pixel_observation

Classes

class prt_rl.env.adapters.pixel_observation.PixelObservationAdapter(env: EnvironmentInterface, pixel_type: Literal['uint8', 'float32'] = 'uint8')

Adapater takes the ‘rgb_array’ pixels from the info dictionary and makes that the observation. The original observation is added back into the info dictionary under the ‘state’ key.

Assumes pixel observation are [H, W, C] numpy arrays of type uint8 and converts to [C, H, W] torch tensors normalized to [0, 1]. Note: This adapter assumes the base environment has render_mode set to “rgb_array”.

Parameters:

env (EnvironmentInterface) – The environment to adapt

close() → None

Closes the environment and cleans up any resources.

get_num_envs() → int

Returns the number of environments in the interface.

Returns:

Number of environments

Return type:

int

get_parameters()

Returns the EnvParams object which contains information about the sizes of observations and actions needed for setting up RL agents.

Returns:

environment parameters object

Return type:

EnvParams

reset(*args, **kwargs)

Resets the environment to the initial state and returns the initial observation.

Parameters:

seed (int | None) – Sets the random seed.

Returns:

Tuple of tensors containing the initial observation and info dictionary

Return type:

Tuple

step(action)

Steps the simulation using the action tensor and returns the new trajectory.

Parameters:

action (torch.Tensor) – Tensor with “action” key that is a tensor with shape (# env, # actions)

Returns:

Tuple of tensors containing the next state, reward, done, and info dictionary

Return type:

Tuple