Drive every actuator with a phase-shifted sinusoid — exercises all DOFs. This is a stronger engine-parity stressor than gravity-only: it pushes t

Set every recorded state on a raw-mujoco model and on the handler; compare FK.

Send actions to the vectorized env.

Asynchronously step the environment.

Asynchronously step the environment.

Step the envs and adapt outputs for SB3.

Wait for the step to complete.

Wait for the step to complete.

Swap and then flatten axes 0 (buffer_size) and 1 (n_envs) to convert shape from [n_steps, n_envs, ...] (when ... is the shape of the

Switch to next entity (TAB key)

Switch to next entity (TAB key)

Switch to the next entity (TAB key).

Convert canonical teleop targets into robot joint control.

Sends all torch tensors in nested dictionary or list or tuple to device @device, and returns a new nested structure. Args: x (di

Converts all torch tensors and numpy arrays in nested dictionary or list or tuple to float type entries, and returns a new nested structure.

Convert tensor to one-hot representation, assuming a certain number of total class labels. Args: tensor (torch.Tensor): tensor conta

Converts a string from camel case to snake case. Args: camel_str: A string in camel case. Returns: A string in snake case (i

Converts a string from camel case to snake case. Args: camel_str: A string in camel case. Returns: A string in snake case (i

Toggle joint control mode (J key)

Toggle joint control mode (J key)

Toggle the joint control mode (J key).

Train RSL-RL PPO

Train RSL-RL PPO

Train PPO for reaching task using RLTaskEnv.

Train PPO for reaching task using RLTaskEnv.

Train PPO for reaching task using RLTaskEnv.

Training procedure for PPO

Transform torso IMU data into pelvis frame given waist yaw and rate.

Adds dimension of size 1 at dimension @dim in all torch tensors and numpy arrays in nested dictionary or list or tuple and returns a new nest

Unsqueeze and expand a tensor at a dimension @dim by @size. Args: x (dict or list or tuple): a possibly nested dictionary or list or

Update observation buffer.

Validate a command payload sent alongside teleoperation packets.

This function is used to visualize the grasp group or grasp.

Block until at least one low state message arrives.

Generate a Python config file in roboverse_pack/tasks/gpt/metacfg/{snake_task_name}.py Class name is {CamelCaseTaskName}Cfg, but the final ru

Write material library file (.mtl) to link texture with OBJ.

Write material library file (.mtl) to link texture with OBJ.

Write material library file (.mtl) to link texture with OBJ.

Write the <robot> element tree to an XML file in a pretty-printed way.

Write the <robot> element tree to an XML file in a pretty-printed way.

Write the <robot> element tree to an XML file in a pretty-printed way.

Apply zero actuator command every step (passive dynamics under gravity).

Hold zero torque until the start signal is received.

Handle incoming low state messages for GO transports.

Handle incoming low state messages for HG transports.

Callable convenience so this object can be used as a callback. It reads the configured keys from env_states.robots[env.name] (or from

Return sensor data whenever `get_extra()` is invoked. * Heavy libraries are imported **inside** the relevant branch only.

Return (N_env, 3) site position whenever `get_extra()` is invoked. * Heavy libraries are imported **inside** the relevant branch only.

Return (N_env, 9) site rotation matrix whenever `get_extra()` is invoked. * Heavy libraries are imported **inside** the relevant branch only.

Return contact info of the two geom.

Dispatch LiDAR queries for the configured simulator backend.

Apply material randomization for provided environment ids.

Apply mass randomization for selected environments.

Calculate the stability reward based on neck height, upright position, and small control forces.

Calculate the locomotion reward based on stability and movement speed.

Vectorised HumanoidBench-style sitting reward.

Compute the stair reward.

Compute the stair reward.

Vectorised HumanoidBench-style sitting reward. Returns: ------- reward : (B,) tensor info : dict # individual ter

Call the query.

Call the query.