AGENT SKILL

artifacts-reports

Use mljar-supervised AutoML artifacts, reports, structured reports, explainability outputs, persistence, and retrain checks.

License 许可
MIT
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
artifacts-reports
license
MIT
metadata
{ "disco-role": "operating" }
description
Use mljar-supervised AutoML artifacts, reports, structured reports, explainability outputs, persistence, and retrain checks.
disable-model-invocation
true

02 资源文件 4

03 SKILL.md 正文

artifacts-reports

Use this sub-skill after an mljar-supervised AutoML run exists or when a task is primarily about saved run artifacts, report extraction, explainability files, reload/prediction behavior, or deciding whether a saved model needs retraining.

Route here for

  • Locating and interpreting an AutoML results_path directory, params.json, leaderboard.csv, top-level report files, and model subdirectories.
  • Loading a trained run with AutoML(results_path=...) or automl.load(path) and predicting after reload.
  • Calling get_leaderboard(...), report(), and report_structured(format="markdown"|"dict"|"json", model_name=...).
  • Explaining why SHAP, permutation-importance, tree, coefficient, or learning-curve artifacts are present or absent.
  • Checking preservation/move requirements for saved runs and using need_retrain(...) on new labeled data.

Start with these references

  1. Artifact layout and persistence for results_path, params.json, model folders, loading, leaderboard, and prediction-after-load checks.
  2. Reporting and explainability for report(), report_structured(...), report payload fields, model_name, and explainability artifacts.
  3. Troubleshooting for stale/conflicting result directories, wrong object loading, missing model files, report format mistakes, optional Graphviz/SHAP gaps, and moved directories.

For package assumptions and cross-cutting install/import issues, use the root references: package overview and package troubleshooting.

Boundaries

  • Route new training configuration, mode/algorithm/time-limit selection, validation design, and metric choice to training-core.
  • Route generated Mercury app workspaces and app(), local_app(), or publish_app() behavior to app-deployment.
  • Route the meaning of fairness metrics, sensitive-feature declarations, privileged/underprivileged groups, and mitigation tradeoffs to fairness-workflows. This sub-skill only explains where fairness fields appear in leaderboards and reports.
  • Do not treat backend learner files as the public persistence API. Standard users should load the full AutoML directory.

Bundled helper

Run scripts/structured_report_smoke.py to train a tiny synthetic model in a temporary or chosen output directory, reload it, call report_structured in multiple formats, optionally request a model-specific report, and print stable pass/fail signals.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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).

★ 351