bounded_recovery_experiment
Run a bounded CMA-ES proxy experiment that validates sampling, selection, CSA, and covariance adaptation evidence.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- bounded_recovery_experiment
- description
- Run a bounded CMA-ES proxy experiment that validates sampling, selection, CSA, and covariance adaptation evidence.
02 资源文件 2
03 SKILL.md 正文
Bounded CMA-ES Recovery Experiment
Use this skill when full-scale paper reproduction is unavailable but soft-mode recovery permits a declared, executable, mechanism-faithful proxy. It should call the parameter, sampling, and adaptation skills rather than duplicate their logic silently.
Inputs
- Paths to generated CMA-ES module skills.
- Attempt directory with module plan and runtime handoff.
- A deterministic objective specification and generation budget.
Outputs
- Recovery metrics, generated objective item, optimization trace, skill invocation evidence, and mechanism checks.
Workflow
- Load the module plan target and runtime handoff.
- Construct a small rotated ellipsoid objective as a black-box callable.
- Initialize CMA-ES defaults, sample/rank offspring, update mean, CSA, and covariance for a bounded number of generations.
- Save before/after losses, parameter changes, covariance diagnostics, and source-boundary evidence.
- Run the Distiller recovery experiment gate.
Validation
Run the recovery harness and validate the attempt with validate_recovery_experiment.py.
Limitations
The output is a reduced/proxy recovery, not a full benchmark claim from the paper.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).