exp(e31): dppo-policy learns Hopper and Walker2d under DPPO in 300 iterations - #75
Merged
Merged
Conversation
E28's dppo-hopper and dppo-walker, unchanged but three times as long (1,228,800 steps). Status rule as E28's: iterations 291-300 at least 500 on 2 of 3. P2 Hopper learns; P3 Walker2d learns, expected falsified.
…erations Mean of iterations 291-300: Hopper 873 / 961 / 812, Walker2d 528 / 622 / 585, 3 of 3 past the bar of 500 on both. P1, P2 and P3 hold; I had expected P3 falsified from E28's flat Walker2d, which E31's first hundred iterations did not repeat (448 / 436 / 393 at 91-100 against E28's 260 / 292 / 274, same code path, first iterations identical).
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
E31, pre-registered at f8ffec0 before any run: E28's two rising
dppo-policy· DPPO cells, unchanged, run to 300 iterations (1,228,800 steps per seed) against the same bar.dppo-policy· DPPO · Hopperdppo-policy· DPPO · Walker2dguangzhao.np.asarray), and every first iteration matches E28's exactly; runs are not bit-reproducible after the first update (E20). FINDINGS says so rather than reading a plateau into one run.Two cells of the coverage figure change from "runs, rising" to "learns"; the figure itself is updated separately, with clips recorded from these checkpoints.
Checkpoints and tensorboards stay on
guangzhao; the logs,verdicts.txtandsummary.tsvare here.