Repository navigation
Measure recording counts through installed simulation qualification - #1
Merged
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Training candidates previously had one-scene evidence and no reproducible way to compare recording counts against installed execution. Add
dvidia-training-benchmark: nested whole-group subsets keep development/test footage fixed; protocol and all candidate bytes freeze before scene evaluation; native installation, exact-scene qualification and matched open-jaw controls retain failures and scoped timing.The published CPU pilot compares 2/4/7 groups across eight nominal and four actuator-stress layouts. Nominal qualifications are 8/8, 7/8 and 8/8; stress qualifications are 3/4 for each, with no successful open-jaw controls. Exact source-layout exclusions are bound to native declarations. This is one composition order, with privileged simulator observations and authored gripper/phase logic, not a minimum video count or human-video/hardware transfer claim.
Validation: 131 local native-enabled tests pass, 17 focused benchmark tests pass, isolated package build and installed entrypoint help pass. The actual offline benchmark completes all 84 teacher/student/control episodes without execution errors. Protocol, models, candidate capsules, results and an evidence archive manifest are included; source runtime/qualification files are unchanged. CI adds base benchmark tests and entrypoint checks across the existing platform matrix.
Also records anonymous verification of the already-public Hugging Face pilot artifacts. Full native traces/replays will be attached as a separate prerelease evidence asset after merge; this does not replace frozen v0.1.0 assets.