forecasting-workflows
Use Darts core forecasting workflows, model selection, covariates, probabilistic prediction, and fit/predict validation.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- forecasting-workflows
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Use Darts core forecasting workflows, model selection, covariates, probabilistic prediction, and fit/predict validation.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/api-reference.md1.9 KB
- 参考文档references/model-selection.md2.8 KB
- 参考文档references/troubleshooting.md2.0 KB
- 参考文档references/workflows.md2.5 KB
- 脚本scripts/forecasting_smoke.py1.8 KB
03 SKILL.md 正文
Forecasting workflows
Use this sub-skill for Darts non-neural/core forecasting: model selection, fit()/predict(), covariate-capable regression/global models, probabilistic forecasts, historical forecasts/backtesting overview, and forecast troubleshooting.
Read first
references/model-selection.mdfor model-family choices, optional dependency boundaries, and covariate support.references/workflows.mdfor baseline/core forecast, probabilistic forecast, lagged regression covariates, and validation patterns.references/api-reference.mdfor fit/predict conventions and common model constructors.references/troubleshooting.mdfor unsupported covariates, wrong horizon, probabilistic output, and optional dependency errors.scripts/forecasting_smoke.pyfor a tiny baseline + ExponentialSmoothing smoke.
Route by task
- First forecast or baseline: start with
NaiveSeasonal,NaiveDrift, orExponentialSmoothingbefore complex models. - No torch allowed: stay in this sub-skill and use core/statistical/regression models; do not route to neural models.
- Covariates with lagged/global models: use
LinearRegressionModelor other regression/global models withlags,lags_past_covariates, and/orlags_future_covariates. - Model rejects covariate arguments: do not force them. Switch to a model family that supports the requested covariate type.
- Probabilistic forecast: verify the model supports samples/likelihood/quantiles, call
predict(..., num_samples=...)when supported, then route metric details to../evaluation-and-explainability/. - Neural model,
darts[torch], GPU, checkpoint, or foundation wrapper: route to../torch-and-foundation-models/.
Safe check
python scripts/forecasting_smoke.py --compact
The smoke uses generated monthly data, validates forecast length and stochastic samples, and does not read repo examples or notebooks.
Boundaries
This sub-skill does not own Darts data construction, preprocessing/covariate generation, PyTorch trainer configuration, anomaly detection, or final metric interpretation. Cross-link instead of duplicating those workflows.
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).