ep_r += r
if miner.memory_counter > opt.memory_capacity:
miner.optimize_model()
if game.get_status() != 0:
print('Ep: ', epoch,
'| Ep_r: ', round(ep_r, 2))
if status != 0:
break
s = s_.copy()
critic_r += ep_r
if (epoch+1) % opt.n_critic == 0:
print('=====evaluation=====')
print('Epochs:', epoch)
print('win number:', win_num)
print('fail number:', fail_num)
print('win rate:', win_num / (win_num + fail_num))
print('total reward:', critic_r)
avg_rewards.append(critic_r)
success.append(win_num)
win_num = 0
fail_num = 0
critic_r = 0
plot_durations(success, 'Accuracy','batch_size16_rmsprop_longdirect.png' )
plot_durations(avg_rewards, 'Critic Reward','batch_size16_rmsprop_longreward.png' )
miner.save_params()
