natural_instruction_schema
Normalize crowdsourcing task instructions into validated Natural Instructions records for cross-task generalization experiments.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- natural_instruction_schema
- description
- Normalize crowdsourcing task instructions into validated Natural Instructions records for cross-task generalization experiments.
02 资源文件 2
03 SKILL.md 正文
Natural Instruction Schema
Use this skill when a recovery or benchmark harness needs to turn task templates into the structured Natural Instructions format from the paper. Do not use it to solve task instances directly or to mix train and evaluation examples.
Inputs
task_id,dataset, andcategorystrings.- Instruction fields:
definition,prompt, optionalthings_to_avoid, optionalemphasis, positive examples, negative examples, and instances. - Examples and instances represented as dictionaries with
inputandoutputstrings.
Outputs
- A JSON-compatible task record with normalized fields.
- Validation errors for missing task identity, absent task-defining text, or missing instances.
Workflow
- Read raw crowdsourcing-template content and split it into minimal subtask fields before calling the script.
- Call
scripts/natural_instruction_schema.pyor importbuild_instruction_record. - Keep examples separate from evaluation instances; examples are instructional content, not predictions.
- Preserve empty negative examples when the source task lacks them, as with QASC in the paper.
- Save the resulting record for split and encoding modules.
Validation
Run python tests/test_natural_instruction_schema.py or validate the whole skill with validate_skill_tree.py --run-tests.
Limitations
This skill does not fetch Natural Instructions data or infer missing instruction text. It only normalizes content supplied by the caller.
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).