원클릭으로
mylab-structure-tuning
Use when the main loop is idea -> implementation -> training -> eval -> analysis for a structural experiment round.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Use when the main loop is idea -> implementation -> training -> eval -> analysis for a structural experiment round.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when a mylab executor needs to start, poll, or inspect a long-running job. Prefer the documented `mylab tool start-job`, `wait-job`, and `tail-job` CLI flow instead of reading source code or inventing alternate entrypoints.
Use when the main loop is parameter generation -> batch experiments -> result collection -> comparison for a parameter experiment round.
Use when the main loop is parameter generation -> batch experiments -> result collection -> comparison for a parameter experiment round.
Use when the main loop is idea -> implementation -> training -> eval -> analysis for a structural experiment round.
| name | mylab-structure-tuning |
| description | Use when the main loop is idea -> implementation -> training -> eval -> analysis for a structural experiment round. |
| trial_kind | structure-tuning |
Use this skill when the user is trying new model structures, new training logic, new loss wiring, or new module combinations.
references/design.md when the code delta is complex.references/analysis.md when the analysis needs a dedicated structure.