AbstractDataSetIterator performance tweak #2823
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Jenkins Build Triggered |
|
Jenkins Build Started |
|
Build finished. 549 tests run, 13 skipped, 460 failed. |
|
Refer to this link for build results (access rights to CI server needed): Build result: FAILURE[...truncated 445.14 KB...] at hudson.maven.MavenModuleSetBuild.run(MavenModuleSetBuild.java:544) at hudson.model.ResourceController.execute(ResourceController.java:98) at hudson.model.Executor.run(Executor.java:405)Caused by: java.io.FileNotFoundException: https://api.github.com/repos/deeplearning4j/deeplearning4j/statuses/305e7289264eb8eb35a38bf3e2915037a83908b0 at sun.reflect.GeneratedConstructorAccessor46.newInstance(Unknown Source) at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45) at java.lang.reflect.Constructor.newInstance(Constructor.java:423) at sun.net.www.protocol.http.HttpURLConnection$10.run(HttpURLConnection.java:1890) at sun.net.www.protocol.http.HttpURLConnection$10.run(HttpURLConnection.java:1885) at java.security.AccessController.doPrivileged(Native Method) at sun.net.www.protocol.http.HttpURLConnection.getChainedException(HttpURLConnection.java:1884) at sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1457) at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1441) at sun.net.www.protocol.https.HttpsURLConnectionImpl.getInputStream(HttpsURLConnectionImpl.java:254) at org.kohsuke.github.Requester.parse(Requester.java:535) at org.kohsuke.github.Requester._to(Requester.java:262) ... 12 moreCaused by: java.io.FileNotFoundException: https://api.github.com/repos/deeplearning4j/deeplearning4j/statuses/305e7289264eb8eb35a38bf3e2915037a83908b0 at sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1836) at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1441) at java.net.HttpURLConnection.getResponseCode(HttpURLConnection.java:480) at sun.net.www.protocol.https.HttpsURLConnectionImpl.getResponseCode(HttpsURLConnectionImpl.java:338) at org.kohsuke.github.Requester.parse(Requester.java:525) ... 13 more |
|
Looks great, thanks! |
AlexDBlack
added a commit
that referenced
this pull request
May 21, 2018
* Additional workspace method * NDArrayFactory: pullRows with result array * Also add new method to sparse factories * Additional debug info in workspace exceptions * Add MemoryWorkspace.setPreviousWorkspace * Remove condition in BaseNDArray.leverageTo * Move WorkspaceMgr to ND4J * Javadoc * Add WorkspaceMgr.hasConfiguration(T) * #2639 Workspace validation on leverage, detach, dup, etc * Add WorkspaceMgr.assertOpen * Additional validation for BaseWorkspaceMgr.notifyScopeBorrowed * Additional Preconditions.checkArgument methods
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The change is too call
Nd4j.create()once instead of multipleNd4j.create()followed byNd4j.vstack(). This tweak gives me up to 10x performance boost - depending on the batch size and the number of features/labels. It's an important improvement for in-memory datasets because these iterators are called during training all the time.