pipeline-basics
Install/import OpenPrompt, inspect the public pipeline APIs, and build the minimal PromptDataLoader/PromptForClassification/PromptForGeneration flow.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- pipeline-basics
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Install/import OpenPrompt, inspect the public pipeline APIs, and build the minimal PromptDataLoader/PromptForClassification/PromptForGeneration flow.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-reference.md4.5 KB
- 参考文档references/troubleshooting.md2.3 KB
- 参考文档references/workflows.md3.8 KB
- 脚本scripts/check_openprompt_install.py7.0 KB
03 SKILL.md 正文
Pipeline Basics
Use this sub-skill for the OpenPrompt package-usage surface: install checks, public exports, InputExample / InputFeatures, PromptDataLoader, PromptForClassification, PromptForGeneration, and load_plm.
Start here:
references/api-reference.mdfor verified signatures and return contracts.references/workflows.mdfor the minimal classification/generation quickstart and the no-download smoke.references/troubleshooting.mdfor dependency and wrapper failures.scripts/check_openprompt_install.pyfor the canonical import/API smoke.
Operating rules:
- Keep to the basic pipeline. Do not expand into prompt-template grammar, prompt-asset curation, dataset acquisition, or training-loop tuning.
- Prefer the smoke script before deeper debugging. It runs from a temporary working directory, imports the installed
openpromptpackage, avoids downloads, and checks the fake-wrapper loader path. PromptDataLoaderaccepts either a ready wrapper object or a wrapper class plus tokenizer; the wrapper class must expose the constructor kwargs filtered by the loader.- Route template/verbalizer design to
../template-verbalizer-design/, data/config work to../data-and-config-workflows/, and training/runtime issues to../training-and-generation/.
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).