model.py

import numpy as np

import torch
import torch.nn as nn
import torch.nn.functional as F

def hidden_init(layer):
    fan_in = layer.weight.data.size()[0]
    lim = 1. / np.sqrt(fan_in)
    return (-lim, lim)

class Actor(nn.Module):
    def __init__(self, state_size, action_size, fc1_units=256, fc2_units=128):
        """
        Params
        ======
            state_size (int): Dimension of each state
            action_size (int): Dimension of each action
            fc1_units (int): Number of neurons in the first hidden layer
            fc2_units (int): Number of neurons in the second hidden layer
        """
        super(Actor, self).__init__()
        self.fc1 = nn.Linear(state_size*2, fc1_units)
        self.fc2 = nn.Linear(fc1_units, fc2_units)
        self.fc3 = nn.Linear(fc2_units, action_size)
        self.reset_parameters()
        
    def reset_parameters(self):
        self.fc1.weight.data.uniform_(*hidden_init(self.fc1))
        self.fc2.weight.data.uniform_(*hidden_init(self.fc2))
        self.fc3.weight.data.uniform_(-3e-3, 3e-3)
        
    def forward(self, state):
        """Build a network that maps state -> action values"""
        x = F.relu(self.fc1(state))
        x = F.relu(self.fc2(x))
        return torch.tanh(self.fc3(x))
    
class Critic(nn.Module):
    def __init__(self, state_size, action_size, fc1_units=256, fc2_units=128):
        """
        Params
        ======
            state_size (int): Dimension of each state
            action_size (int): Dimension of each action
            fc1_units (int): Number of neurons in the first hidden layer
            fc2_units (int): Number of neurons in the second hidden layer
        """
        super(Critic, self).__init__()
        self.fc1 = nn.Linear(2 * (state_size + action_size), fc1_units)
        self.fc2 = nn.Linear(fc1_units, fc2_units)
        self.fc3 = nn.Linear(fc2_units, 1)
        self.bn = nn.BatchNorm1d(fc1_units)
        self.reset_parameters()
        
    def reset_parameters(self):
        self.fc1.weight.data.uniform_(*hidden_init(self.fc1))
        self.fc2.weight.data.uniform_(*hidden_init(self.fc2))
        self.fc3.weight.data.uniform_(-3e-3, 3e-3)
        
    def forward(self, state, action):
        """Build a network that maps (state, action) -> Q values"""
        x = F.leaky_relu(self.fc1(torch.cat((state, action.float()), dim=1)))
        x = self.bn(x)
        x = F.leaky_relu(self.fc2(x))
        return self.fc3(x)