def train():
record = 0
agent = Agent()
game = Game(True)
counter: int = 0
game.set_game_time(1)
while True:
state_old = game.get_state()
move = agent.get_action(state_old)
if move == [1, 0, 0]:
final_move = TriDirections.STRAIGHT
elif move == [0, 1, 0]:
final_move = TriDirections.RIGHT
else:
final_move = TriDirections.LEFT
game.set_direction(final_move)
reward, done, score = game.game_logic()
if reward == 0:
counter += 1
else:
counter = 0
state_new = game.get_state()
if counter > 100:
done = True
reward = -10
if counter > 100:
game.set_game_time(5)
agent.train_short_memory(state_old, move, reward, state_new, done)
agent.remember(state_old, move, reward, state_new, done)
if done:
counter = 0
game.reset()
agent.n_games += 1
agent.train_long_memory()
if score > record:
record = score
agent.model.save()
print('Game', agent.n_games, 'Score', score, 'Record:', record)
from main import Game, select_action
import sys
states = {}
with open(sys.argv[1], "r") as f:
line = f.readline()
while line:
data = line[:-1].split("#")
states[eval(data[0])] = eval(data[1])
line = f.readline()
game = Game()
while 1:
state = game.get_state()
if not state in states.keys():
states[state] = {}
for i in game.get_actions():
states[state][i] = 0.0
action = max(states[state], key=states[state].get)
# print(states[state])
game.do_action(action, 1)
next_state = game.get_state()
if game.is_end() != -1:
game.print()
print(f"game end {game.is_end()}")
break
actions = game.get_actions()
