Fix priority sampling #57
Merged
Conversation
This file contains hidden or 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
Codecov Report
@@ Coverage Diff @@
## main #57 +/- ##
==========================================
+ Coverage 73.54% 74.50% +0.96%
==========================================
Files 15 15
Lines 756 765 +9
==========================================
+ Hits 556 570 +14
+ Misses 200 195 -5
... and 1 file with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
jeremiahpslewis
approved these changes
Sep 13, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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 fixes the issue of the unlikely but not impossible sampling of invalid steps with prioritized replay buffers.
Instead of relying on the internal sum_tree, I moved to the FrequencyWeights of StatsBase. SumTree is still used to maintain and update the priorities, but for sampling it was much simpler to use StatsBase (also this is a mature package in the Julian ecosystem).
As far as I know, Frequency Sampling weights the elements identically to the current sumtree sampling. I do not know about the efficiency but I did not observe a slow down in the experiments.
This PR relies on #56 so it must be merged afterward.