github-actions[bot] commented on PR #14898:
URL: https://github.com/apache/lucene/pull/14898#issuecomment-3034648288
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
nehemiaharchives opened a new pull request, #14898:
URL: https://github.com/apache/lucene/pull/14898
### Description
Fixes geX() to getX()
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to
github-actions[bot] commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3034170967
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
HUSTERGS commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3034034888
>I know we've had to be careful with the vector API at times as playing some
tricks may get even slower than scalar code on some hardware.
Yeah, I aggree with that too, I've seen
kaivalnp commented on PR #14874:
URL: https://github.com/apache/lucene/pull/14874#issuecomment-3033793168
Okay I ran things _slightly_ differently for 300d vectors. All runs are
without `-reindex`, but I'm deleting the index between runs of `main` and this
PR to create a fresh one
`m
kaivalnp commented on PR #14874:
URL: https://github.com/apache/lucene/pull/14874#issuecomment-3033770135
> if the results hold up for 768d vectors
Thanks @msokolov I quantized the Cohere 768d vectors by:
1. normalizing
2. scaling each dimension by 256
3. clipping between \[-1
vigyasharma commented on PR #14729:
URL: https://github.com/apache/lucene/pull/14729#issuecomment-3033747542
Thanks for the reviews and suggestions, everyone. The latest revision
incorporates all feedback so far.
It includes a `LateInteractionRescorer` for reranking top-N hits from
vigyasharma commented on code in PR #14729:
URL: https://github.com/apache/lucene/pull/14729#discussion_r2183767177
##
lucene/core/src/java/org/apache/lucene/search/LateInteractionFloatValuesSource.java:
##
@@ -0,0 +1,194 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
jpountz commented on PR #14827:
URL: https://github.com/apache/lucene/pull/14827#issuecomment-3033680943
I worked on improving the `ScorerUtil` test so that it would catch this
problem. It helped me find another problem. I pushed directly to the branch. I
think we're good now.
--
This is
vigyasharma commented on code in PR #14729:
URL: https://github.com/apache/lucene/pull/14729#discussion_r2183766249
##
lucene/core/src/java/org/apache/lucene/search/LateInteractionFloatValuesSource.java:
##
@@ -0,0 +1,194 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
jpountz commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3033677309
Wow, this is a big speedup! I'd like to get opinions on
`INT_FOR_DOUBLE_SPECIES` from folks who are familiar with the vector API, maybe
@uschindler or @ChrisHegarty. I know we've had to
github-actions[bot] commented on PR #14897:
URL: https://github.com/apache/lucene/pull/14897#issuecomment-3033659500
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
dweiss opened a new pull request, #14897:
URL: https://github.com/apache/lucene/pull/14897
(no comment)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-ma
github-actions[bot] commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3033043863
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
HUSTERGS commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3033023288
BTW, benchmark result showed above runs on a machine with
cpu
`Intel(R) Xeon(R) Platinum 8336C CPU @ 2.30GHz`
and flags
`fpu vme de pse tsc msr pae mce cx8 apic sep mtrr p
github-actions[bot] commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3033019807
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
kaivalnp commented on PR #14847:
URL: https://github.com/apache/lucene/pull/14847#issuecomment-3033015253
@mikemccand I stumbled upon a way to allocate a `long[]` in native memory
using a specific byte order (`LITTLE_ENDIAN`) -- which we use in a filtered
search (i.e. if an explicit filter
github-actions[bot] commented on PR #14847:
URL: https://github.com/apache/lucene/pull/14847#issuecomment-3033002093
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
github-actions[bot] commented on PR #14896:
URL: https://github.com/apache/lucene/pull/14896#issuecomment-3032970722
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
HUSTERGS opened a new pull request, #14896:
URL: https://github.com/apache/lucene/pull/14896
### Description
This PR is a follow-up of the
[comment](https://github.com/apache/lucene/pull/14827#issuecomment-3018852667)
from #14827 , trying to vectorize the `filterCompetitiveHits` func
jpountz commented on PR #14827:
URL: https://github.com/apache/lucene/pull/14827#issuecomment-3032847633
The seed did not reproduce for me, but I think I understand the problem. The
code assumes that if `a + b > c` then `a - ε + b <= c` (`ε > 0`). However this
is not true with floating-poin
jpountz commented on PR #14827:
URL: https://github.com/apache/lucene/pull/14827#issuecomment-3032803703
I'm taking a look now.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific commen
dweiss commented on PR #14827:
URL: https://github.com/apache/lucene/pull/14827#issuecomment-3032637198
Builds are failing after this has been merged?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go
thecoop merged PR #14873:
URL: https://github.com/apache/lucene/pull/14873
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apa
msokolov commented on PR #14892:
URL: https://github.com/apache/lucene/pull/14892#issuecomment-3032526748
BTW if you use a current checkout of `luceneutil` `knnPerfTest.py` will
produce an HTML file with a graph of the test run - would love to see that here
if possible?
--
This is a
benwtrent merged PR #14895:
URL: https://github.com/apache/lucene/pull/14895
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.a
thecoop opened a new pull request, #14895:
URL: https://github.com/apache/lucene/pull/14895
Backport #14797 to 10.x
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsub
romseygeek commented on issue #14756:
URL: https://github.com/apache/lucene/issues/14756#issuecomment-3032374484
The max of the clause count from the different variants sounds good to me.
And yes, doing the special handling inside the num clauses check Visitor feels
like the right place.
msokolov commented on PR #14874:
URL: https://github.com/apache/lucene/pull/14874#issuecomment-3032167570
I would be curious to see if the results hold up for 768d vectors. Maybe
300 not being divisible by 8 is problematic, although I don't believe it. But
... the results look so weird, I
msokolov commented on code in PR #14893:
URL: https://github.com/apache/lucene/pull/14893#discussion_r2182703723
##
lucene/core/src/java/org/apache/lucene/index/TieredMergePolicy.java:
##
@@ -139,9 +139,9 @@ public double getMaxMergedSegmentMB() {
* amplification factor will
msokolov commented on issue #13883:
URL: https://github.com/apache/lucene/issues/13883#issuecomment-3032121241
> Using the same thread pool for indexing and merging. This way if the
thread pool gets full of merges, this will naturally push back on indexing.
--
This is an automated message
jpountz merged PR #14827:
URL: https://github.com/apache/lucene/pull/14827
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apa
jpountz commented on PR #14827:
URL: https://github.com/apache/lucene/pull/14827#issuecomment-3031917436
Thank you, we may want to look into making this better in follow-ups, but
this looks good enough for me so I merged. :+1:
--
This is an automated message from the Apache Git Service.
T
jpountz commented on issue #13883:
URL: https://github.com/apache/lucene/issues/13883#issuecomment-3031906059
Yes, this is the same idea indeed! I had thumbs-up'ed it. :) But then the
discussion went back to iterating on CMS, which feels both more complicated
(N-N feedback loops between the
javanna commented on issue #14756:
URL: https://github.com/apache/lucene/issues/14756#issuecomment-3031555727
@romseygeek I vaguely remember chatting about this problem, and possible
solutions.
We don't know when doing the counting which of the query variants will be
used. Yet it se
dweiss commented on PR #14880:
URL: https://github.com/apache/lucene/pull/14880#issuecomment-3031419131
Thank you. I could reproduce the issue.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the
dweiss merged PR #14880:
URL: https://github.com/apache/lucene/pull/14880
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apac
37 matches
Mail list logo
