Add SPR implementation to atari_100k lab #184
Open
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Outdated suggestions cannot be applied.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.
I'd like to add support for SPR to the Atari 100k lab project. My implementation is mostly siloed to minimize changes to the existing algorithms. SPR needs a custom model-based style replay buffer that returns subtrajectories, and a version of noisy nets that allows noise to be toggled on and off within a function, both of which are included.
This implementation performs a bit better than the original version; I found median 0.45 over 100 seeds, compared to 0.395 for the original.
@psc-g @agarwl