WaterHyacinthInNANHU
- 23 skills
- 0 followers
- 6 hours ago last updated
- ▌ Paper Rob Openpi · waterhyacinthinnanhu bundlePhysical Intelligence VLA foundation models (pi0, pi0-FAST, pi0.5) for generalist robot control
- ▌ Paper Rob Resfit · waterhyacinthinnanhu bundleResFiT: Residual Off-Policy RL for Finetuning Behavior Cloning Policies
- ▌ Paper 3d Concerto · waterhyacinthinnanhu bundleConcerto — Joint 2D-3D Self-Supervised Learning for Spatial Representations
- ▌ Paper Rob Robocasa · waterhyacinthinnanhu bundleRoboCasa -- large-scale kitchen simulation framework (365 tasks, 2500 scenes) for training generalist robots, built on robosuite/MuJoCo
- ▌ Paper Rob Any3d Vla · waterhyacinthinnanhu bundleAny3D-VLA -- Enhancing VLA Robustness via Diverse Point Clouds
- ▌ Paper Rob 3d Diffusion Policy · waterhyacinthinnanhu bundle3D Diffusion Policy (DP3) -- visuomotor imitation learning via sparse point cloud representations and diffusion-based action generation
- ▌ Rlinf · waterhyacinthinnanhu bundleTroubleshooting guide and operational knowledge for RLinf RL training framework on HPCC.
- ▌ Create Skill · waterhyacinthinnanhu bundleGuide for creating well-structured Claude Code skills following best practices.
- ▌ Topic Survey · waterhyacinthinnanhuSurvey a research field interactively, exploring papers by topic and writing a structured literature review.
- ▌ Idea Refinery · waterhyacinthinnanhu bundleIteratively refine a research idea through literature survey, validation, and branching exploration until convergence.
- ▌ Auto Experiment · waterhyacinthinnanhu bundleAutomated experiment workflow for code + data experiments with structured planning, execution, and reporting.
- ▌ Ucr Hpcc Cluster · waterhyacinthinnanhu bundleHelp users work with the UCR HPCC and BCOE department clusters. Provides commands for connecting, submitting jobs, managing software, and handling data storage.
- ▌ Experiment Report · waterhyacinthinnanhu bundleWrite structured experiment reports for robot learning and AI research.
- ▌ Paper Related Works · waterhyacinthinnanhuGiven a paper, find its related works (cited references and successor/citing papers) and produce a structured map.
- ▌ Paper Rl Qc · waterhyacinthinnanhu bundleQ-Chunking -- action chunking for TD-based offline-to-online RL on long-horizon sparse-reward tasks
- ▌
- ▌ Paper Rob Bpp · waterhyacinthinnanhu bundleBPP (Big Picture Policies) -- long-context robot imitation learning via VLM-detected keyframes to reduce history distribution shift
- ▌ Paper Rob Rdd · waterhyacinthinnanhu bundleTraining-free sub-task decomposition for long-horizon robot tasks via retrieval and dynamic programming.
- ▌ Python Project Layout · waterhyacinthinnanhu bundleRefactor Python projects to follow modern best-practice directory structure using src layout and pyproject.toml.
- ▌ Paper Rob Ibrl · waterhyacinthinnanhu bundleIBRL uses Q-based selection between RL and BC actions for exploration and bootstrapping to achieve sample-efficient robotic manipulation.
- ▌
- ▌ Create Skill With Paper · waterhyacinthinnanhu bundleTurn an academic paper into a reusable Claude Code skill for quick reference and reproduction in future projects.
- ▌ Paper 3d Utonia · waterhyacinthinnanhu bundleUtonia - unified cross-domain point cloud encoder (PTv3-based, self-supervised, 5 domains)