DQN with experience replay and target network

memory.py

from collections import deque
from random import sample


class ReplayMemory:
    def __init__(self, capacity):
        self.samples = deque([], maxlen=capacity)

    def store(self, exp):
        self.samples.append(exp)
        pass

    def get_batch(self, n):
        n_samples = min(n, len(self.samples))
        samples = sample(self.samples, n_samples)
        return samples