[GitHub] sandeep-krishnamurthy commented on issue #8999: how to train data on hadoop ?

sandeep-krishnamurthy commented on issue #8999: how to train data on hadoop ? URL: https://github.com/apache/incubator-mxnet/issues/8999#issuecomment-372098998 Can you please usage related query at - https://discuss.mxnet.io/ Please file a GitHub issue if you encounter any issue. -

[GitHub] sandeep-krishnamurthy closed issue #8999: how to train data on hadoop ?

sandeep-krishnamurthy closed 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 lo

[GitHub] sandeep-krishnamurthy commented on issue #8933: How to train data with multi-class label

sandeep-krishnamurthy commented on issue #8933: How to train data with multi-class label URL: https://github.com/apache/incubator-mxnet/issues/8933#issuecomment-372099535 Can you please ask usage related questions at - https://discuss.mxnet.io/

[GitHub] sandeep-krishnamurthy closed issue #8933: How to train data with multi-class label

sandeep-krishnamurthy closed issue #8933: How to train data with multi-class label URL: https://github.com/apache/incubator-mxnet/issues/8933 This is an automated message from the Apache Git Service. To respond to the messag

[GitHub] sandeep-krishnamurthy closed issue #8885: path_imglist of mx.io.ImageRecordIter

sandeep-krishnamurthy closed issue #8885: path_imglist of mx.io.ImageRecordIter URL: https://github.com/apache/incubator-mxnet/issues/8885 This is an automated message from the Apache Git Service. To respond to the message, p

[GitHub] sandeep-krishnamurthy commented on issue #8885: path_imglist of mx.io.ImageRecordIter

sandeep-krishnamurthy commented on issue #8885: path_imglist of mx.io.ImageRecordIter URL: https://github.com/apache/incubator-mxnet/issues/8885#issuecomment-372099837 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reopen if you encounter a

[GitHub] sandeep-krishnamurthy commented on issue #8868: Not able to able to understand mx.symbol.take()? Example

sandeep-krishnamurthy commented on issue #8868: Not able to able to understand mx.symbol.take()? Example URL: https://github.com/apache/incubator-mxnet/issues/8868#issuecomment-37216 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reope

[GitHub] sandeep-krishnamurthy closed issue #8868: Not able to able to understand mx.symbol.take()? Example

sandeep-krishnamurthy closed issue #8868: Not able to able to understand mx.symbol.take()? Example URL: https://github.com/apache/incubator-mxnet/issues/8868 This is an automated message from the Apache Git Service. To resp

[GitHub] sandeep-krishnamurthy commented on issue #8828: Change layer or weight name in Gluon

sandeep-krishnamurthy commented on issue #8828: Change layer or weight name in Gluon URL: https://github.com/apache/incubator-mxnet/issues/8828#issuecomment-372100319 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reopen if you encounter an

[GitHub] sandeep-krishnamurthy closed issue #8828: Change layer or weight name in Gluon

sandeep-krishnamurthy closed issue #8828: Change layer or weight name in Gluon URL: https://github.com/apache/incubator-mxnet/issues/8828 This is an automated message from the Apache Git Service. To respond to the message, pl

[GitHub] sandeep-krishnamurthy closed issue #8823: About the speed of data parallel training on different machines

sandeep-krishnamurthy closed issue #8823: About the speed of data parallel training on different machines URL: https://github.com/apache/incubator-mxnet/issues/8823 This is an automated message from the Apache Git Service. T

[GitHub] sandeep-krishnamurthy commented on issue #8823: About the speed of data parallel training on different machines

sandeep-krishnamurthy commented on issue #8823: About the speed of data parallel training on different machines URL: https://github.com/apache/incubator-mxnet/issues/8823#issuecomment-372100334 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to

[GitHub] sandeep-krishnamurthy closed issue #8789: I had two pairs of set of images and one ground truth image. How to train in mxnet?

sandeep-krishnamurthy closed issue #8789: I had two pairs of set of images and one ground truth image. How to train in mxnet? URL: https://github.com/apache/incubator-mxnet/issues/8789 This is an automated message from the A

[GitHub] sandeep-krishnamurthy commented on issue #8789: I had two pairs of set of images and one ground truth image. How to train in mxnet?

sandeep-krishnamurthy commented on issue #8789: I had two pairs of set of images and one ground truth image. How to train in mxnet? URL: https://github.com/apache/incubator-mxnet/issues/8789#issuecomment-372100543 Can you please ask usage related questions at -?https://discuss.mxnet.io/

[GitHub] sandeep-krishnamurthy commented on issue #8783: how to print each loss&acc for every classes

sandeep-krishnamurthy commented on issue #8783: how to print each loss&acc for every classes URL: https://github.com/apache/incubator-mxnet/issues/8783#issuecomment-372100567 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reopen if you enco

[GitHub] sandeep-krishnamurthy closed issue #8783: how to print each loss&acc for every classes

sandeep-krishnamurthy closed issue #8783: how to print each loss&acc for every classes URL: https://github.com/apache/incubator-mxnet/issues/8783 This is an automated message from the Apache Git Service. To respond to the me

[GitHub] sandeep-krishnamurthy commented on issue #8768: how to use mx.io.extract()

sandeep-krishnamurthy commented on issue #8768: how to use mx.io.extract() URL: https://github.com/apache/incubator-mxnet/issues/8768#issuecomment-372100630 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reopen if you encounter any issues.

[GitHub] sandeep-krishnamurthy closed issue #8768: how to use mx.io.extract()

sandeep-krishnamurthy closed issue #8768: how to use mx.io.extract() URL: https://github.com/apache/incubator-mxnet/issues/8768 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] sandeep-krishnamurthy closed issue #8745: Utilities for estimating dense optical flow in mxnet?

sandeep-krishnamurthy closed issue #8745: Utilities for estimating dense optical flow in mxnet? URL: https://github.com/apache/incubator-mxnet/issues/8745 This is an automated message from the Apache Git Service. To respond

[GitHub] sandeep-krishnamurthy commented on issue #8745: Utilities for estimating dense optical flow in mxnet?

sandeep-krishnamurthy commented on issue #8745: Utilities for estimating dense optical flow in mxnet? URL: https://github.com/apache/incubator-mxnet/issues/8745#issuecomment-372100763 Can you please ask usage related questions at -?https://discuss.mxnet.io/ Please feel free to reopen if

[GitHub] piiswrong commented on issue #9931: Add axes support to Dropout for variational dropout in NLP

piiswrong commented on issue #9931: Add axes support to Dropout for variational dropout in NLP URL: https://github.com/apache/incubator-mxnet/pull/9931#issuecomment-372101730 @yzhliu @zhanghang1989 Please make sure we don't merge code without test coverage next time. --

[GitHub] piiswrong commented on issue #9958: Parallelization for ROIpooling OP

piiswrong commented on issue #9958: Parallelization for ROIpooling OP URL: https://github.com/apache/incubator-mxnet/pull/9958#issuecomment-372102097 @cjolivier01 Shouldn't the omp pragma get number of threads from OpenMP::Get()? -

[GitHub] piiswrong commented on issue #9932: Fixes for profiler

piiswrong commented on issue #9932: Fixes for profiler URL: https://github.com/apache/incubator-mxnet/pull/9932#issuecomment-372103844 @cjolivier01 ping This is an automated message from the Apache Git Service. To respond to t

[GitHub] piiswrong commented on issue #9982: Unary ops logcdf_normal, derivlogcdf_normal [MXNET-39]

piiswrong commented on issue #9982: Unary ops logcdf_normal, derivlogcdf_normal [MXNET-39] URL: https://github.com/apache/incubator-mxnet/pull/9982#issuecomment-372103968 I think putting it in contrib is fine for now until we decide if we want a mx.distributions name space ---

[GitHub] marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update

marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update URL: https://github.com/apache/incubator-mxnet/pull/10062#issuecomment-372104528 Very nice catch! Do you have an estimation how much overall speedup this could bring? We could highlight this in the release n

[GitHub] marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update

marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update URL: https://github.com/apache/incubator-mxnet/pull/10062#issuecomment-372104639 By the way, could you add the benchmark at tests/python/benchmark so we can use them leter on? --

[GitHub] marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update

marcoabreu commented on issue #10062: [MXNET-72] [WIP] Improve sparse.adam_update URL: https://github.com/apache/incubator-mxnet/pull/10062#issuecomment-372104639 By the way, could you add the benchmark at tests/python/benchmark so we can use them later on? --

[GitHub] cjolivier01 commented on issue #9958: Parallelization for ROIpooling OP

cjolivier01 commented on issue #9958: Parallelization for ROIpooling OP URL: https://github.com/apache/incubator-mxnet/pull/9958#issuecomment-372104910 it probably wouldn?t hurt. I don?t think it?s critical, though since channels tends to be a small number and it doesn?t look like the inten

[GitHub] kobenaxie closed issue #8886: Bug in mx.nd.NDArray.reshape ~

kobenaxie closed issue #8886: Bug in mx.nd.NDArray.reshape ~ URL: https://github.com/apache/incubator-mxnet/issues/8886 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub

[GitHub] cjolivier01 closed pull request #9932: Fixes for profiler

cjolivier01 closed pull request #9932: Fixes for profiler URL: https://github.com/apache/incubator-mxnet/pull/9932 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

[incubator-mxnet] branch master updated: Fixes for profiler (#9932)

This is an automated email from the ASF dual-hosted git repository. cjolivier01 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 94f68fc Fixes for profiler (#9932

[GitHub] hsddlz opened a new issue #10065: Error in operator reshape

hsddlz opened a new issue #10065: Error in operator reshape URL: https://github.com/apache/incubator-mxnet/issues/10065 when I bind the data .It told me like this: simple_bind error. Arguments: data: (32, 1L, 32L, 286L) softmax_label: (32, 21L) Error in operator reshape0: [1

[GitHub] EternalSaga commented on issue #9944: MXNet MinGW-w64 build error

EternalSaga commented on issue #9944: MXNet MinGW-w64 build error URL: https://github.com/apache/incubator-mxnet/issues/9944#issuecomment-372108671 I encountered the similar error when I was building the cpp package examples in VS2015 --

[GitHub] hsddlz closed issue #10065: Error in operator reshape

hsddlz closed issue #10065: Error in operator reshape URL: https://github.com/apache/incubator-mxnet/issues/10065 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and

[GitHub] hsddlz commented on issue #10065: Error in operator reshape

hsddlz commented on issue #10065: Error in operator reshape URL: https://github.com/apache/incubator-mxnet/issues/10065#issuecomment-372108989 shape=(batch_size,1,rnn_length, rnn_dimen)sorry This is an automated message

[GitHub] leopd commented on issue #8835: Python crashes (core-dump) instead of a graceful error message when GPU context is used on a CPU-only instance (EC2 x1.32xlarge)

leopd commented on issue #8835: Python crashes (core-dump) instead of a graceful error message when GPU context is used on a CPU-only instance (EC2 x1.32xlarge) URL: https://github.com/apache/incubator-mxnet/issues/8835#issuecomment-372110162 I'm still seeing this on latest the AWS DL AMI

[GitHub] asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation

asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation URL: https://github.com/apache/incubator-mxnet/pull/10048#discussion_r173655956 ## File path: tests/python/unittest/test_random.py ## @@ -552,6 +554,79 @@ def compute_expected

[GitHub] asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation

asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation URL: https://github.com/apache/incubator-mxnet/pull/10048#discussion_r173656027 ## File path: tests/python/unittest/test_random.py ## @@ -552,6 +554,79 @@ def compute_expected

[GitHub] asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation

asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation URL: https://github.com/apache/incubator-mxnet/pull/10048#discussion_r173656043 ## File path: tests/python/unittest/test_random.py ## @@ -552,6 +554,79 @@ def compute_expected

[GitHub] asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation

asitstands commented on a change in pull request #10048: [MXNET-68] Random shuffle implementation URL: https://github.com/apache/incubator-mxnet/pull/10048#discussion_r173656043 ## File path: tests/python/unittest/test_random.py ## @@ -552,6 +554,79 @@ def compute_expected

[GitHub] mike07026 commented on issue #8360: How to bind different input shape to executor in c++?

mike07026 commented on issue #8360: How to bind different input shape to executor in c++? URL: https://github.com/apache/incubator-mxnet/issues/8360#issuecomment-372128240 Yes, I meet the same problem when using mtcnn(face detector). I hope C/C++ interface support this case better. @szha

[GitHub] mike07026 commented on issue #8360: How to bind different input shape to executor in c++?

mike07026 commented on issue #8360: How to bind different input shape to executor in c++? URL: https://github.com/apache/incubator-mxnet/issues/8360#issuecomment-372128240 Yes, I meet the same problem when using mtcnn(face detector). I hope C/C++ interface will support this case better. @

[GitHub] mike07026 commented on issue #8360: How to bind different input shape to executor in c++?

mike07026 commented on issue #8360: How to bind different input shape to executor in c++? URL: https://github.com/apache/incubator-mxnet/issues/8360#issuecomment-372128240 Yes, I meet the same problem when using mtcnn(face detector). I hope C/C++ interface will support this case better in

[GitHub] cloudfool opened a new issue #10066: WarpCTC loss output

cloudfool opened a new issue #10066: WarpCTC loss output URL: https://github.com/apache/incubator-mxnet/issues/10066 Hi, I want to output the warpctc loss values for each batch. Do I have to write a new metric for this by myself ? or is there any work I can refer ? Please giv

[GitHub] ericvlognow opened a new issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow opened a new issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067 i am trying to use mxnet to analyze the keras imdb dataset. using attached python script (imdb_mx2.py), i

[GitHub] ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067#issuecomment-372144334 following are outputs when maxLen = 150 dyld: warning, LC_RPATH ${ORIGIN} in /Li

[GitHub] ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067#issuecomment-372144369 following are outputs when maxLen = 500 dyld: warning, LC_RPATH ${ORIGIN} in /Li

[GitHub] ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067#issuecomment-372144527 i tried it in keras with tensor_flow back end, even with max_len = 500, it still g

[GitHub] ShootingSpace opened a new issue #10068: rnn.encode_sentences deals with unknown token

ShootingSpace opened a new issue #10068: rnn.encode_sentences deals with unknown token URL: https://github.com/apache/incubator-mxnet/issues/10068 ## Description For rnn.encode_sentences(), could mxnet provides user self-defined behavior when vocab dictionary is given, instead of just r

[GitHub] Jerryzcn commented on issue #10042: Gluon dataloader crash on speech recognition training

Jerryzcn commented on issue #10042: Gluon dataloader crash on speech recognition training URL: https://github.com/apache/incubator-mxnet/issues/10042#issuecomment-372152880 Seems like related to multiprocessing. when num_worker=0 problem is resolved. --

[GitHub] zheng-da commented on a change in pull request #9552: [REQUEST FOR REVIEW | DO NOT MERGE] Model Quantization with Calibration

zheng-da 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_r173673443 ## File path: src/operator/quantization/quantized_pooling.cu ## @@ -

[GitHub] zheng-da commented on a change in pull request #9552: [REQUEST FOR REVIEW | DO NOT MERGE] Model Quantization with Calibration

zheng-da 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_r173673398 ## File path: src/operator/quantization/quantized_conv.cu ## @@ -0,0

[GitHub] szha commented on issue #10042: Gluon dataloader crash on speech recognition training

szha commented on issue #10042: Gluon dataloader crash on speech recognition training URL: https://github.com/apache/incubator-mxnet/issues/10042#issuecomment-372168514 This isn't actionable since we don't have your code. Please attach code. ---

[GitHub] szha commented on issue #9705: Added unittest for benchmarking metric performance

szha commented on issue #9705: Added unittest for benchmarking metric performance URL: https://github.com/apache/incubator-mxnet/pull/9705#issuecomment-372170145 Checking in on the public nightly build results, is it still on track? -

[GitHub] szha commented on a change in pull request #10025: Language model with Google's billion words dataset

szha commented on a change in pull request #10025: Language model with Google's billion words dataset URL: https://github.com/apache/incubator-mxnet/pull/10025#discussion_r173680183 ## File path: example/rnn/large_word_lm/data.py ## @@ -0,0 +1,202 @@ +# Licensed to the Apa

[GitHub] marcoabreu commented on issue #9705: Added unittest for benchmarking metric performance

marcoabreu commented on issue #9705: Added unittest for benchmarking metric performance URL: https://github.com/apache/incubator-mxnet/pull/9705#issuecomment-372170806 I don't think so - at least not from my side. We have been resource constrained and managing the Nightly CI does not fit i

[GitHub] szha commented on a change in pull request #10025: Language model with Google's billion words dataset

szha commented on a change in pull request #10025: Language model with Google's billion words dataset URL: https://github.com/apache/incubator-mxnet/pull/10025#discussion_r173680648 ## File path: example/rnn/large_word_lm/model.py ## @@ -0,0 +1,181 @@ +# Licensed to the Ap

[GitHub] szha commented on a change in pull request #10025: Language model with Google's billion words dataset

szha commented on a change in pull request #10025: Language model with Google's billion words dataset URL: https://github.com/apache/incubator-mxnet/pull/10025#discussion_r173680708 ## File path: python/mxnet/gluon/contrib/rnn/rnn_cell.py ## @@ -181,3 +181,126 @@ def unrol

[GitHub] Jerryzcn commented on issue #10042: Gluon dataloader crash on speech recognition training

Jerryzcn commented on issue #10042: Gluon dataloader crash on speech recognition training URL: https://github.com/apache/incubator-mxnet/issues/10042#issuecomment-372171903 will produce minimal reproducible code soon This i

[GitHub] szha commented on a change in pull request #10025: Language model with Google's billion words dataset

szha commented on a change in pull request #10025: Language model with Google's billion words dataset URL: https://github.com/apache/incubator-mxnet/pull/10025#discussion_r173681036 ## File path: src/operator/nn/fully_connected-inl.h ## @@ -95,11 +95,16 @@ void FCForward(c

[GitHub] szha commented on issue #9705: Added unittest for benchmarking metric performance

szha commented on issue #9705: Added unittest for benchmarking metric performance URL: https://github.com/apache/incubator-mxnet/pull/9705#issuecomment-372172430 In that case, let's put the test in unittest for now. @safrooze could you resolve conflict? ---

[GitHub] zhanghang1989 commented on issue #9931: Add axes support to Dropout for variational dropout in NLP

zhanghang1989 commented on issue #9931: Add axes support to Dropout for variational dropout in NLP URL: https://github.com/apache/incubator-mxnet/pull/9931#issuecomment-372183180 ? Got it. My bad. This is an automated message

[GitHub] anirudhacharya commented on a change in pull request #9963: [MXNET-34] Onnx Module to import onnx models into mxnet

anirudhacharya commented on a change in pull request #9963: [MXNET-34] Onnx Module to import onnx models into mxnet URL: https://github.com/apache/incubator-mxnet/pull/9963#discussion_r173687805 ## File path: python/mxnet/contrib/onnx/_import/import_onnx.py ## @@ -0,0 +1,1

[GitHub] anirudhacharya commented on a change in pull request #9963: [MXNET-34] Onnx Module to import onnx models into mxnet

anirudhacharya commented on a change in pull request #9963: [MXNET-34] Onnx Module to import onnx models into mxnet URL: https://github.com/apache/incubator-mxnet/pull/9963#discussion_r173690286 ## File path: example/onnx/test_super_resolution.py ## @@ -0,0 +1,112 @@ +# Li

[GitHub] ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067#issuecomment-372144527 i tried it in keras with tensor_flow back end, even with max_len = 500, it generat

[GitHub] samhodge commented on issue #9989: Cannot train example gluon style transfer

samhodge commented on issue #9989: Cannot train example gluon style transfer URL: https://github.com/apache/incubator-mxnet/issues/9989#issuecomment-372190518 @zhanghang1989 Thanks for looking into this. I realised that moving L82 out of the autorecord messed up the resulting model, I was

[GitHub] ericvlognow closed issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow closed issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067 This is an automated message from the Apache Gi

[GitHub] ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500

ericvlognow commented on issue #10067: keras imdb data set is not converged when maximum size of the word sequence is 500 URL: https://github.com/apache/incubator-mxnet/issues/10067#issuecomment-372201693 i found the issue. the pre-padding instead of post-padding solved the issue.