[jira] [Updated] (MXNET-1249) Fix Performance issue with GPU of Object Detector
[ https://issues.apache.org/jira/browse/MXNET-1249?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Qing Lan updated MXNET-1249: Assignee: Qing Lan Status: In Progress (was: To Do) > Fix Performance issue with GPU of Object Detector > - > > Key: MXNET-1249 > URL: https://issues.apache.org/jira/browse/MXNET-1249 > Project: Apache MXNet > Issue Type: New Feature > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Qing Lan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1249) Fix Performance issue with GPU of Object Detector
[ https://issues.apache.org/jira/browse/MXNET-1249?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Qing Lan updated MXNET-1249: Status: In Review (was: In Progress) > Fix Performance issue with GPU of Object Detector > - > > Key: MXNET-1249 > URL: https://issues.apache.org/jira/browse/MXNET-1249 > Project: Apache MXNet > Issue Type: New Feature > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Qing Lan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1249) Fix Performance issue with GPU of Object Detector
Qing Lan created MXNET-1249: --- Summary: Fix Performance issue with GPU of Object Detector Key: MXNET-1249 URL: https://issues.apache.org/jira/browse/MXNET-1249 Project: Apache MXNet Issue Type: New Feature Components: Apache MXNet Java API, Apache MXNet Scala API Reporter: Qing Lan -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Deleted] (MXNET-1247) Introduce Travis CI for automation of Scala/Java packages on Mac OSX
[ https://issues.apache.org/jira/browse/MXNET-1247?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko deleted MXNET-1247: --- > Introduce Travis CI for automation of Scala/Java packages on Mac OSX > > > Key: MXNET-1247 > URL: https://issues.apache.org/jira/browse/MXNET-1247 > Project: Apache MXNet > Issue Type: Story >Reporter: Denis Davydenko >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1248) Prepare Travis build for MXNet Scala/Java package
Qing Lan created MXNET-1248: --- Summary: Prepare Travis build for MXNet Scala/Java package Key: MXNET-1248 URL: https://issues.apache.org/jira/browse/MXNET-1248 Project: Apache MXNet Issue Type: New Feature Components: Apache MXNet Java API, Apache MXNet Scala API Reporter: Qing Lan Due to the needs of Mac OS build/test for Scala/Java package, we decided to introduce the Travis build pipeline. # Clean up and apply the existing build code # Enable monitoring -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1247) Introduce Travis CI for automation of Scala/Java packages
Denis Davydenko created MXNET-1247: -- Summary: Introduce Travis CI for automation of Scala/Java packages Key: MXNET-1247 URL: https://issues.apache.org/jira/browse/MXNET-1247 Project: Apache MXNet Issue Type: Story Reporter: Denis Davydenko -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1247) Introduce Travis CI for automation of Scala/Java packages on Mac OSX
[ https://issues.apache.org/jira/browse/MXNET-1247?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1247: --- Summary: Introduce Travis CI for automation of Scala/Java packages on Mac OSX (was: Introduce Travis CI for automation of Scala/Java packages) > Introduce Travis CI for automation of Scala/Java packages on Mac OSX > > > Key: MXNET-1247 > URL: https://issues.apache.org/jira/browse/MXNET-1247 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Denis Davydenko >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1247) Introduce Travis CI for automation of Scala/Java packages
[ https://issues.apache.org/jira/browse/MXNET-1247?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1247: --- Component/s: Apache MXNet Scala API Apache MXNet Java API > Introduce Travis CI for automation of Scala/Java packages > - > > Key: MXNET-1247 > URL: https://issues.apache.org/jira/browse/MXNET-1247 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Denis Davydenko >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1241) Scala/Java Packages in Maven for CentOS
[ https://issues.apache.org/jira/browse/MXNET-1241?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1241: -- Assignee: Zach Kimberg > Scala/Java Packages in Maven for CentOS > --- > > Key: MXNET-1241 > URL: https://issues.apache.org/jira/browse/MXNET-1241 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Naveen Swamy >Assignee: Zach Kimberg >Priority: Major > > Some users are interested to use the Scala package on CentOS. > We should enable the new maven packages(with all dependencies) to be used on > CentOS by adding suitable tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1241) Scala/Java Packages in Maven for CentOS
[ https://issues.apache.org/jira/browse/MXNET-1241?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1241: --- Sprint: Scala/Java 10 (12/3-12/17) > Scala/Java Packages in Maven for CentOS > --- > > Key: MXNET-1241 > URL: https://issues.apache.org/jira/browse/MXNET-1241 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Naveen Swamy >Priority: Major > > Some users are interested to use the Scala package on CentOS. > We should enable the new maven packages(with all dependencies) to be used on > CentOS by adding suitable tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-376) Hardmax
[ https://issues.apache.org/jira/browse/MXNET-376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708145#comment-16708145 ] Anirudh Acharya commented on MXNET-376: --- yea sure. The formula I had written was not a generic formula but it was written for illustration purpose for a certain axis value. It will have to be modified for generic axis values. While implementing the op, make sure you implement the correct definition. > Hardmax > --- > > Key: MXNET-376 > URL: https://issues.apache.org/jira/browse/MXNET-376 > Project: Apache MXNet > Issue Type: Sub-task >Reporter: Hao Jin >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Comment Edited] (MXNET-376) Hardmax
[ https://issues.apache.org/jira/browse/MXNET-376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708069#comment-16708069 ] Chaitanya Prakash Bapat edited comment on MXNET-376 at 12/4/18 3:23 AM: [~anirudhacharya] It looks like the above code doesn't give the same result as Tensorflow does For the same input {code:python} import numpy as np data= np.random.rand(2,3,4) print(data) [[[0.18651888 0.29437149 0.45589573 0.55928469] [0.1951702 0.50387834 0.38563502 0.49604304] [0.12411261 0.67440557 0.50830552 0.67146303]] [[0.52552375 0.06349533 0.78590277 0.36544202] [0.6055249 0.317296 0.42477562 0.34462548] [0.55422445 0.68032915 0.93749125 0.89412014]]] {code} Tensorflow output is {code:python} print(sess.run(tf.contrib.seq2seq.hardmax(tf.Variable(data [[[0. 0. 0. 1.] [0. 1. 0. 0.] [0. 1. 0. 0.]] [[0. 0. 1. 0.] [1. 0. 0. 0.] [0. 0. 1. 0.]]] {code} However, your code gives {code:python} xn = mx.nd.array(data) xn_r = mx.nd.reshape(xn, shape=(2,12)) xn_e = mx.nd.eye(xn_r.shape[1], dtype=data.dtype)[mx.nd.argmax(xn_r, axis=1)] hardmax_output = mx.nd.reshape(xn_e, shape=xn.shape) print(hardmax_output) [[[0. 0. 0. 0.] [0. 1. 0. 0.] [0. 0. 0. 0.]] [[0. 1. 0. 0.] [0. 0. 0. 0.] [0. 0. 0. 0.]]] {code} That is, it finds the max across the first dimension (2 in [2,3,4]) But tensor flow on the other hand finds the max for all rows (being sequence to sequence based (for NLP applications)) What do you reckon? In that case I have implemented this Hardmax operator (similar to Tensorflow's) - https://github.com/apache/incubator-mxnet/pull/13083 Hence, In my opinion, to get equivalent output we should be doing {code:python} xn_r = mx.nd.reshape(data, shape=(6,4)) {code} instead of {code:python} xn_r = mx.nd.reshape(data, shape=(2,12)) {code} was (Author: bapac): [~anirudhacharya] It looks like the above code doesn't give the same result as Tensorflow does For the same input {code:python} import numpy as np data= np.random.rand(2,3,4) print(data) [[[0.18651888 0.29437149 0.45589573 0.55928469] [0.1951702 0.50387834 0.38563502 0.49604304] [0.12411261 0.67440557 0.50830552 0.67146303]] [[0.52552375 0.06349533 0.78590277 0.36544202] [0.6055249 0.317296 0.42477562 0.34462548] [0.55422445 0.68032915 0.93749125 0.89412014]]] {code} Tensorflow output is {code:python} print(sess.run(tf.contrib.seq2seq.hardmax(tf.Variable(data [[[0. 0. 0. 1.] [0. 1. 0. 0.] [0. 1. 0. 0.]] [[0. 0. 1. 0.] [1. 0. 0. 0.] [0. 0. 1. 0.]]] {code} However, your code gives {code:python} xn = mx.nd.array(data) xn_r = mx.nd.reshape(xn, shape=(2,12)) xn_e = mx.nd.eye(xn_r.shape[1], dtype=data.dtype)[mx.nd.argmax(xn_r, axis=1)] hardmax_output = mx.nd.reshape(xn_e, shape=xn.shape) print(hardmax_output) [[[0. 0. 0. 0.] [0. 1. 0. 0.] [0. 0. 0. 0.]] [[0. 1. 0. 0.] [0. 0. 0. 0.] [0. 0. 0. 0.]]] {code} That is, it finds the max across the first dimension (2 in [2,3,4]) But tensor flow on the other hand finds the max for all rows (being sequence to sequence based (for NLP applications)) What do you reckon? In that case I have implemented this Hardmax operator (similar to Tensorflow's) - https://github.com/apache/incubator-mxnet/pull/13083 > Hardmax > --- > > Key: MXNET-376 > URL: https://issues.apache.org/jira/browse/MXNET-376 > Project: Apache MXNet > Issue Type: Sub-task >Reporter: Hao Jin >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Comment Edited] (MXNET-376) Hardmax
[ https://issues.apache.org/jira/browse/MXNET-376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708069#comment-16708069 ] Chaitanya Prakash Bapat edited comment on MXNET-376 at 12/4/18 1:51 AM: [~anirudhacharya] It looks like the above code doesn't give the same result as Tensorflow does For the same input {code:python} import numpy as np data= np.random.rand(2,3,4) print(data) [[[0.18651888 0.29437149 0.45589573 0.55928469] [0.1951702 0.50387834 0.38563502 0.49604304] [0.12411261 0.67440557 0.50830552 0.67146303]] [[0.52552375 0.06349533 0.78590277 0.36544202] [0.6055249 0.317296 0.42477562 0.34462548] [0.55422445 0.68032915 0.93749125 0.89412014]]] {code} Tensorflow output is {code:python} print(sess.run(tf.contrib.seq2seq.hardmax(tf.Variable(data [[[0. 0. 0. 1.] [0. 1. 0. 0.] [0. 1. 0. 0.]] [[0. 0. 1. 0.] [1. 0. 0. 0.] [0. 0. 1. 0.]]] {code} However, your code gives {code:python} xn = mx.nd.array(data) xn_r = mx.nd.reshape(xn, shape=(2,12)) xn_e = mx.nd.eye(xn_r.shape[1], dtype=data.dtype)[mx.nd.argmax(xn_r, axis=1)] hardmax_output = mx.nd.reshape(xn_e, shape=xn.shape) print(hardmax_output) [[[0. 0. 0. 0.] [0. 1. 0. 0.] [0. 0. 0. 0.]] [[0. 1. 0. 0.] [0. 0. 0. 0.] [0. 0. 0. 0.]]] {code} That is, it finds the max across the first dimension (2 in [2,3,4]) But tensor flow on the other hand finds the max for all rows (being sequence to sequence based (for NLP applications)) What do you reckon? In that case I have implemented this Hardmax operator (similar to Tensorflow's) - https://github.com/apache/incubator-mxnet/pull/13083 was (Author: bapac): [~anirudhacharya] It looks like the above code doesn't give the same result as tensor flow does For the same input {code:python} import numpy as np data= np.random.rand(2,3,4) print(data) [[[0.18651888 0.29437149 0.45589573 0.55928469] [0.1951702 0.50387834 0.38563502 0.49604304] [0.12411261 0.67440557 0.50830552 0.67146303]] [[0.52552375 0.06349533 0.78590277 0.36544202] [0.6055249 0.317296 0.42477562 0.34462548] [0.55422445 0.68032915 0.93749125 0.89412014]]] {code} Tensorflow output is {code:python} print(sess.run(tf.contrib.seq2seq.hardmax(tf.Variable(data [[[0. 0. 0. 1.] [0. 1. 0. 0.] [0. 1. 0. 0.]] [[0. 0. 1. 0.] [1. 0. 0. 0.] [0. 0. 1. 0.]]] {code} However, your code gives {code:python} xn = mx.nd.array(data) xn_r = mx.nd.reshape(xn, shape=(2,12)) xn_e = mx.nd.eye(xn_r.shape[1], dtype=data.dtype)[mx.nd.argmax(xn_r, axis=1)] hardmax_output = mx.nd.reshape(xn_e, shape=xn.shape) print(hardmax_output) [[[0. 0. 0. 0.] [0. 1. 0. 0.] [0. 0. 0. 0.]] [[0. 1. 0. 0.] [0. 0. 0. 0.] [0. 0. 0. 0.]]] {code} That is, it finds the max across the first dimension (2 in [2,3,4]) But tensor flow on the other hand finds the max for all rows (being sequence to sequence based (for NLP applications)) What do you reckon? In that case I have implemented this Hardmax operator (similar to Tensorflow's) - https://github.com/apache/incubator-mxnet/pull/13083 > Hardmax > --- > > Key: MXNET-376 > URL: https://issues.apache.org/jira/browse/MXNET-376 > Project: Apache MXNet > Issue Type: Sub-task >Reporter: Hao Jin >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-376) Hardmax
[ https://issues.apache.org/jira/browse/MXNET-376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708069#comment-16708069 ] Chaitanya Prakash Bapat commented on MXNET-376: --- [~anirudhacharya] It looks like the above code doesn't give the same result as tensor flow does For the same input {code:python} import numpy as np data= np.random.rand(2,3,4) print(data) [[[0.18651888 0.29437149 0.45589573 0.55928469] [0.1951702 0.50387834 0.38563502 0.49604304] [0.12411261 0.67440557 0.50830552 0.67146303]] [[0.52552375 0.06349533 0.78590277 0.36544202] [0.6055249 0.317296 0.42477562 0.34462548] [0.55422445 0.68032915 0.93749125 0.89412014]]] {code} Tensorflow output is {code:python} print(sess.run(tf.contrib.seq2seq.hardmax(tf.Variable(data [[[0. 0. 0. 1.] [0. 1. 0. 0.] [0. 1. 0. 0.]] [[0. 0. 1. 0.] [1. 0. 0. 0.] [0. 0. 1. 0.]]] {code} However, your code gives {code:python} xn = mx.nd.array(data) xn_r = mx.nd.reshape(xn, shape=(2,12)) xn_e = mx.nd.eye(xn_r.shape[1], dtype=data.dtype)[mx.nd.argmax(xn_r, axis=1)] hardmax_output = mx.nd.reshape(xn_e, shape=xn.shape) print(hardmax_output) [[[0. 0. 0. 0.] [0. 1. 0. 0.] [0. 0. 0. 0.]] [[0. 1. 0. 0.] [0. 0. 0. 0.] [0. 0. 0. 0.]]] {code} That is, it finds the max across the first dimension (2 in [2,3,4]) But tensor flow on the other hand finds the max for all rows (being sequence to sequence based (for NLP applications)) What do you reckon? In that case I have implemented this Hardmax operator (similar to Tensorflow's) - https://github.com/apache/incubator-mxnet/pull/13083 > Hardmax > --- > > Key: MXNET-376 > URL: https://issues.apache.org/jira/browse/MXNET-376 > Project: Apache MXNet > Issue Type: Sub-task >Reporter: Hao Jin >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1200) simplify the LD_LIBRARY_PATH
[ https://issues.apache.org/jira/browse/MXNET-1200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Yuxi Hu updated MXNET-1200: --- Status: Done (was: In Progress) > simplify the LD_LIBRARY_PATH > > > Key: MXNET-1200 > URL: https://issues.apache.org/jira/browse/MXNET-1200 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1166) Fix and add unit tests in Horovod
[ https://issues.apache.org/jira/browse/MXNET-1166?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Yuxi Hu updated MXNET-1166: --- Status: In Progress (was: To Do) > Fix and add unit tests in Horovod > - > > Key: MXNET-1166 > URL: https://issues.apache.org/jira/browse/MXNET-1166 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1166) Fix and add unit tests in Horovod
[ https://issues.apache.org/jira/browse/MXNET-1166?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1166: Sprint: Daniel Bernoulli 2 > Fix and add unit tests in Horovod > - > > Key: MXNET-1166 > URL: https://issues.apache.org/jira/browse/MXNET-1166 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1166) Fix and add unit tests in Horovod
[ https://issues.apache.org/jira/browse/MXNET-1166?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan reassigned MXNET-1166: --- Assignee: Yuxi Hu > Fix and add unit tests in Horovod > - > > Key: MXNET-1166 > URL: https://issues.apache.org/jira/browse/MXNET-1166 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1246) Create benchmark AI test
[ https://issues.apache.org/jira/browse/MXNET-1246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan reassigned MXNET-1246: --- Assignee: Yuxi Hu > Create benchmark AI test > > > Key: MXNET-1246 > URL: https://issues.apache.org/jira/browse/MXNET-1246 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1200) simplify the LD_LIBRARY_PATH
[ https://issues.apache.org/jira/browse/MXNET-1200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan reassigned MXNET-1200: --- Assignee: Yuxi Hu (was: Lin Yuan) > simplify the LD_LIBRARY_PATH > > > Key: MXNET-1200 > URL: https://issues.apache.org/jira/browse/MXNET-1200 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1220) Add a mnist example
[ https://issues.apache.org/jira/browse/MXNET-1220?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1220: Sprint: Daniel Bernoulli, Daniel Bernoulli 2 (was: Daniel Bernoulli) > Add a mnist example > --- > > Key: MXNET-1220 > URL: https://issues.apache.org/jira/browse/MXNET-1220 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Lin Yuan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1246) Create benchmark AI test
[ https://issues.apache.org/jira/browse/MXNET-1246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1246: Sprint: Daniel Bernoulli 2 > Create benchmark AI test > > > Key: MXNET-1246 > URL: https://issues.apache.org/jira/browse/MXNET-1246 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1246) Create benchmark AI test
Lin Yuan created MXNET-1246: --- Summary: Create benchmark AI test Key: MXNET-1246 URL: https://issues.apache.org/jira/browse/MXNET-1246 Project: Apache MXNet Issue Type: Task Components: Apache MXNet Backend Reporter: Lin Yuan -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1110) Merge final API design to Horovod repository
[ https://issues.apache.org/jira/browse/MXNET-1110?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1110: Sprint: Daniel Bernoulli, Daniel Bernoulli 2 (was: Daniel Bernoulli) > Merge final API design to Horovod repository > > > Key: MXNET-1110 > URL: https://issues.apache.org/jira/browse/MXNET-1110 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Lin Yuan >Priority: Major > Time Spent: 4h > Remaining Estimate: 0h > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1200) simplify the LD_LIBRARY_PATH
[ https://issues.apache.org/jira/browse/MXNET-1200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1200: Sprint: Daniel Bernoulli, Daniel Bernoulli 2 (was: Daniel Bernoulli) > simplify the LD_LIBRARY_PATH > > > Key: MXNET-1200 > URL: https://issues.apache.org/jira/browse/MXNET-1200 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Lin Yuan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1219) Add a cloudformation template for launching multiple instances
[ https://issues.apache.org/jira/browse/MXNET-1219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan reassigned MXNET-1219: --- Assignee: Yuxi Hu (was: Lin Yuan) > Add a cloudformation template for launching multiple instances > -- > > Key: MXNET-1219 > URL: https://issues.apache.org/jira/browse/MXNET-1219 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Yuxi Hu >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1219) Add a cloudformation template for launching multiple instances
[ https://issues.apache.org/jira/browse/MXNET-1219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lin Yuan updated MXNET-1219: Assignee: Lin Yuan (was: Yuxi Hu) Status: In Progress (was: To Do) > Add a cloudformation template for launching multiple instances > -- > > Key: MXNET-1219 > URL: https://issues.apache.org/jira/browse/MXNET-1219 > Project: Apache MXNet > Issue Type: Task > Components: Apache MXNet Backend >Reporter: Lin Yuan >Assignee: Lin Yuan >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1245) Deploy Label Bot model into Production
[ https://issues.apache.org/jira/browse/MXNET-1245?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1245: --- Sprint: Scala/Java 10 (12/3-12/17) > Deploy Label Bot model into Production > -- > > Key: MXNET-1245 > URL: https://issues.apache.org/jira/browse/MXNET-1245 > Project: Apache MXNet > Issue Type: New Feature > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Harsh Patel >Priority: Major > > # put models into production > # code clean up > # optional: enable travis CI -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1245) Deploy Label Bot model into Production
Denis Davydenko created MXNET-1245: -- Summary: Deploy Label Bot model into Production Key: MXNET-1245 URL: https://issues.apache.org/jira/browse/MXNET-1245 Project: Apache MXNet Issue Type: New Feature Reporter: Denis Davydenko # put models into production # code clean up # optional: enable travis CI -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1244) Refactor JNI code for Scala
[ https://issues.apache.org/jira/browse/MXNET-1244?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1244: -- Assignee: Andrew Ayres (was: Piyush Ghai) > Refactor JNI code for Scala > --- > > Key: MXNET-1244 > URL: https://issues.apache.org/jira/browse/MXNET-1244 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Andrew Ayres >Priority: Major > > This refactoring should address number of github issues with Scala, make it > easier to debug and troubleshoot Scala code. Also as part of this story - > implement unit tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1245) Deploy Label Bot model into Production
[ https://issues.apache.org/jira/browse/MXNET-1245?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1245: -- Assignee: Harsh Patel > Deploy Label Bot model into Production > -- > > Key: MXNET-1245 > URL: https://issues.apache.org/jira/browse/MXNET-1245 > Project: Apache MXNet > Issue Type: New Feature > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Harsh Patel >Priority: Major > > # put models into production > # code clean up > # optional: enable travis CI -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1245) Deploy Label Bot model into Production
[ https://issues.apache.org/jira/browse/MXNET-1245?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1245: --- Component/s: Apache MXNet Java API > Deploy Label Bot model into Production > -- > > Key: MXNET-1245 > URL: https://issues.apache.org/jira/browse/MXNET-1245 > Project: Apache MXNet > Issue Type: New Feature > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Harsh Patel >Priority: Major > > # put models into production > # code clean up > # optional: enable travis CI -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1244) Refactor JNI code for Scala
[ https://issues.apache.org/jira/browse/MXNET-1244?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1244: -- Assignee: Piyush Ghai (was: Andrew Ayres) > Refactor JNI code for Scala > --- > > Key: MXNET-1244 > URL: https://issues.apache.org/jira/browse/MXNET-1244 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Piyush Ghai >Priority: Major > > This refactoring should address number of github issues with Scala, make it > easier to debug and troubleshoot Scala code. Also as part of this story - > implement unit tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1244) Refactor JNI code for Scala
[ https://issues.apache.org/jira/browse/MXNET-1244?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1244: -- Assignee: Andrew Ayres > Refactor JNI code for Scala > --- > > Key: MXNET-1244 > URL: https://issues.apache.org/jira/browse/MXNET-1244 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Andrew Ayres >Priority: Major > > This refactoring should address number of github issues with Scala, make it > easier to debug and troubleshoot Scala code. Also as part of this story - > implement unit tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1242) Introduce maven wrapper to Scala benchmarking scripts
[ https://issues.apache.org/jira/browse/MXNET-1242?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1242: -- Assignee: Piyush Ghai > Introduce maven wrapper to Scala benchmarking scripts > - > > Key: MXNET-1242 > URL: https://issues.apache.org/jira/browse/MXNET-1242 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Piyush Ghai >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1244) Refactor JNI code for Scala
Denis Davydenko created MXNET-1244: -- Summary: Refactor JNI code for Scala Key: MXNET-1244 URL: https://issues.apache.org/jira/browse/MXNET-1244 Project: Apache MXNet Issue Type: Improvement Components: Apache MXNet Scala API Reporter: Denis Davydenko This refactoring should address number of github issues with Scala, make it easier to debug and troubleshoot Scala code. Also as part of this story - implement unit tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1244) Refactor JNI code for Scala
[ https://issues.apache.org/jira/browse/MXNET-1244?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1244: --- Sprint: Scala/Java 10 (12/3-12/17) > Refactor JNI code for Scala > --- > > Key: MXNET-1244 > URL: https://issues.apache.org/jira/browse/MXNET-1244 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Andrew Ayres >Priority: Major > > This refactoring should address number of github issues with Scala, make it > easier to debug and troubleshoot Scala code. Also as part of this story - > implement unit tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1243) Introduce Float64/Long support into Scala/Java
Denis Davydenko created MXNET-1243: -- Summary: Introduce Float64/Long support into Scala/Java Key: MXNET-1243 URL: https://issues.apache.org/jira/browse/MXNET-1243 Project: Apache MXNet Issue Type: Improvement Components: Apache MXNet Java API, Apache MXNet Scala API Reporter: Denis Davydenko -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1243) Introduce Float64/Long support into Scala/Java
[ https://issues.apache.org/jira/browse/MXNET-1243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1243: -- Assignee: Piyush Ghai > Introduce Float64/Long support into Scala/Java > -- > > Key: MXNET-1243 > URL: https://issues.apache.org/jira/browse/MXNET-1243 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Piyush Ghai >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1243) Introduce Float64/Long support into Scala/Java
[ https://issues.apache.org/jira/browse/MXNET-1243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1243: --- Sprint: Scala/Java 10 (12/3-12/17) > Introduce Float64/Long support into Scala/Java > -- > > Key: MXNET-1243 > URL: https://issues.apache.org/jira/browse/MXNET-1243 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Piyush Ghai >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1242) Introduce maven wrapper to Scala benchmarking scripts
Denis Davydenko created MXNET-1242: -- Summary: Introduce maven wrapper to Scala benchmarking scripts Key: MXNET-1242 URL: https://issues.apache.org/jira/browse/MXNET-1242 Project: Apache MXNet Issue Type: Story Components: Apache MXNet Java API Reporter: Denis Davydenko -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1242) Introduce maven wrapper to Scala benchmarking scripts
[ https://issues.apache.org/jira/browse/MXNET-1242?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1242: --- Sprint: Scala/Java 10 (12/3-12/17) > Introduce maven wrapper to Scala benchmarking scripts > - > > Key: MXNET-1242 > URL: https://issues.apache.org/jira/browse/MXNET-1242 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Piyush Ghai >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1239) Investigate how to track amount of installs of Scala/Java package from Maven and incorporate that into tracked KPIs
[ https://issues.apache.org/jira/browse/MXNET-1239?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1239: --- Resolution: Fixed Status: Done (was: To Do) This is not possible. > Investigate how to track amount of installs of Scala/Java package from Maven > and incorporate that into tracked KPIs > --- > > Key: MXNET-1239 > URL: https://issues.apache.org/jira/browse/MXNET-1239 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Denis Davydenko >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1226) Update links to Maven Central repo in all Java API docs
[ https://issues.apache.org/jira/browse/MXNET-1226?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1226: --- Sprint: (was: Scala/Java 10 (12/3-12/17)) > Update links to Maven Central repo in all Java API docs > --- > > Key: MXNET-1226 > URL: https://issues.apache.org/jira/browse/MXNET-1226 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-1231) Help user make some as optional field
[ https://issues.apache.org/jira/browse/MXNET-1231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16707715#comment-16707715 ] Denis Davydenko commented on MXNET-1231: As part of this story please also add unit tests to verify code generated by macros. > Help user make some as optional field > - > > Key: MXNET-1231 > URL: https://issues.apache.org/jira/browse/MXNET-1231 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Qing Lan >Priority: Major > > Currently, the usage of some in api. is pretty awful, users > have to do Some(Symbol) and then .get to use the arithmetic method and then > wrapped up using another Some() to make it pass to the next layer. We should > figure out a way to simplify the process. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1231) Help user make some as optional field
[ https://issues.apache.org/jira/browse/MXNET-1231?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1231: -- Assignee: Qing Lan > Help user make some as optional field > - > > Key: MXNET-1231 > URL: https://issues.apache.org/jira/browse/MXNET-1231 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Qing Lan >Priority: Major > > Currently, the usage of some in api. is pretty awful, users > have to do Some(Symbol) and then .get to use the arithmetic method and then > wrapped up using another Some() to make it pass to the next layer. We should > figure out a way to simplify the process. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-715) Scalapkg: Automation of Build and Distribution
[ https://issues.apache.org/jira/browse/MXNET-715?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-715: -- Sprint: Scala/Java 1 (07/30-08/13), Scala/Java 2 (08/13-08/27), Scala/Java 4 (09/10-09/24), Scala/Java 5 (09/24-10/08), Scala/Java 6 (10/08 - 10/22), Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 1 (07/30-08/13), Scala/Java 2 (08/13-08/27), Scala/Java 4 (09/10-09/24), Scala/Java 5 (09/24-10/08), Scala/Java 6 (10/08 - 10/22), Scala/Java 9 (11/19-12/3)) > Scalapkg: Automation of Build and Distribution > -- > > Key: MXNET-715 > URL: https://issues.apache.org/jira/browse/MXNET-715 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Zach Kimberg >Priority: Major > > MXNet Scala API package should be easily distributable to users who would > like to use Scala API for inference or training. Scope of this story is to > provide users with complete package of MXNet engine, all its necessary > dependencies and ScalaAPI bindings. Scala API supports three flavors of > MXNet: MacOSX CPU, Linux CPU and Linux GPU, these flavors should be reflected > in corresponding packages. > As a result of this story users will be able to choose and download maven > package that would have everything needed for inference or training using > Scala API included. User will be able to start using MXNet with Scala API > bindings right away after that. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1225) Update Scala Install instructions
[ https://issues.apache.org/jira/browse/MXNET-1225?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1225: --- Sprint: Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 9 (11/19-12/3)) > Update Scala Install instructions > - > > Key: MXNET-1225 > URL: https://issues.apache.org/jira/browse/MXNET-1225 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Zach Kimberg >Priority: Major > Time Spent: 1h 40m > Remaining Estimate: 0h > > Current Scala Install instructions for building it from Source (not for Maven > download) are not very easy to follow, they require navigating to multiple > pages and figuring out how to build and stitch multiple pieces together to > make it work. Also there are multiple paths users could end up following in > order to build Scala from source. One of them could lead users to the trap > that we experienced while building Scala from source for performance testing. > The trap, in more details, pertains to instructions to build MXNet from > source where there is no copying params into config.mk file. As a result – > MXNet could be built with OpenBLAS but consequent Scala package build won’t > have config.mk file to use and thus won’t have same set of params that were > used during MXNet build (namely – it could miss USE_BLAS=OPENBLAS flag and > thus build Scala package without it, exactly the trap we experienced). Please > update Build from Source instructions to ensure this trap path is avoided. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1224) Improve MXNet Scala build
[ https://issues.apache.org/jira/browse/MXNET-1224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1224: --- Assignee: Denis Davydenko (was: Frank Liu) Status: In Progress (was: To Do) > Improve MXNet Scala build > - > > Key: MXNET-1224 > URL: https://issues.apache.org/jira/browse/MXNET-1224 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Denis Davydenko >Priority: Major > > Currently, Scala builds from makefile require all build flags to be inherited > from the Engine build. It should be as simple as linking libmxnet.so and > libmxnet.a since these two files are already embedded with all dependencies > required. > * first define why we are missing use_blas flag (and other flags) during > build of .so file from JNI > * second identify what is necessary in order to build .so file for Scala > package using the same flags as were used for original .so file of MXNet. It > could be just copying flags config and preserve building .so file for Scala > package, or it could be stop building .so file for Scala package code and use > original .so file of MXNet as the only .so file in the package. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-731) Memory Management in Scala binding
[ https://issues.apache.org/jira/browse/MXNET-731?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-731: -- Sprint: Scala/Java 1 (07/30-08/13), Scala/Java 2 (08/13-08/27), Scala/Java 4 (09/10-09/24), Scala/Java 5 (09/24-10/08), Scala/Java 6 (10/08 - 10/22), Scala/Java 7 (10/22-11/5), Scala/Java 8 (11/5 - 11/19), Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 1 (07/30-08/13), Scala/Java 2 (08/13-08/27), Scala/Java 4 (09/10-09/24), Scala/Java 5 (09/24-10/08), Scala/Java 6 (10/08 - 10/22), Scala/Java 7 (10/22-11/5), Scala/Java 8 (11/5 - 11/19), Scala/Java 9 (11/19-12/3)) > Memory Management in Scala binding > -- > > Key: MXNET-731 > URL: https://issues.apache.org/jira/browse/MXNET-731 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Scala API >Reporter: Denis Davydenko >Assignee: Naveen Swamy >Priority: Major > > Due to C++ nature of MXNet, Scala binding has to take care about memory > management and garbage collector for its users. Success criteria: Scala API > users should not have to take special care about disposing objects created by > MXNet and hence should not see memory leaks with their implementation of > MXNet using Scala API binding. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1224) Improve MXNet Scala build
[ https://issues.apache.org/jira/browse/MXNET-1224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1224: --- Sprint: Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 9 (11/19-12/3)) > Improve MXNet Scala build > - > > Key: MXNET-1224 > URL: https://issues.apache.org/jira/browse/MXNET-1224 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Frank Liu >Priority: Major > > Currently, Scala builds from makefile require all build flags to be inherited > from the Engine build. It should be as simple as linking libmxnet.so and > libmxnet.a since these two files are already embedded with all dependencies > required. > * first define why we are missing use_blas flag (and other flags) during > build of .so file from JNI > * second identify what is necessary in order to build .so file for Scala > package using the same flags as were used for original .so file of MXNet. It > could be just copying flags config and preserve building .so file for Scala > package, or it could be stop building .so file for Scala package code and use > original .so file of MXNet as the only .so file in the package. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1216) View Label Bot Logs across accounts
[ https://issues.apache.org/jira/browse/MXNET-1216?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1216: --- Sprint: Scala/Java 8 (11/5 - 11/19), Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 8 (11/5 - 11/19), Scala/Java 9 (11/19-12/3)) > View Label Bot Logs across accounts > --- > > Key: MXNET-1216 > URL: https://issues.apache.org/jira/browse/MXNET-1216 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Scala API >Reporter: Harsh Patel >Assignee: Harsh Patel >Priority: Major > > Label bot's logs are not viewable to other AWS accounts - enabling > cross-account access would allow for potential issues in the future to be > tracked down easily -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1208) Add Scala/Java Benchmark Memory Monitoring toolkit
[ https://issues.apache.org/jira/browse/MXNET-1208?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko updated MXNET-1208: --- Sprint: Scala/Java 9 (11/19-12/3), Scala/Java 10 (12/3-12/17) (was: Scala/Java 9 (11/19-12/3)) > Add Scala/Java Benchmark Memory Monitoring toolkit > -- > > Key: MXNET-1208 > URL: https://issues.apache.org/jira/browse/MXNET-1208 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Qing Lan >Priority: Major > > User should be able to use toolkit that is available in Scala/Java to monitor > the CPU/GPU RAM usage while running the benchmarks. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Assigned] (MXNET-1224) Improve MXNet Scala build
[ https://issues.apache.org/jira/browse/MXNET-1224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Denis Davydenko reassigned MXNET-1224: -- Assignee: Frank Liu (was: Denis Davydenko) > Improve MXNet Scala build > - > > Key: MXNET-1224 > URL: https://issues.apache.org/jira/browse/MXNET-1224 > Project: Apache MXNet > Issue Type: Improvement > Components: Apache MXNet Java API, Apache MXNet Scala API >Reporter: Qing Lan >Assignee: Frank Liu >Priority: Major > > Currently, Scala builds from makefile require all build flags to be inherited > from the Engine build. It should be as simple as linking libmxnet.so and > libmxnet.a since these two files are already embedded with all dependencies > required. > * first define why we are missing use_blas flag (and other flags) during > build of .so file from JNI > * second identify what is necessary in order to build .so file for Scala > package using the same flags as were used for original .so file of MXNet. It > could be just copying flags config and preserve building .so file for Scala > package, or it could be stop building .so file for Scala package code and use > original .so file of MXNet as the only .so file in the package. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1187) Create mxnet.io documentation for Java API
[ https://issues.apache.org/jira/browse/MXNET-1187?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Andrew Ayres updated MXNET-1187: Status: Done (was: In Review) > Create mxnet.io documentation for Java API > -- > > Key: MXNET-1187 > URL: https://issues.apache.org/jira/browse/MXNET-1187 > Project: Apache MXNet > Issue Type: Story > Components: Apache MXNet Java API >Reporter: Denis Davydenko >Assignee: Andrew Ayres >Priority: Critical > Time Spent: 12h 10m > Remaining Estimate: 0h > > Create mxnet.io documentation for Java API. Documents to be added are similar > to ones we have for Scala: > > - [http://mxnet.incubator.apache.org/api/scala/index.html] > - [http://mxnet.incubator.apache.org/tutorials/index.html#scala-tutorials] > - > [http://mxnet.incubator.apache.org/install/index.html?platform=Linux&language=Python&processor=CPU] > (Aaron leads this one but please provide him with instructions) > - home page cart about Java API > Documentation should include Java API functionality and examples/tutorials. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Updated] (MXNET-1154) Update Module.Scala to resolve Memory Leaks
[ https://issues.apache.org/jira/browse/MXNET-1154?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Andrew Ayres updated MXNET-1154: Status: In Progress (was: To Do) > Update Module.Scala to resolve Memory Leaks > --- > > Key: MXNET-1154 > URL: https://issues.apache.org/jira/browse/MXNET-1154 > Project: Apache MXNet > Issue Type: Sub-task > Components: Apache MXNet Scala API >Reporter: Naveen Swamy >Assignee: Andrew Ayres >Priority: Major > > Module.Scala throws a quite a few warnings for undisposed NDArrays. Although > they will be cleaned up at some point due to the Phantom Reference feature, > it could be cause OOM if the GC does not run frequently and also it is bad > experience to see a lot of Warnings with Stack traces. > Reach out to me offline and I will discuss some of the pointers I have found > and possible solutions. > This is one of the issues that a Clojure user reported, which is applicable > to both Scala and Java users. > https://discuss.mxnet.io/t/dealing-with-memory-leaks-ndarray/1767 > > > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Created] (MXNET-1241) Scala/Java Packages in Maven for CentOS
Naveen Swamy created MXNET-1241: --- Summary: Scala/Java Packages in Maven for CentOS Key: MXNET-1241 URL: https://issues.apache.org/jira/browse/MXNET-1241 Project: Apache MXNet Issue Type: Improvement Components: Apache MXNet Java API, Apache MXNet Scala API Reporter: Naveen Swamy Some users are interested to use the Scala package on CentOS. We should enable the new maven packages(with all dependencies) to be used on CentOS by adding suitable tests. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org