Skip to content

[SPARK-32907][ML] adaptively blockify instances - revert blockify gmm - #29782

Closed
zhengruifeng wants to merge 3 commits into
apache:masterfrom
zhengruifeng:unblockify_gmm
Closed

[SPARK-32907][ML] adaptively blockify instances - revert blockify gmm#29782
zhengruifeng wants to merge 3 commits into
apache:masterfrom
zhengruifeng:unblockify_gmm

Conversation

@zhengruifeng

Copy link
Copy Markdown
Contributor

What changes were proposed in this pull request?

revert blockify gmm

Why are the changes needed?

@WeichenXu123 and I thought we should use memory size instead of number of rows to blockify instance; then if a buffer's size is large and determined by number of rows, we should discard it.
In GMM, we found that the pre-allocated memory maybe too large and should be discarded:

@transient private lazy val auxiliaryPDFMat = DenseMatrix.zeros(blockSize, numFeatures)

We had some offline discuss and thought it is better to revert blockify GMM.

Does this PR introduce any user-facing change?

blockSize added in master branch will be removed

How was this patch tested?

existing testsuites

@SparkQA

Copy link
Copy Markdown

Test build #128806 has finished for PR 29782 at commit 3e616ef.

  • This patch fails Python style tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

@SparkQA

Copy link
Copy Markdown

Test build #128807 has finished for PR 29782 at commit d4d3b14.

  • This patch passes all tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

return self._set(aggregationDepth=value)

@since("3.1.0")
def setBlockSize(self, value):

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@srowensrowen left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems OK if this is just reverting something that went into 3.1.

@srowen

Copy link
Copy Markdown
Member

Feel free to merge

@WeichenXu123

Copy link
Copy Markdown
Contributor

Jenkins retest this

@WeichenXu123

Copy link
Copy Markdown
Contributor

@zhengruifeng

Copy link
Copy Markdown
ContributorAuthor

@WeichenXu123 Good catch! Thanks

@SparkQA

Copy link
Copy Markdown

Test build #129002 has finished for PR 29782 at commit 8d7113c.

  • This patch passes all tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

@zhengruifeng

Copy link
Copy Markdown
ContributorAuthor

Merged to master. Thanks all for reviewing!

@zhengruifeng
zhengruifeng deleted the unblockify_gmm branch September 23, 2020 07:56
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants

@zhengruifeng@SparkQA@srowen@WeichenXu123@HyukjinKwon