Log warnings for numIterations * miniBatchFraction < 1.0 #13265
Closed
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 test this please |
| @@ -197,6 +197,11 @@ object GradientDescent extends Logging { | |||
| "< 1.0 can be unstable because of the stochasticity in sampling.") | |||
| } | |||
|
|
|||
| if (numIterations * miniBatchFraction < 1.0) { | |||
| logWarning("Not all examples will be used if numIterations * miniBatchFraction " + | |||
There was a problem hiding this comment.
Use string interpolation to add the actual values into this warning message
|
Test build #59233 has finished for PR 13265 at commit
|
|
OK merged to master/2.0 |
asfgit
pushed a commit
that referenced
this pull request
May 25, 2016
## What changes were proposed in this pull request? Add a warning log for the case that `numIterations * miniBatchFraction <1.0` during gradient descent. If the product of those two numbers is less than `1.0`, then not all training examples will be used during optimization. To put this concretely, suppose that `numExamples = 100`, `miniBatchFraction = 0.2` and `numIterations = 3`. Then, 3 iterations will occur each sampling approximately 6 examples each. In the best case, each of the 6 examples are unique; hence 18/100 examples are used. This may be counter-intuitive to most users and led to the issue during the development of another Spark ML model: zhengruifeng/spark-libFM#11. If a user actually does not require the training data set, it would be easier and more intuitive to use `RDD.sample`. ## How was this patch tested? `build/mvn -DskipTests clean package` build succeeds Author: Gio Borje <[email protected]> Closes #13265 from Hydrotoast/master. (cherry picked from commit 589cce9) Signed-off-by: Sean Owen <[email protected]>
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.
What changes were proposed in this pull request?
Add a warning log for the case that
numIterations * miniBatchFraction <1.0during gradient descent. If the product of those two numbers is less than1.0, then not all training examples will be used during optimization. To put this concretely, suppose thatnumExamples = 100,miniBatchFraction = 0.2andnumIterations = 3. Then, 3 iterations will occur each sampling approximately 6 examples each. In the best case, each of the 6 examples are unique; hence 18/100 examples are used.This may be counter-intuitive to most users and led to the issue during the development of another Spark ML model: zhengruifeng/spark-libFM#11. If a user actually does not require the training data set, it would be easier and more intuitive to use
RDD.sample.How was this patch tested?
build/mvn -DskipTests clean packagebuild succeeds