def plot_rewards_toy_mr():
num_lines = 100
data_1 = parse_results.parse_results_file(
'./data/hadooqn_toy_mr_sectors.txt', max_lines=num_lines)
data_2 = parse_results.parse_results_file('./data/cts_toy_mr.txt',
max_lines=num_lines)
data_3 = parse_results.parse_results_file('./data/cts_toy_mr_lives.txt',
max_lines=num_lines)
data_4 = parse_results.parse_results_file(
'./data/cts_toy_mr_lives_repeat_action.txt', max_lines=num_lines)
data_5 = parse_results.parse_results_file(
'./data/double_dqn_toy_mr_fake.txt', max_lines=num_lines)
data_1 = data_1['reward'].values.astype(np.float)
data_2 = data_2['reward'].values.astype(np.float)
data_3 = data_3['reward'].values.astype(np.float)
data_4 = data_4['reward'].values.astype(np.float)
data_5 = data_5['reward'].values.astype(np.float)
labels = [
'DAQN', 'Intrinsic', 'Intrinsic+L', 'Intrinsic+L+S', 'Double DQN'
]
plot_data.plot_data(list(range(num_lines)),
[data_1, data_2, data_3, data_4, data_5],
'Reward',
'Millions of Frames',
'Average Test Reward',
labels=labels,
ylim=[-.1, 1.1],
yticks=[0, 1],
save_file='./figures/toy_mr_reward.png',
legend_loc='lower right')
def plot_rewards_coin():
data_1 = parse_results.parse_results_file('./data/hadooqn_coin_fake.txt')['reward'].values.astype(np.float)
data_2 = parse_results.parse_results_file('./data/cts_coin_fake.txt')['reward'].values.astype(np.float)
data_3 = parse_results.parse_results_file('./data/double_dqn_coin_fake.txt')['reward'].values.astype(np.float)
labels = ['DAQN', 'Intrinsic', 'Double DQN']
plot_data.plot_data(range(0, 200), [data_1, data_2, data_3], 'Reward', 'Millions of Frames', 'Average Test Reward',
labels=labels, save_file='./figures/coin_reward.png', legend_loc='lower right')
def plot_rooms_4_rooms():
data_1 = parse_results.parse_results_file('./data/hadooqn_4_rooms_fake.txt')['rooms'].values.astype(
np.float)
data_2 = parse_results.parse_results_file('./data/cts_4_rooms_fake.txt')['rooms'].values.astype(np.float)
data_3 = parse_results.parse_results_file('./data/double_dqn_4_rooms_fake.txt')['rooms'].values.astype(np.float)
labels = ['DAQN', 'Intrinsic', 'Double DQN']
plot_data.plot_data(range(0, 200), [data_1, data_2, data_3], 'Rooms Discovered', 'Millions of Frames', 'Rooms Discovered',
labels=labels, ylim=None, yticks=None, save_file='./figures/4_rooms_rooms.png', legend_loc='lower right')
def plot_rooms_toy_mr():
num_lines = 100
data_1 = parse_results.parse_results_file('./data/hadooqn_toy_mr_sectors.txt', max_lines=num_lines)
data_2 = parse_results.parse_results_file('./data/cts_toy_mr.txt', max_lines=num_lines)
data_3 = parse_results.parse_results_file('./data/cts_toy_mr_lives.txt', max_lines=num_lines)
data_4 = parse_results.parse_results_file('./data/cts_toy_mr_lives_repeat_action.txt', max_lines=num_lines)
data_5 = parse_results.parse_results_file('./data/double_dqn_toy_mr_fake.txt', max_lines=num_lines)
data_1 = data_1['rooms'].values.astype(np.float)
data_2 = data_2['rooms'].values.astype(np.float)
data_3 = data_3['rooms'].values.astype(np.float)
data_4 = data_4['rooms'].values.astype(np.float)
data_5 = data_5['rooms'].values.astype(np.float)
labels = ['DAQN', 'Intrinsic', 'Intrinsic+L', 'Intrinsic+L+S', 'Double DQN']
plot_data.plot_data(range(num_lines), [data_1, data_2, data_3, data_4, data_5], 'Rooms Discovered', 'Millions of Frames', 'Rooms Discovered',
labels=labels, ylim=None, yticks=None, save_file='./figures/toy_mr_rooms.png', legend_loc='upper right')