update

train_dqn.py

@@ -139,7 +139,7 @@ for episode in range(n_episode):
             f'num_no_capacity: {num_no_capacity[episode]}, accepted orders: {accepted_orders[episode]}')
 
 
-print(f"Training time for {n_episode} episodes: {timer() - start_time}")
+print(f"Training time for {args.n_episode} episodes: {timer() - start_time}")
 
 # save the model parameters
 torch.save(dqn.state_dict(), os.path.join(OUT_PATH, 'dqn_{}.pk1'.format(n_episode)))
@@ -186,7 +186,7 @@ plt.savefig(os.path.join(OUT_PATH, 'penalties.png'), dpi=1200, transparent=True,
 plt.close()
 
 # Plot loss
-loss= plt.plot(losses, label= 'Loss')
+loss= plt.plot(sliding_window(losses, len(losses)), label= 'Loss')
 plt.title('Loss')
 plt.xlabel('Steps')
 plt.ylabel('Loss')

train_reinforce.py

@@ -178,6 +178,8 @@ for episode in range(n_episode):
 
 #writer.close()
 
+print(f"Training time for {args.n_episode} episodes: {timer() - start_time}")
+
 # save the model parameters
 torch.save(policy_net.state_dict(), os.path.join(OUT_PATH, 'policy_{}.pk1'.format(n_episode)))
 torch.save(value_net.state_dict(), os.path.join(OUT_PATH, 'value_{}.pk1'.format(n_episode)))