MOSI
Datasets
All datasets matching “MOSI”CMU-MOSIMOSIV
Overview
MOSIV is a synthetic dataset designed for the task of multi-object system identification from videos, where the goal is to recover both the 4D geometry (time-varying 3D shape) and the physical properties (e.g., stiffness, friction, plasticity) of multiple interacting objects directly from multi-view video observations.
MOSIV features contact-rich, multi-object interactions with diverse materials and complex dynamics. Each sequence contains multiple objects undergoing… See the full description on the dataset page: https://huggingface.co/datasets/Hanibel/MOSIV.LiveClawbench-trajectoriesLiveClawBench: Benchmarking LLM Agents on Complex, Real-World Assistant Tasks
Overview
LLM agents are increasingly expected to handle real-world assistant tasks — booking flights, managing emails, debugging code, curating knowledge bases — yet existing benchmarks evaluate them under isolated difficulty sources. LiveClawBench addresses this gap by introducing a Triple-Axis Complexity Framework and building a benchmark of 134 manually constructed tasks with explicit factor… See the full description on the dataset page: https://huggingface.co/datasets/Mosi-AI/LiveClawbench-trajectories.mtmc-dog-detectionMoSim_Dataset
🗂️ MoSim Dataset
Official release of the dataset from the paper:Neural Motion Simulator: Pushing the Limit of World Models in Reinforcement Learning
This dataset contains sequential state-action trajectories for training and evaluating MoSim (Neural Motion Simulator) world models.All trajectories are collected from random policies in classical control and locomotion environments.
📦 Dataset Overview
Format: .npz (NumPy compressed arrays)
Contents:… See the full description on the dataset page: https://huggingface.co/datasets/wujiss1/MoSim_Dataset.mosim-humanoid-walk-tdmpc2-ar10-dt005
MoSim Humanoid Walk TD-MPC2 AR10, dt=0.005
Humanoid Walk state transition dataset collected using a TD-MPC2 policy. Rows are recorded every physics step at dt=0.005 seconds; the policy action is refreshed every 10 physics rows.
Format
Each .npz contains:
data: float32 array with shape (num_trajectories, 1000, state_dim + action_dim + state_dim).
metadata: object with state_dim, action_dim, and DMC qpos/qvel sizes when available.
Each transition row is [s_t, a_t… See the full description on the dataset page: https://huggingface.co/datasets/Weyl09/mosim-humanoid-walk-tdmpc2-ar10-dt005.