← Robot Manipulation

Papers with Code · CC BY-SA 4.0

RLBench

아카이브 평가 18건 · 마지막 적재/변경 2026-10-30

원본 데이터셋 링크 1

평가 기록

과거 수집 기록이며 최신 순위가 아닙니다. 데이터셋 버전·split·실험 조건과 지표 방향이 검증되지 않은 결과는 직접 비교할 수 없습니다. 날짜는 논문 날짜이며 최신 평가일을 뜻하지 않습니다.

표가 화면보다 넓으면 좌우로 스크롤하여 논문과 구현 링크를 확인하세요.

원본 기록 순서 · 순위 미부여
모델	원본 지표·값	논문	구현
EquAct	Succ. Rate (18 tasks, 100 demo/task) 89.4 Inference Speed (fps) 1.4 Input Image Size 256 Training Time (A100 x hour) 240 Succ. Rate (18 tasks, 10 demo/task) 60.1	EquAct: An SE(3)-Equivariant Multi-Task Transformer for Open-Loop Robotic Manipulation 2025-05-27	연결 없음
SAM2Act	Succ. Rate (18 tasks, 100 demo/task) 86.8 Input Image Size 128 Training Time (A100 x hour) 200	SAM2Act: Integrating Visual Foundation Model with A Memory Architecture for Robotic Manipulation 2025-01-30	코드 1
ARP+	Succ. Rate (18 tasks, 100 demo/task) 84.9	Autoregressive Action Sequence Learning for Robotic Manipulation 2024-10-04	코드 1
3D-LOTUS	Succ. Rate (18 tasks, 100 demo/task) 83.1 Training Time (V100 x 8 x day) 0.28 Inference Speed (fps) 9.5 Input Image Size 256 Training Time (A100 x hour) 40	Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policy 2024-10-02	코드 1
RVT-2	Succ. Rate (18 tasks, 100 demo/task) 81.4 Training Time (V100 x 8 x day) 0.83 Inference Speed (fps) 20.6 Input Image Size 128	RVT-2: Learning Precise Manipulation from Few Demonstrations 2024-06-12	코드 1
3D Diffuser Actor	Succ. Rate (18 tasks, 100 demo/task) 81.3 Training Time (V100 x 8 x day) 8 Input Image Size 256 Training Time (A100 x hour) 936	3D Diffuser Actor: Policy Diffusion with 3D Scene Representations 2024-02-18	코드 1
Mini Diffuser	Succ. Rate (18 tasks, 100 demo/task) 77.6 Training Time (V100 x 8 x day) 0.24 Input Image Size 256 Training Time (A100 x hour) 24	Mini Diffuser: Fast Multi-task Diffusion Policy Training Using Two-level Mini-batches 2025-05-14	코드 1
SAM-E	Succ. Rate (18 tasks, 100 demo/task) 70.6	SAM-E: Leveraging Visual Foundation Model with Sequence Imitation for Embodied Manipulation 2024-05-30	연결 없음
Act3D	Succ. Rate (18 tasks, 100 demo/task) 65 Training Time (V100 x 8 x day) 5 Input Image Size 256 Succ. Rate (18 tasks, 10 demo/task) 48	Act3D: 3D Feature Field Transformers for Multi-Task Robotic Manipulation 2023-06-30	코드 1 · 코드 2
RVT	Succ. Rate (18 tasks, 100 demo/task) 62.9 Training Time (V100 x 8 x day) 1 Inference Speed (fps) 11.6 Input Image Size 128	RVT: Robotic View Transformer for 3D Object Manipulation 2023-06-26	코드 1
PerAct (Evaluated in RVT)	Succ. Rate (18 tasks, 100 demo/task) 49.4 Training Time (V100 x 8 x day) 16 Inference Speed (fps) 4.9 Input Image Size 128	Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation 2022-09-12	코드 1
PolarNet	Succ. Rate (18 tasks, 100 demo/task) 46.4 Training Time (V100 x 8 x day) 5 Succ. Rate (10 tasks, 100 demos/task) 89.8 Succ. Rate (74 tasks, 100 demos/task) 60.3 Input Image Size 128	PolarNet: 3D Point Clouds for Language-Guided Robotic Manipulation 2023-09-27	코드 1
Hiveformer	Succ. Rate (18 tasks, 100 demo/task) 45.3 Succ. Rate (10 tasks, 100 demos/task) 83.3	Instruction-driven history-aware policies for robotic manipulations 2022-09-11	코드 1 · 코드 2
PerAct	Succ. Rate (18 tasks, 100 demo/task) 42.7 Training Time (V100 x 8 x day) 16 Input Image Size 128 Succ. Rate (18 tasks, 10 demo/task) 30	Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation 2022-09-12	코드 1
C2FARM-BC (Evaluated in PerAct)	Succ. Rate (18 tasks, 100 demo/task) 20.1 Input Image Size 128	Coarse-to-Fine Q-attention: Efficient Learning for Visual Robotic Manipulation via Discretisation 2021-06-23	코드 1
Image-BC VIT	Succ. Rate (18 tasks, 100 demo/task) 1.3 Input Image Size 128	Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation 2022-09-12	코드 1
Image-BC CNN	Succ. Rate (18 tasks, 100 demo/task) 1.3 Input Image Size 128	Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation 2022-09-12	코드 1
Auto-λ	Succ. Rate (10 tasks, 100 demos/task) 69.3	Auto-Lambda: Disentangling Dynamic Task Relationships 2022-02-07	코드 1