reward = 0
reward_list = []
for eval_episode in range(evaluate_num):
print('\rEPISODE {} - Eval {} over {} - Number of game played {} - {}'.format(episode, eval_episode,
evaluate_num,
total_game_played,
time_difference_good_format(
seconds,
time.time())),
end='')
_, payoffs = eval_env.run(is_training=False)
total_game_played += 1
reward_list.append(payoffs[0])
reward += payoffs[0]
logger.log('\n########## Evaluation - Episode {} ##########'.format(episode))
logger.log('Timestep: {} Average reward is {}'.format(env.timestep, float(reward) / evaluate_num))
# Add point to logger
logger.add_point(x=env.timestep, y=float(reward) / evaluate_num)
# Make plot
if episode % save_plot_every == 0 and episode > 0:
logger.make_plot(save_path=figure_path + str(episode) + '.png')
logger.make_plot_hist(save_path_1=figure_path + str(episode) + '_hist.png',
save_path_2=figure_path + str(episode) + '_freq.png', reward_list=reward_list)
# Make the final plot
logger.make_plot(save_path=figure_path + 'final_' + str(episode) + '.png')
logger.make_plot_hist(save_path_1=figure_path + str(episode) + '_hist.png',
save_path_2=figure_path + str(episode) + '_freq.png', reward_list=reward_list)
.format(episode))
logger_random.log(
'Timestep: {} Average reward against random is {}'.format(
env.timestep,
float(reward_random) / evaluate_num))
# Add point to logger
logger_random.add_point(x=env.timestep,
y=float(reward_random) / evaluate_num)
# Make plot
logger_random.make_plot(save_path=figure_path_random +
str(episode) + '.png')
logger_random.make_plot_hist(
save_path_1=figure_path_random + str(episode) + '_hist.png',
save_path_2=figure_path_random + str(episode) + '_freq.png',
reward_list=reward_random_list,
taking_list=taking_list)
# Eval against last agent
reward_opponent = 0
reward_opponent_list = []
taking_list = []
eval_env.set_agents([agent] + [opponent_agent] *
(env.player_num - 1))
for eval_episode in range(evaluate_num):
print(
'\rEPISODE {} - Eval Opponent {} over {} - Number of game played {} - {}'
.format(episode, eval_episode, evaluate_num,
total_game_played,
time_difference_good_format(seconds, time.time())),
