piiswrong commented on a change in pull request #9981: [fix issue#9976] The
assignment problem in NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#discussion_r172040979
##
File path: src/ndarray/ndarray.cc
##
@@ -1128,7 +1128,7 @@ void CopyFromToImpl(const
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 94d3c06 Fix test_cast (#9951)
94d3c06
Ldpe2G opened a new pull request #9984: add reshape predicator function to
c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984
## Description ##
add a function to c_predict_api which can change the input shape of an
already created predicator .
related issue:
pengzhao-intel commented on issue #9977: Cpu lstm inference
URL: https://github.com/apache/incubator-mxnet/pull/9977#issuecomment-370213650
@Jerryzcn Thanks for the info. So, I think it's better to merge this PR
as-is.
Our LSTM/GRU will be ready in this month and we will submit the
pengzhao-intel commented on a change in pull request #9958: Parallelization for
ROIpooling OP
URL: https://github.com/apache/incubator-mxnet/pull/9958#discussion_r172045940
##
File path: src/operator/roi_pooling.cc
##
@@ -74,7 +79,13 @@ inline void ROIPoolForward(const
pengzhao-intel commented on issue #9939: add multi proposal operator (cpu
version) and fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370225375
@wkcn Nice work for the parallelization by OMP. **2X** speedup is a good
start
xinyu-intel commented on a change in pull request #9958: Parallelization for
ROIpooling OP
URL: https://github.com/apache/incubator-mxnet/pull/9958#discussion_r172048022
##
File path: src/operator/roi_pooling.cc
##
@@ -98,30 +109,26 @@ inline void ROIPoolForward(const
xinyu-intel commented on issue #9939: add multi proposal operator (cpu version)
and fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370230842
@wkcn @pengzhao-intel I think there are so much units in multi_proposal.cc.
We can
yajiedesign opened a new pull request #9985: add OPT:REF,Reduce the size of the
binary file.
URL: https://github.com/apache/incubator-mxnet/pull/9985
## Description ##
add Gy
add OPT:REF
add OPT:ICF
in msvc
it reduce the size of the binary file.
## Checklist ##
###
cjolivier01 commented on a change in pull request #9985: add OPT:REF,Reduce the
size of the windows binary file.
URL: https://github.com/apache/incubator-mxnet/pull/9985#discussion_r172057234
##
File path: CMakeLists.txt
##
@@ -583,8 +590,8 @@ endif()
if(USE_CUDA)
cjolivier01 commented on a change in pull request #9985: add OPT:REF,Reduce the
size of the windows binary file.
URL: https://github.com/apache/incubator-mxnet/pull/9985#discussion_r172057243
##
File path: CMakeLists.txt
##
@@ -583,8 +590,8 @@ endif()
if(USE_CUDA)
cjolivier01 commented on a change in pull request #9984: add reshape predicator
function to c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984#discussion_r172057455
##
File path: src/c_api/c_predict_api.cc
##
@@ -243,6 +249,95 @@ int
cjolivier01 commented on a change in pull request #9984: add reshape predicator
function to c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984#discussion_r172057545
##
File path: src/c_api/c_predict_api.cc
##
@@ -243,6 +249,95 @@ int
cjolivier01 commented on a change in pull request #9984: add reshape predicator
function to c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984#discussion_r172057426
##
File path: src/c_api/c_predict_api.cc
##
@@ -243,6 +249,95 @@ int
cjolivier01 commented on a change in pull request #9984: add reshape predicator
function to c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984#discussion_r172057371
##
File path: src/c_api/c_predict_api.cc
##
@@ -243,6 +249,95 @@ int
reminisce commented on a change in pull request #9981: [fix issue#9976] The
assignment problem in NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#discussion_r172065630
##
File path: src/ndarray/ndarray.cc
##
@@ -1128,7 +1128,7 @@ void CopyFromToImpl(const
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172065968
##
File path: tests/ci_build/Dockerfile.build_cuda8_cudnn7
yajiedesign commented on a change in pull request #9985: add OPT:REF,Reduce the
size of the windows binary file.
URL: https://github.com/apache/incubator-mxnet/pull/9985#discussion_r172065972
##
File path: CMakeLists.txt
##
@@ -583,8 +590,8 @@ endif()
if(USE_CUDA)
b0noI opened a new issue #9987: Implement an ability to use NCCL with the
distributed training
URL: https://github.com/apache/incubator-mxnet/issues/9987
Currently one can either use dist kvstore or nccl kvstore, however there is
no way to specify NCCL mode for the distributed training.
jeremiedb commented on issue #9803: R Metrics
URL: https://github.com/apache/incubator-mxnet/pull/9803#issuecomment-370259029
Added a metric_cpu option defaulting to TRUE to have metrics computed on
CPU, but allowing computation on GPU when needed. There's now no performance
tradeoff.
kenfehling commented on issue #9501: LSTM doesn't work with Symbols and
HybridBlocks
URL:
https://github.com/apache/incubator-mxnet/issues/9501#issuecomment-370260477
OK, thanks!
This is an automated message from the Apache
szha commented on issue #9501: LSTM doesn't work with Symbols and HybridBlocks
URL:
https://github.com/apache/incubator-mxnet/issues/9501#issuecomment-370260366
Our LSTM cell doesn't support full sequence hybridization yet. The
hybridization happens only at cell level. Thus, it will still
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063709
##
File path: src/operator/quantization/quantize_graph_pass.cc
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063725
##
File path: src/operator/quantization/quantize_graph_pass.cc
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063832
##
File path: src/operator/quantization/quantize_graph_pass.cc
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064056
##
File path: example/quantization/imagenet_inference.py
##
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064414
##
File path: include/mxnet/op_attr_types.h
##
@@ -261,6
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064407
##
File path: include/mxnet/op_attr_types.h
##
@@ -261,6
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064559
##
File path: python/mxnet/contrib/quantization.py
##
@@
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064592
##
File path: python/mxnet/contrib/quantization.py
##
@@
szha opened a new pull request #9986: gluon language modeling dataset and text
token reader
URL: https://github.com/apache/incubator-mxnet/pull/9986
## Description ##
gluon language modeling dataset and text token reader
## Checklist ##
### Essentials ###
- [x] Passed code
szha commented on issue #9986: gluon language modeling dataset and text token
reader
URL: https://github.com/apache/incubator-mxnet/pull/9986#issuecomment-370260251
the PR for vocabulary, vectors, and model zoo models will happen separately.
kenfehling commented on issue #9501: LSTM doesn't work with Symbols and
HybridBlocks
URL:
https://github.com/apache/incubator-mxnet/issues/9501#issuecomment-370260239
I have a question about its performance. The documentation says:
> Recurrent cells allows fine-grained control when
szha commented on issue #9977: Cpu lstm inference
URL: https://github.com/apache/incubator-mxnet/pull/9977#issuecomment-370261454
@pengzhao-intel will the elman RNN be part of your PR? cudnn currently
supports it, so we support it as part of the RNN layers.
KellenSunderland commented on a change in pull request #9918: Update mkldnn to
the newest & Add clang build test with mkldnn.
URL: https://github.com/apache/incubator-mxnet/pull/9918#discussion_r172063496
##
File path: tests/ci_build/Dockerfile.cpu_clang
##
@@ -19,3 +19,9
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063626
##
File path: src/operator/quantization/quantize_graph_pass.cc
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063832
##
File path: src/operator/quantization/quantize_graph_pass.cc
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172063979
##
File path: example/quantization/imagenet_gen_qsym.py
##
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064304
##
File path: example/quantization/launch_quantize.sh
##
@@
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064298
##
File path: example/quantization/launch_inference.sh
##
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064374
##
File path: include/mxnet/c_api.h
##
@@ -1386,8 +1386,37
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064452
##
File path: include/mxnet/op_attr_types.h
##
@@ -261,6
KellenSunderland commented on a change in pull request #9552: [REQUEST FOR
REVIEW | DO NOT MERGE] Model Quantization with Calibration
URL: https://github.com/apache/incubator-mxnet/pull/9552#discussion_r172064433
##
File path: include/mxnet/op_attr_types.h
##
@@ -261,6
Ldpe2G commented on a change in pull request #9984: add reshape predicator
function to c_predict_api
URL: https://github.com/apache/incubator-mxnet/pull/9984#discussion_r172070078
##
File path: src/c_api/c_predict_api.cc
##
@@ -243,6 +249,95 @@ int
cgraywang commented on issue #9572: Return a NaN when using operator ( ** ) on
Windows version with GPU
URL:
https://github.com/apache/incubator-mxnet/issues/9572#issuecomment-370275278
@Feywell ok, I will be looking into it. Since I will need to setting up the
windows, it may take
jeremiedb commented on issue #9803: R Metrics
URL: https://github.com/apache/incubator-mxnet/pull/9803#issuecomment-370259029
Added a metric_cpu option defaulting to TRUE to have metrics computed on
CPU, but allowing computation on GPU when needed. There's now no performance
tradeoff.
samhodge commented on issue #9989: Cannot train example gluon style transfer
URL:
https://github.com/apache/incubator-mxnet/issues/9989#issuecomment-370296496
@zhanghang1989 Do you have any input?
This is an automated
wkcn commented on issue #9939: add multi proposal operator (cpu version) and
fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370320488
For the cpu/gpu consistency, it's interesting that the number of valid
anchors in the CPU
wkcn commented on a change in pull request #9981: [fix issue#9976] The
assignment problem in NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#discussion_r172075822
##
File path: tests/python/unittest/test_ndarray.py
##
@@ -1099,6 +1099,39 @@ def
wkcn commented on issue #9939: add multi proposal operator (cpu version) and
fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370285916
@pengzhao-intel @xinyu-intel
Thank you! I will have a try.
HaoLiuHust closed issue #9962: how to group 2 or more outputs in gluon just
like sym.Group()
URL: https://github.com/apache/incubator-mxnet/issues/9962
This is an automated message from the Apache Git Service.
To respond to
HaoLiuHust commented on issue #9962: how to group 2 or more outputs in gluon
just like sym.Group()
URL:
https://github.com/apache/incubator-mxnet/issues/9962#issuecomment-370288388
@orchidmajumder Thanks, get your point
TaoLv commented on issue #9977: Cpu lstm inference
URL: https://github.com/apache/incubator-mxnet/pull/9977#issuecomment-370287856
@szha To make it easier to review, we would like to split the whole RNN
implementation on CPU into several PRs. Firstly, We will submit code for
single-layer
samhodge opened a new issue #9989: Cannot train example gluon style transfer
URL: https://github.com/apache/incubator-mxnet/issues/9989
Note: Providing complete information in the most concise form is the best
way to get help. This issue template serves as the checklist for essential
szha commented on issue #9977: Cpu lstm inference
URL: https://github.com/apache/incubator-mxnet/pull/9977#issuecomment-370295377
@TaoLv sounds good. What timeline are we looking at for feature parity with
cudnn?
This is an
sergeykolychev commented on issue #9988: [Perl] Sparse feature.
URL: https://github.com/apache/incubator-mxnet/pull/9988#issuecomment-370290337
@tlby Rob, please take a look, thanks!
This is an automated message from the
wkcn commented on issue #9939: add multi proposal operator (cpu version) and
fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370320488
For the cpu/gpu consistency test, it's interesting that the number of valid
anchors in the
ZiyueHuang commented on issue #9887: Non-blocking row_sparse_pull. Fix
incorrect indices generated by device kvstore.row_sparse_pull
URL: https://github.com/apache/incubator-mxnet/pull/9887#issuecomment-370323043
@eric-haibin-lin Thanks for the fix!
Looks good.
asitstands opened a new pull request #9991: Random shuffle implementation
URL: https://github.com/apache/incubator-mxnet/pull/9991
## Description ##
This operator randomly shuffles the elements of an NDArray along the last
axis, i.e., it performs a batched shuffling. For example, if
piiswrong commented on issue #9981: [fix issue#9976] The assignment problem in
NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#issuecomment-370287669
@marcoabreu retrigger test?
This is an automated message
TaoLv commented on issue #9977: Cpu lstm inference
URL: https://github.com/apache/incubator-mxnet/pull/9977#issuecomment-370287856
@szha To make it easier to review, we would like to split the whole RNN
implementation on CPU into several PRs. Firstly, We will submit code for
single-layer
arundasan91 opened a new pull request #9990: Update tensorboard.py
URL: https://github.com/apache/incubator-mxnet/pull/9990
Adding `step` to Tensorboard. Currently only horizontal axis setting of
`Relative` or `Wall` are available without any info about the epoch.
## Description ##
wkcn commented on a change in pull request #9981: [fix issue#9976] The
assignment problem in NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#discussion_r172075839
##
File path: tests/python/unittest/test_ndarray.py
##
@@ -1099,6 +1099,39 @@ def
wkcn commented on issue #9939: add multi proposal operator (cpu version) and
fix the bug in proposal op (gpu version)
URL: https://github.com/apache/incubator-mxnet/pull/9939#issuecomment-370320488
For the cpu/gpu consistency test, it's interesting that the number of valid
anchors in the
wkcn commented on issue #9981: [fix issue#9976] The assignment problem in
NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#issuecomment-370288088
@piiswrong Thank you! I will change the code and then trigger test.
wkcn commented on a change in pull request #9981: [fix issue#9976] The
assignment problem in NDArray
URL: https://github.com/apache/incubator-mxnet/pull/9981#discussion_r172075502
##
File path: tests/python/unittest/test_ndarray.py
##
@@ -1099,6 +1099,39 @@ def
eric-haibin-lin commented on issue #9887: Non-blocking row_sparse_pull
URL: https://github.com/apache/incubator-mxnet/pull/9887#issuecomment-370303541
This PR also fixed a bug in the original GPU "Unique" implementation, where
the same pointer is passed as both the input/output to
eric-haibin-lin commented on a change in pull request #9887: Non-blocking
row_sparse_pull
URL: https://github.com/apache/incubator-mxnet/pull/9887#discussion_r172087240
##
File path: tests/python/unittest/test_kvstore.py
##
@@ -76,7 +76,7 @@ def check_row_sparse_pull(kv,
eric-haibin-lin commented on issue #9887: Non-blocking row_sparse_pull
URL: https://github.com/apache/incubator-mxnet/pull/9887#issuecomment-370303883
@ZiyueHuang please view the changes for gpu unique
This is an automated
eric-haibin-lin commented on issue #9887: Non-blocking row_sparse_pull
URL: https://github.com/apache/incubator-mxnet/pull/9887#issuecomment-370303541
This PR also fixed a bug in the original GPU "Unique" implementation, where
the same pointer is passed as both the input/output to
II-Matto closed issue #9012: Implementing New Operators
URL: https://github.com/apache/incubator-mxnet/issues/9012
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and
xinyu-intel commented on a change in pull request #9918: Update mkldnn to the
newest & Add clang build test with mkldnn.
URL: https://github.com/apache/incubator-mxnet/pull/9918#discussion_r172090005
##
File path: tests/ci_build/Dockerfile.cpu_clang
##
@@ -19,3 +19,9 @@
xinyu-intel commented on a change in pull request #9918: Update mkldnn to the
newest & Add clang build test with mkldnn.
URL: https://github.com/apache/incubator-mxnet/pull/9918#discussion_r172092504
##
File path: tests/ci_build/Dockerfile.cpu_clang
##
@@ -19,3 +19,9 @@
sergeykolychev opened a new pull request #9988: [Perl] Sparse feature.
URL: https://github.com/apache/incubator-mxnet/pull/9988
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
- [ ] Passed code style checking (`make lint`)
74 matches
Mail list logo