import torch
import matplotlib.pyplot as plt
import agent
from unityagents import UnityEnvironment
from configs import get_dqn_cfg_defaults
plt.ion()
cfgs = get_dqn_cfg_defaults().HYPER_PARAMETER
def show_results(env, brain_name, agent):
# load the weights from file
agent.qnetwork_local.load_state_dict(
torch.load('./results/checkpoint_dqn_vec.pth',
map_location=lambda storage, loc: storage))
with torch.no_grad():
for i in range(3):
score = 0
env_info = env.reset(train_mode=False)[brain_name]
state = env_info.vector_observations[0]
while True:
action = agent.act(state)
env_info = env.step(action)[brain_name]
next_state = env_info.vector_observations[
0] # get the next state
reward = env_info.rewards[0] # get the reward
done = env_info.local_done[0] # see if episode has finished
state = next_state
score += reward
if done:
break
import torch
import torch.nn as nn
import torch.nn.functional as F
from configs import get_dqn_cfg_defaults
cfgs_model = get_dqn_cfg_defaults().MODEL_PARAMETER
class QNetwork(nn.Module):
"""Actor (Policy) Model."""
def __init__(self, state_size, action_size, seed):
"""Initialize parameters and build model.
Params
======
state_size (int): Dimension of each state
action_size (int): Dimension of each action
seed (int): Random seed
"""
super(QNetwork, self).__init__()
self.seed = torch.manual_seed(seed)
self.state_size = state_size
self.action_size = action_size
self.fc1 = nn.Linear(self.state_size, cfgs_model.H1)
self.fc2 = nn.Linear(cfgs_model.H1, cfgs_model.H2)
self.fc3 = nn.Linear(cfgs_model.H2, cfgs_model.H3)
self.fc4 = nn.Linear(cfgs_model.H3, self.action_size)
def forward(self, state):
"""Build a network that maps state -> action values."""
x = F.relu(self.fc1(state))
x = F.relu(self.fc2(x))