull request #8957: Adding the steps to build
MXNet with NCCL
URL: https://github.com/apache/incubator-mxnet/pull/8957#discussion_r155832343
##
File path: docs/install/build_from_source.md
##
@@ -317,6 +317,40 @@ Use
cjolivier01 opened a new pull request #9002: Misc performance changes
URL: https://github.com/apache/incubator-mxnet/pull/9002
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
- [ ] Passed code style checking (`make lint`)
cjolivier01 commented on issue #8302: Refactor operators & MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#issuecomment-350313457
Please make sure all googletest tests in tests/cpp still build and pass
This is
cjolivier01 commented on issue #8302: Refactor operators & MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#issuecomment-350313321
Why is operator refactor and MKLDNN in the same PR?
This is an automated
cjolivier01 commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155818915
##
File path: include/mxnet/ndarray.h
##
@@ -35,12 +35,12 @@
#include
#include
#include
+#if
cjolivier01 commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155819722
##
File path: prepare_mkldnn.sh
##
@@ -0,0 +1,121 @@
+#!/bin/bash
+
+# Licensed to the Apache Software
cjolivier01 commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155821751
##
File path: src/operator/nn/batch_norm-inl.h
##
@@ -208,155 +127,127 @@ class BatchNormOp : public
cjolivier01 commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155819361
##
File path: include/mxnet/ndarray.h
##
@@ -820,20 +798,14 @@ class NDArray {
// storage shape is also
cjolivier01 commented on a change in pull request #8302: Refactor operators &
MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#discussion_r155819873
##
File path: python/mxnet/ndarray/mkldnn.py
##
@@ -0,0 +1,113 @@
+# Licensed to the Apache Software
szha commented on issue #7759: How to use cpu kernel of an operator (which
provides only cpu implementation) in gpu context?
URL:
https://github.com/apache/incubator-mxnet/issues/7759#issuecomment-350251485
This issue is closed due to lack of activity in the last 90 days. Feel free
to
szha closed issue #7759: How to use cpu kernel of an operator (which provides
only cpu implementation) in gpu context?
URL: https://github.com/apache/incubator-mxnet/issues/7759
This is an automated message from the Apache
3bhady commented on issue #8358: 'module' object has no attribute 'nd'
URL:
https://github.com/apache/incubator-mxnet/issues/8358#issuecomment-350250900
@QQQYang Thanks for replying, glad you solved yours.
But sadly this isn't the case for me i get this error on the validating
program
marcoabreu commented on issue #8763: [WIP] Add text apis
URL: https://github.com/apache/incubator-mxnet/pull/8763#issuecomment-350233958
Hello @astonzhang, please rebase your PR.
This is an automated message from the Apache
marcoabreu commented on issue #8302: Refactor operators & MKLDNN
URL: https://github.com/apache/incubator-mxnet/pull/8302#issuecomment-350233854
Hello @zheng-da, please rebase your PR.
This is an automated message from the
marcoabreu commented on issue #8993: Fix compatibility of examples with py3
URL: https://github.com/apache/incubator-mxnet/pull/8993#issuecomment-350233620
@rahul003 We're using G3.8xlarge with 8Gb of CUDA-memory - that should be
more than sufficient. I've submitted a PR to disable this
marcoabreu opened a new pull request #9001: Disable test_optimizers:test_sgd
(Temp fix #9000)
URL: https://github.com/apache/incubator-mxnet/pull/9001
See https://github.com/apache/incubator-mxnet/issues/9000
This is an
marcoabreu commented on issue #9000: Flaky test OOM: test_optimizers:test_sgd
URL:
https://github.com/apache/incubator-mxnet/issues/9000#issuecomment-350229681
We've got 8Gb of CUDA-memory. Considering MXNet should also run on Edge
devices, we should definitely reduce the memory usage.
marcoabreu opened a new issue #9000: Flaky test OOM: test_optimizers:test_sgd
URL: https://github.com/apache/incubator-mxnet/issues/9000
tests/ci_build/ci_build.sh gpu_mklml PYTHONPATH=./python/ nosetests-3.4
--with-timer --verbose tests/python/gpu
```
test_operator_gpu.test_sgd
janelu9 opened a new issue #8999: how to train data on hadoop ?
URL: https://github.com/apache/incubator-mxnet/issues/8999
This is an automated message from the Apache Git Service.
To respond to the message, please log on
101 - 170 of 170 matches
Mail list logo