def validate(players, n_games):
game = Game(players=players,
dealer_type=PerfectDealer,
verbose=False,
collect_stats=True)
# turn training off
for player in players:
player.training = False
for _ in range(n_games):
game.start_round()
# turn training back on
for player in players:
player.training = True
# average the dealer reward
dealer_reward = sum(game.collector.rewards['dealer']) / n_games
# average player rewards
player_reward = 0
for player in players:
player_reward += sum(game.collector.rewards[player.name]) / n_games
player_reward /= len(players)
return player_reward, dealer_reward
with open('validation.pickle', 'rb') as f:
validation = pickle.load(f)
except FileNotFoundError:
validation = []
game = Game(players=players,
dealer_type=PerfectDealer,
verbose=False,
collect_stats=False)
# collect validation results here
stats = []
start_time = time.time()
for i in range(TRAINING_ROUNDS):
game.start_round()
# validate
if VALIDATE_EVERY > 0 and i % VALIDATE_EVERY == VALIDATE_EVERY - 1:
print('VALIDATING...', end='', flush=True)
stats.append(validate(players, VALIDATION_ROUNDS))
print('PLAYER:', stats[-1][0], 'DEALER:', stats[-1][1])
# also save stats
with open('validation.pickle', 'wb') as f:
pickle.dump(stats, f)
# save
if i % SAVE_EVERY == SAVE_EVERY - 1:
print('SAVING --', i + 1)
save_values(players, SAVE_DIR)