Repository navigation
eval: commit the AutoSketch-vs-ASAP accuracy curves (#138) - #194
Merged
Merged
Conversation
The eval's inputs README lists the saturation study's accuracy grid (out_grid_1e7_cost) and targeted 1e9 points (out_1e9), but .gitignore's *.csv kept their six CSVs out of #138, so the eval couldn't be rerun from the repo. Add a narrow exception for that directory and commit them (about 1 MB): the curves the committed results were planned with. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FhcWJExZcmVqS6r6rEtjis
This was referenced Oct 10, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up to #138.
rqe-optimizer/results/autosketch-vs-asap-inputs/README.mdsays the eval's inputs include the saturation study's accuracy curves:saturation/out_grid_1e7_cost/: the accuracy grid to N = 1e7 with merge curves;saturation/out_1e9/: the targeted points past 1e7.But
.gitignorehas*.csv, so those six files never made it into #138. Onlyoptimizer_cost/rqe_atomic_costs.jsonwas committed, and the eval can't be rerun from the repo alone.Change
.gitignore: a narrow exception,!rqe-optimizer/results/autosketch-vs-asap-inputs/saturation/**/*.csv. CSVs stay ignored everywhere else.saturation.csv,saturation_curve.csv,saturation_merge_curve.csvin each run directory), about 1 MB in total. They are the files the committed results were planned with: rerunning the runner from these inputs reproduced the committed results exactly, as reported in eval: AutoSketch vs. ASAP on the synthetic mixed set (cost by use; frontier and SLA versions) #138.Check
cargo run --release -p rqe-optimizer --example small_problem -- --milp --saturation-dir rqe-optimizer/results/autosketch-vs-asap-inputs/saturationloads the curves and plans from the repo copy alone.small_problemrefuses to plan if a cost-table row disagrees with the curves, and here it doesn't refuse.The KLL memory in
optimizer_cost/rqe_atomic_costs.jsonpredates #192. Regenerating it, and rerunning the eval, is the separate follow-up listed in #192.🤖 Generated with Claude Code
https://claude.ai/code/session_01FhcWJExZcmVqS6r6rEtjis