import numpy as np
from ppo_torch import Agent
import os
from utils import plot_learning_curve
if __name__ == "__main__":
env = gym.make("CartPole-v0")
N = 20
batch_size = 5
n_epochs = 4
alpha = 0.0003
agent = Agent(
n_actions=env.action_space.n,
batch_size=batch_size,
alpha=alpha,
n_epochs=n_epochs,
input_dims=env.observation_space.shape,
)
n_games = 300
if not os.path.exists("plots"):
os.makedirs("plots")
figure_file = os.path.join("plots", "cartpole.png")
best_score = env.reward_range[0]
score_history = []
learn_iters = 0
avg_score = 0
import gym
import numpy as np
from ppo_torch import Agent
from utils import plot_learning_curve
from ppo_torch import ActorNetwork
from newmaze import Maze
if __name__ == '__main__':
#env = gym.make('FrozenLake-v0', is_slippery=False )
env = Maze()
N = 20
batch_size = 5
n_epochs = 4
alpha = 0.0003
agent = Agent(n_actions=env.action_space.n,
batch_size=batch_size,
alpha=alpha,
n_epochs=n_epochs,
input_dims=(1, ))
agent.load_models()
observation = env.reset()
done = False
env.render()
while not done:
action, prob, val = agent.choose_action(observation)
observation_, reward, done, info = env.step(action)
env.render()
N = 20
batch_size = 5
n_epochs = 4
alpha = 0.0003
best_score = env.reward_range[0]
perturbation = p
score_book = {}
actor_loss_book = {}
critic_loss_book = {}
total_loss_book = {}
for trial in range(n_trials):
print('\nTrial:', trial+1)
agent = Agent(n_actions=env.action_space.n, batch_size=batch_size, alpha=alpha,
n_epochs=n_epochs, input_dims=env.observation_space.shape,
fc1_dims=256, fc2_dims=256, chkpt_dir='tmp/ppo')
advAgent = adv_Agent(input_dims=env.observation_space.shape, layer1_size=200, layer2_size=200,
env=env, n_actions=env.action_space.n, gamma=0.99, max_size=1000000, tau=0.005, ent_alpha=0.5, batch_size=100,
reward_scale = 1, chkpt_dir='advModels/SACd/tmp/adv')
score_history = []
avg_score_history = []
loss = []
actor_loss = []
critic_loss = []
total_loss = []
learn_iters = 0
avg_score = 0
best_score = env.reward_range[0]
layer_1_dim = 128
layer_2_dim = 128
# Final results
score_book = {}
actor_loss_book = {}
critic_loss_book = {}
total_loss_book = {}
for trial in range(n_trials):
print('\nTrial:', trial + 1)
agent = Agent(n_actions=env.action_space.n,
batch_size=batch_size,
alpha=alpha,
n_epochs=n_epochs,
input_dims=env.observation_space.shape,
fc1_dims=layer_1_dim,
fc2_dims=layer_2_dim,
chkpt_dir=chkpt_dir)
# Initialize storage pointers
score_history = []
avg_score_history = []
loss = []
actor_loss = []
critic_loss = []
total_loss = []
# Initialize the run
learn_iters = 0
avg_score = 0