21 datasets found

A near real-world benchmark for offline RL, which contains datasets from various domains with controlled sizes, and extra test datasets for policy validation.

The HandManipulateBlock environment from OpenAI gym robotics suite

The FetchPickAndPlace and HandManipulateBlock environments from OpenAI gym robotics suite

The FetchPush, FetchPickAndPlace and HandManipulateBlock environments from OpenAI gym robotics suite

The Funnel board task is a domain where a ball falls through a grid of obstacles onto one of five platforms. Every other row of obstacles consists of funnel-shaped objects,...

The Room runner task is a domain where an agent moves through a randomly generated map of rooms, which are observed in 2D from above. The agent follows the policy of always...

The dataset used in the paper is a simulated environment for event camera-based reinforcement learning. The dataset includes a car-like robot equipped with an event camera, and...

The dataset used in the paper is a collection of robot morphologies, each with a unique set of actuators and joints.

The dataset used in the paper is a collection of environments with multiple confounders, including mass, length, damping, and a crippled leg. The dataset is used to evaluate the...

The dataset used in the paper is a simulated self-driving car environment, a mountain car environment, and a robotic environment.

The dataset used in the paper is an Inverted Pendulum dataset, which is a standard benchmark system in control and reinforcement learning.

The dataset used for training a robotic agent to play the theremin instrument using time-dependent goals.

The dataset used in the paper is a set of motor babbling data, which is used to initialize the dynamics model and to optimize the control policies.

The dataset used in the paper is RLBench, a standard benchmark for vision-based robotics which has been shown to serve as a proxy for real-robot experiments.

A benchmark for evaluating task compositionality and long-horizon tasks through object manipulation, with language serving as the mechanism for goal specification.

The dataset used in the paper is a set of real-world robot learning tasks, including Create-Reacher, UR5-VisualReacher, Vector-ChargerDetector, and Franka-VisualReacher.

The dataset used in the paper is not explicitly described, but it is mentioned that the authors used a combination of reinforcement learning and generative models to solve...

The DeepMind Control Suite is a collection of 20 robotic manipulation tasks, each with 5 different environments and 5 different robot parameters. The tasks are designed to test...

The dataset used in the paper is a collection of simulated environments, including blockworld, canada, cart, cartpole, cartpoleBalance, cartpoleParallelDouble,...

The dataset used in the paper is not explicitly described, but it is mentioned that the authors used several continuous control environments from the OpenAI Gym.