prehj/humandata-conf-labels-v3
실제 Franka 압축 가능성 라벨 (human_data v3) 각 순간의 모션을 더 빠르게 지나가도 되는지 VLM 으로 매긴 라벨. 만든 방법 판정기 gemini-3.8-flash (reasoning low), 온도 0 입력: 그 순간의 카메라 2장 + 지시문 + 계획된 액션에서 계산한 사실 출력: 문항을 1~5 등급으로. prompts/humandata_v3.txt 가 나간 전문 그대로다 SYSTEM 메시지를 쓰지 않는다. role: user 하나이고 가이던스가 본문 머리에 온다 청크 길이 16, stride 16. 아래 "중간 시점" 참고 게이트는 VLM 신뢰도 하나다. 접촉 열을 쓰지 않는다 실제 Franka 데모 세 벌 · 3,446 청크. 셋 다 같은 프롬프트로 만들었다 (전문·문항·부호·가중이 글자까지 같다). 데이터셋 청크 에피 fps 목적지 conf 평균 pnp_task 874 101… See the full description on the dataset page: https://huggingface.co/datasets/prehj/humandata-conf-labels-v3.
Conversations for this repository live on Hugging Face.
CoolFace shows imported repositories read-only. Posting into someone else’s repository from here would need an authorised integration and the account holder’s consent, so the link goes to the source instead.
Open discussions on Hugging Face