game_end = time.time()
# Train the agent off the game we just played
if len(agent.get_memory()) > batch_size:
agent.replay(batch_size)
train_end = time.time()
last_play_time = (int((game_end - time_start) / 60 * 10000)) / 10000
last_train_time = (int((train_end - game_end) / 60 * 10000)) / 10000
print("Playing took: " + str(last_play_time) + " minutes.")
print("Training took: " + str(last_train_time) + " minutes.")
if len(winners_window) == window_size:
win_count_1 = winners_window.count(game.get_player_1().get_name())
win_count_2 = winners_window.count(game.get_player_2().get_name())
p1_win_ratio = win_count_1 / window_size
p2_win_ratio = win_count_2 / window_size
winners_window = []
print("Player 1 win ratio: " + str(p1_win_ratio))
print("Player 2 win ratio: " + str(p2_win_ratio))
logger.add_game({
"winner":
"Player 1"
if game.get_winner() == game.get_player_1() else "Player 2",
"play_time":
last_play_time,
"train_time":
game_end = time.time()
# Train the agent off the game we just played
if len(agent.get_memory()) > batch_size:
agent.replay(batch_size)
train_end = time.time()
last_play_time = (int((game_end-time_start) / 60 * 10000)) / 10000
last_train_time = (int((train_end-game_end) / 60 * 10000)) / 10000
print("Playing took: " + str(last_play_time) + " minutes.")
print("Training took: " + str(last_train_time) + " minutes.")
if len(winners_window) == window_size:
win_count_1 = winners_window.count(game.get_player_1().get_name())
win_count_2 = winners_window.count(game.get_player_2().get_name())
p1_win_ratio = win_count_1/window_size
p2_win_ratio = win_count_2/window_size
winners_window = []
print("Player 1 win ratio: " + str(p1_win_ratio))
print("Player 2 win ratio: " + str(p2_win_ratio))
logger.add_game({
"winner": "Player 1" if game.get_winner() == game.get_player_1() else "Player 2",
"play_time": last_play_time,
"train_time": last_train_time,
"epsilon": agent.get_epsilon(),
"player_1_health": game.get_player_1().get_health(),
"player_2_health": game.get_player_2().get_health(),
