/벤치마크, 평가, 보정 /jev-benchmarks

jev-benchmarks

타입 지정 결정 모델을 위한 확률 인지 평가입니다: 보정, 선택적 위험, 지연 시간을 재현 가능하게 다룹니다.

이 페이지는 영문 목록에서 생성했습니다.

readme에서

jev-benchmarks Probability-aware evaluation for typed decision models. jev-benchmarks measures more than whether a model selects the right label. It evaluates whether the reported probabilities are calibrated enough to support automation, how much work can be accepted at a fixed error budget, what resources each decision uses, and how long it takes end to …

상세

섹션
벤치마크, 평가, 보정
owner
AbdelStark
스타
12
포크
0
최근 커밋
2026-09-17
언어
Python
라이선스
Apache-2.0

어떤 프로젝트인가

형태
데이터셋 또는 벤치마크
호스트 에이전트
단독 실행
대상
연구자
성숙도
문서
실측치를 보고합니다

이럴 때 유용합니다

가장 가까운 용도