AGENT SKILL

offline_dataset_protocol

Validate fixed offline reinforcement learning transition datasets and summarize D4RL-style provenance before recovery experiments.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
offline_dataset_protocol
description
Validate fixed offline reinforcement learning transition datasets and summarize D4RL-style provenance before recovery experiments.

02 资源文件 2

03 SKILL.md 正文

Offline Dataset Protocol

Use this skill when a recovery or benchmark task needs to verify that training data is a fixed offline RL transition dataset. Do not use it to collect online rollouts, run simulators, or claim that synthetic proxy data is a real D4RL dataset.

Inputs

  • A JSON file containing either a list of transition records or an object with a transitions list.
  • Each transition must contain observation, action, reward, next_observation, terminal, and timeout.
  • Optional metadata fields such as source, domain, variant, and is_resource_derived.

Outputs

  • A JSON validation report with ok, counts, quality_tags, errors, and metadata.
  • The report preserves whether the dataset is synthetic proxy data, resource-derived data, or a real benchmark item.

Workflow

  1. Load the dataset JSON without mutating it.
  2. Confirm that every transition has the required keys.
  3. Coerce rewards to floats and terminal/timeout markers to booleans.
  4. Count transitions, terminal endings, timeout endings, and combined episode endings.
  5. Add the fixed_dataset tag and optional tags from metadata.
  6. Reject empty datasets or missing fields.

Validation

Run python scripts/validate_offline_dataset.py <dataset.json> for CLI validation. Run python -m pytest tests or the Distiller validate_skill_tree.py --run-tests command for deterministic tests.

Limitations

This skill validates dataset structure and provenance; it does not implement a full D4RL environment, simulator, or benchmark download.

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