Repository navigation
Anytime generator search: time_budget and per-stage timings - #49
Merged
Merged
Conversation
…stage Co-authored-by: Cursor <cursoragent@cursor.com>
…search Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com> # Conflicts: # CHANGELOG.md
Co-authored-by: Cursor <cursoragent@cursor.com> # Conflicts: # CHANGELOG.md
3 tasks done
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
generative_model,generative_model_from_bidirectionalandminimal_wyner_modeltaketime_budget(seconds). The pipeline checks the budget before each stage, between line-split seeds and between coupled-search sizes. Once it's spent, the remaining stages are logged asskipped: time budget exhaustedand the best verified bracket so far is returned, withGenerativeModelResult.budget_exhausted = True(also in the reliability summary). A single long stage can still overrun the budget.elapsedseconds since the search started, so slow stages show up in survey records.Motivation: in a random 4-state survey, 30 of the 39 cases that reached the full pipeline hit the 1500 s hard timeout and lost both bounds.
Test plan
time_budget=0on a survey process returns the epsilon-machine bracket withbudget_exhausted, skips the expensive stages, and timestamps every entry.test_generative_model.py; ruff; ty (no new warnings).Made with Cursor