Papers with Code 아카이브
Vision-Language Navigation
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
- 하위 태스크
- 1
- 평가표
- 1
- 보존된 평가 기록
- 3
Vision-language navigation (VLN) is the task of navigating an embodied agent to carry out natural language instructions inside real 3D environments.
<span style="color:grey; opacity: 0.6">( Image credit: Learning to Navigate Unseen Environments: Back Translation with Environmental Dropout )</span>
코드
관련 논문과 평가 결과에 연결된 구현입니다. 출처에서 확인된 저자 구현에는 공식 코드 표시가 있습니다.
https://github.com/DriveVLA/OpenDriveVLA
공식 코드OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model
https://github.com/HCI-LMC/VLN-SUSA
공식 코드https://github.com/HanqingWangAI/Active_VLN
공식 코드Active Visual Information Gathering for Vision-Language Navigation
https://github.com/HanqingWangAI/SSM-VLN
공식 코드https://github.com/Homagn/MOVILAN
공식 코드https://github.com/Kelym/FAST
공식 코드Tactical Rewind: Self-Correction via Backtracking in Vision-and-Language Navigation
https://github.com/LukeLIN-web/VOTE
공식 코드VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting
https://github.com/ManifoldRG/MultiNet
공식 코드Benchmarking Vision, Language, & Action Models on Robotic Learning Tasks
https://github.com/OpenHelix-robot/OpenHelix
공식 코드https://github.com/Physical-Intelligence/real-time-chunking-kinetix
공식 코드https://github.com/RoboDita/Dita
공식 코드https://github.com/Smart-Wheelchair-RRC/o3d-sim
공식 코드Open-Set 3D Semantic Instance Maps for Vision Language Navigation -- O3D-SIM
https://github.com/Zhangwenyao1/DreamVLA
공식 코드DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge
https://github.com/aburns4/MoTIF
공식 코드A Dataset for Interactive Vision-Language Navigation with Unknown Command Feasibility
https://github.com/ahydchh/impromptu-vla
공식 코드Impromptu VLA: Open Weights and Open Data for Driving Vision-Language-Action Models
https://github.com/airsplay/R2R-EnvDrop
공식 코드Learning to Navigate Unseen Environments: Back Translation with Environmental Dropout
https://github.com/alibaba-damo-academy/worldvla
공식 코드https://github.com/allenxuuu/disco
공식 코드DISCO: Embodied Navigation and Interaction via Differentiable Scene Semantics and Dual-level Control
https://github.com/aoqunjin/awesome-vla-post-training
공식 코드Parallels Between VLA Model Post-Training and Human Motor Learning: Progress, Challenges, and Trends
https://github.com/cccedric/conrft
공식 코드ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policy
https://github.com/chenjinyubuaa/sevol
공식 코드Reinforced Structured State-Evolution for Vision-Language Navigation
https://github.com/chihyaoma/regretful-agent
공식 코드The Regretful Navigation Agent for Vision-and-Language Navigation
https://github.com/chihyaoma/selfmonitoring-agent
공식 코드Self-Monitoring Navigation Agent via Auxiliary Progress Estimation
https://github.com/cseeyangchen/c2vl
공식 코드https://github.com/defaultrui/bev-scene-graph
공식 코드https://github.com/defaultrui/vln-ver
공식 코드Volumetric Environment Representation for Vision-Language Navigation
https://github.com/eliotjones1/robogcg
공식 코드Adversarial Attacks on Robotic Vision Language Action Models
https://github.com/embodied-generalist/embodied-generalist
공식 코드https://github.com/expectorlin/DR-Attacker
공식 코드Adversarial Reinforced Instruction Attacker for Robust Vision-Language Navigation
https://github.com/expectorlin/console
공식 코드Correctable Landmark Discovery via Large Models for Vision-Language Navigation
https://github.com/gabinsane/multi-vaes-in-robotics
공식 코드Bridging Language, Vision and Action: Multimodal VAEs in Robotic Manipulation Tasks
https://github.com/ghy0501/awesome-continual-learning-in-generative-models
공식 코드A Comprehensive Survey on Continual Learning in Generative Models
https://github.com/google-research/valan
공식 코드Environment-agnostic Multitask Learning for Natural Language Grounded Navigation
https://github.com/guanxinglu/vlarl
공식 코드VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learning
https://github.com/hanqingwangai/ccc-vln
공식 코드출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.