Changed training to update values-file during training

tic_tac_toe.py

@@ -12,19 +12,9 @@ REPORT_INTERVAL = 1000
 report_counter = REPORT_INTERVAL
 
 
-def report(k):
-	global report_counter
-	if report_counter > 0:
-		report_counter -= 1
-	else:
-		report_counter = REPORT_INTERVAL-1
-		print(f"Iteration: {k:8d}")
-
-
 def play(player_provider: PlayerProvider, k_max=10000, with_debug=False):
 	player_provider.set_debug(with_debug)
 	for k in range(0, k_max):
-		report(k)
 		state = create_empty_state()
 		players = player_provider.choose(state)
 		move = 1
@@ -82,16 +72,21 @@ p_hx = HumanPlayer(mark='X', name="Jens")
 p_mx = MachinePlayer(mark='X', params=MachinePlayer.Params(p_explore=0.1, alpha=0.1), values=x_values)
 p_mo = MachinePlayer(mark='O', params=MachinePlayer.Params(p_explore=0.1, alpha=0.1), values=o_values)
 
+K = 50
+N = 1000
 if do_training:
-	play(PlayerProvider(p_mx, p_mo), 5000, False)
+	for k in range(0, K):
+		play(PlayerProvider(p_mx, p_mo), N, False)
 
-	# Values after training
-	# Convert and write JSON object to file
-	with open("values_x.json", "w") as fp:
-		json.dump(p_mx.values, fp, indent=0)
+		# Values after training
+		# Convert and write JSON object to file
+		with open("values_x.json", "w") as fp:
+			json.dump(p_mx.values, fp, indent=0)
 
-	with open("values_o.json", "w") as fp:
-		json.dump(p_mo.values, fp, indent=0)
+		with open("values_o.json", "w") as fp:
+			json.dump(p_mo.values, fp, indent=0)
+
+		print(f"Iteration: {k*N:8d}")
 
 if do_human_player:
 	pp = PlayerProvider(p_hx, p_mo)