relayrl_framework 0.4.52

A system-oriented, distributed reinforcement learning framework using a Rust backend with Python interfaces.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

from abc import ABC, abstractmethod


class ApplicationAbstract(ABC):
    """
    Abstract class for designing and implementing new applications that can interface with RelayRL.
    """
    def __init__(self, *args, **kwargs):
        super(ApplicationAbstract, self).__init__(*args, **kwargs)

    @abstractmethod
    def run_application(self, *args, **kwargs):
        """
        Run the application's main loop where observations are collected, actions are taken,
        and rewards are calculated.
        """
        pass

    @abstractmethod
    def build_observation(self, *args, **kwargs):
        """
        Build the observation for the current step to be passed through the RelayRL Agent in run_application.
        """
        pass

    @abstractmethod
    def calculate_performance_return(self, *args, **kwargs):
        """
        Calculate the performance return for RelayRL's flag_last_action reward value.
        """
        pass