def __init__(self,
N_STATES,
N_ACTIONS,
H1Size,
H2Size,
done=None,
reward=None):
super(EnvModel, self).__init__()
# build network layers
self.fc1 = nn.Linear(N_STATES + N_ACTIONS, H1Size)
self.fc2 = nn.Linear(H1Size, H2Size)
self.statePrime = nn.Linear(H2Size, N_STATES)
init_list = [self.fc1, self.fc2, self.statePrime]
if reward is None:
self.reward = nn.Linear(N_STATES, 1)
self.reward_layer = True
init_list.append(self.reward)
else: #reward function has been given
self.reward = reward
self.reward_layer = False
if done is None:
self.done = nn.Linear(N_STATES, 1)
self.done_layer = True
init_list.append(self.done)
else: #done function has been given
self.done = done
self.done_layer = False
# initialize layers
utils.weights_init_normal(init_list, 0.0, 0.1)
def __init__(self, N_STATES, N_ACTIONS, H1Size, H2Size):
super(Q_Net, self).__init__()
# build network layers
self.fc1 = nn.Linear(N_STATES, H1Size)
self.fc2 = nn.Linear(H1Size, H2Size)
self.out = nn.Linear(H2Size, N_ACTIONS)
# initialize layers
utils.weights_init_normal([self.fc1, self.fc2, self.out], 0.0, 0.1)
def __init__(self, N_STATES):
super().__init__()
self.fc1 = nn.Linear(N_STATES, N_STATES)
self.out = nn.Linear(N_STATES, 1)
utils.weights_init_normal([self.fc1, self.out], 0.0, 0.1)