data-preparation
Validate and explain TimeMixer dataset layouts, file names, splits, date and frequency features, and benchmark-specific data conventions.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- data-preparation
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Validate and explain TimeMixer dataset layouts, file names, splits, date and frequency features, and benchmark-specific data conventions.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/data-formats.md5.9 KB
- 参考文档references/troubleshooting.md4.5 KB
- 脚本scripts/validate_timemixer_data.py20.3 KB
03 SKILL.md 正文
TimeMixer Data Preparation
Use this sub-skill for dataset layout questions, file checks, and small validation tasks.
Use when
- a custom CSV must be checked for
date, target, and numeric feature columns - an ETT, M4, PEMS, Solar, anomaly, or UEA dataset needs layout confirmation
- split sizes, windowability, or time-feature conventions need to be explained
- a future agent needs a safe validator instead of a training or download run
Do not use when
- the task is about command construction for forecasting experiments
- the task is about model tensor shapes, embeddings, or architecture internals
- the task requires dataset downloads, benchmark training, or native example execution
What this sub-skill covers
- dataset-key to loader mapping
- required file names and column conventions
- split behavior and window-length implications
- date parsing,
freqhandling, and generated time features - benchmark-specific quirks for custom CSV, M4, PEMS, Solar, anomaly CSV/NPY, and UEA
.tsfiles
Bundled assets
references/data-formats.mdreferences/troubleshooting.mdscripts/validate_timemixer_data.py
Quick workflow
- Read
references/data-formats.mdfor the expected layout. - Run
python scripts/validate_timemixer_data.py --helpto inspect supported checks. - Validate the candidate dataset tree with the bundled script.
- If the issue is actually about forecast commands or model internals, hand off to the matching sub-skill.
Routing
- Forecast command assembly and benchmark recipe questions:
forecasting-experiments - Model classes, tensor shapes, and architecture behavior:
model-architecture - Imputation, anomaly logic, or classification workflow details:
universal-tasks
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).