def __init__(self, state_dim, action_dim):
"""
:param state_dim: Dimension of input state (int)
:param action_dim: Dimension of input action (int)
:return:
"""
super(Critic, self).__init__()
self.state_dim = state_dim
self.action_dim = action_dim
print("dimmmmmmmmmm", state_dim, state_dim + action_dim)
self.fc1 = nn.Linear(state_dim + action_dim, 400)
self.fc1.weight.data = utls.fanin_init(self.fc1.weight.data.size())
self.fc2 = nn.Linear(400, 300)
self.fc2.weight.data = utls.fanin_init(self.fc2.weight.data.size())
self.fc3 = nn.Linear(300, 1)
self.fc3.weight.data.uniform_(-EPS, EPS)
# self.fc3.weight.data = utls.fanin_init(self.fc3.weight.data.size())
self.fca1 = nn.Linear(state_dim + action_dim, 400)
self.fca1.weight.data = utls.fanin_init(self.fca1.weight.data.size())
self.fca2 = nn.Linear(400, 300)
self.fca2.weight.data = utls.fanin_init(self.fca2.weight.data.size())
self.fca3 = nn.Linear(300, 1)
self.fca3.weight.data.uniform_(-EPS, EPS)
def __init__(self, state_dim, action_dim, action_lim=1.):
"""
:param state_dim: Dimension of input state (int)
:param action_dim: Dimension of output action (int)
:param action_lim: Used to limit action in [-action_lim,action_lim]
:return:
"""
super(Actor, self).__init__()
self.state_dim = state_dim
self.action_dim = action_dim
self.action_lim = action_lim
self.fc1 = nn.Linear(state_dim, 256)
self.fc1.weight.data = utls.fanin_init(self.fc1.weight.data.size())
self.fc2 = nn.Linear(256, 128)
self.fc2.weight.data = utls.fanin_init(self.fc2.weight.data.size())
self.fc3 = nn.Linear(128, 100)
self.fc3.weight.data = utls.fanin_init(self.fc3.weight.data.size())
self.fc4 = nn.Linear(100, 64)
self.fc4.weight.data = utls.fanin_init(self.fc4.weight.data.size())
self.fc5 = nn.Linear(64, action_dim)
self.fc5.weight.data.uniform_(-EPS, EPS)