Issue 340: Simplify LatentDelay #388
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
seabbs
changed the title
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #388 +/- ##
==========================================
+ Coverage 93.23% 93.33% +0.10%
==========================================
Files 53 53
Lines 532 540 +8
==========================================
+ Hits 496 504 +8
Misses 36 36 ☔ View full report in Codecov by Sentry. |
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rt-without-renewal/Rt-without-renewalJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rt-without-renewal/Rt-without-renewalJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rt-without-renewal/Rt-without-renewalJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
|
|
Benchmarks are a bit noisy (we need to look at tuning down the statistical threshold for flagging) but appears to show a clear speed up for ascertainment and latent delay methods. However, there is no apparent speed up for compiled reverse auto-diff. The ascertainment result is a bit of a poser given that we haven't changed anything related to that model here and the benchmarks aren't related: Noting that I still see gradient issues in the benchmarks (but still not fixed Tova source) so this doesn't close the target issue even if it speeds things up a little. |
|
Thinking about it I assume the big variance in the ascertainment model is related to #357. |
|
@SamuelBrand1 thoughts on this? |
Urgh... I hit edit rather than quote reply... I've edited it back now and point above stands and is consistent with @seabbs views here. |
|
Yes agree memoization or a turing side resolution to this would be preferable longer term |
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.
This PR targets but may not close #340. It may also help with #357.
It simplifies the
LatentDelayapproach to use a dot product-based method rather than the current sparse matrix approach. It keeps the sparse matrix constructor as this may be useful in #386 (and generally changes here are temporary as #386 will likely improve this implementation substantially.I initially implemented this as a simple map but then decided to implement using the
accumulate_scanframework as a test. That resulted in a bit more code overhead and complexity but maybe worth it for consistency. It definitely works in this setting though and confirms my belief that it will work for composable modelling in #385