Basic Model
Basic model class which is inherited by all models supported by the FRobs_RL library.
- class basic_model.BasicModel(env, save_model_path, log_path, ns='/', load_trained=False)[source]
Base class for all the algorithms supported by the frobs_rl library.
- Parameters
env – The environment to be used.
save_model_path – The path to save the model.
log_path – The path to save the log.
ns – The namespace of the parameters.
load_trained – Whether or not to load a trained model.
- check_env() bool[source]
Use the stable-baselines check_env method to check the environment.
- Returns
True if the environment is correct, False otherwise.
- Return type
bool
- close_env() bool[source]
Use the env close method to close the environment.
- Returns
True if the environment was closed, False otherwise.
- Return type
bool
- predict(observation, state=None, mask=None, deterministic=False)[source]
Get the current action based on the observation, state or mask
- Parameters
observation (ndarray) – The enviroment observation
state (ndarray) – The previous states of the enviroment, used in recurrent policies.
mask (ndarray) – The mask of the last states, used in recurrent policies.
deterministic (bool) – Whether or not to return deterministic actions.
- Returns
The action to be taken and the next state(for recurrent policies)
- Return type
ndarray, ndarray
- save_model() bool[source]
Function to save the model.
- Returns
True if the model was saved, False otherwise.
- Return type
bool
- save_replay_buffer() bool[source]
Funtion to save the replay buffer, to be used the training must be finished or an error will be raised.
- Returns
True if the replay buffer was saved, False otherwise.
- Return type
bool
- set_model_logger() bool[source]
Function to set the logger of the model.
- Returns
True if the logger was set, False otherwise.
- Return type
bool
- train() bool[source]
Function to train the model the number of steps specified in the ROS parameter server. The function will automatically save the model after training.
- Returns
True if the training was successful, False otherwise.
- Return type
bool