advanced-workflows
Use Lazy Predict optional advanced workflows for tuning, search spaces, explainability, feature importance, SHAP, Optuna, FLAML, visualization pointers, and safe advanced dependency checks.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- advanced-workflows
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Use Lazy Predict optional advanced workflows for tuning, search spaces, explainability, feature importance, SHAP, Optuna, FLAML, visualization pointers, and safe advanced dependency checks.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/advanced-workflows.md3.9 KB
- 参考文档references/troubleshooting.md2.3 KB
- 脚本scripts/smoke_advanced.py3.5 KB
03 SKILL.md 正文
Advanced Workflows
Use this sub-skill when the task goes beyond a first Lazy Predict leaderboard: hyperparameter tuning, search-space inspection, permutation or SHAP explainability, FLAML/Optuna decisions, optional visualization dependencies, or advanced smoke checks.
Start here
-
Run a small core benchmark first with the supervised or time-series sub-skill. Advanced APIs usually need fitted models or score tables.
-
Check optional dependencies before promising SHAP, Optuna, FLAML, InterpretML, or plotting.
-
Keep trials, top-k counts, and timeouts bounded. Advanced workflows can be much more expensive than the initial Lazy Predict sweep.
-
Run the bundled advanced smoke helper for a quick environment check:
python scripts/smoke_advanced.py --json
What to read
- references/advanced-workflows.md covers supervised and time-series tuning, search-space registries, permutation/SHAP explainability, and dependency choices.
- references/troubleshooting.md covers missing optional packages, invalid tuning parameters, slow or empty tuning runs, SHAP compatibility, and unfitted model errors.
Common routes
- For
LazyClassifier(..., tune=True)orLazyRegressor(..., tune=True), read the supervised tuning section and then return to supervised-benchmarking for the base fit contract. - For
LazyForecaster(..., tune=True), read the time-series tuning section and then return to time-series-forecasting for forecasting data and metrics. - For feature importance after supervised fitting, prefer permutation
importance as the base dependency path; use SHAP only when the
shapextra is installed and the fitted model family is compatible. - For MLflow, Dask/PySpark, Spark, CLI, or GPU environment checks, use cli-and-integrations.
Do not use this sub-skill for
- A first quick model comparison without tuning or explanation.
- Repository maintenance, release automation, or documentation publishing.
- Full production AutoML orchestration outside Lazy Predict's exposed APIs.
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).