Preparation

masus04 · Sep 16, 2018 · a9c96b0 · a9c96b0
1 parent 6688828
commit a9c96b0
Show file tree

Hide file tree

Showing 4 changed files with 5 additions and 6 deletions.
diff --git a/Othello/experiments/trainBaselinePlayerVsBest.py b/Othello/experiments/trainBaselinePlayerVsBest.py
@@ -79,8 +79,8 @@ def run(self, lr, milestones=False, silent=False):
 
 if __name__ == '__main__':
 
-    MILESTONES = False
-    GAMES = 1000000
+    MILESTONES = True
+    GAMES = 3000000
     EVALUATIONS = GAMES//100
     LR = random()*1e-9 + 1e-3
 

diff --git a/Othello/experiments/trainBaselinePlayerVsSelf.py b/Othello/experiments/trainBaselinePlayerVsSelf.py
@@ -70,7 +70,7 @@ def run(self, lr, milestones=False, silent=False):
 if __name__ == '__main__':
 
     MILESTONES = True
-    GAMES = 1000000
+    GAMES = 3000000
     EVALUATIONS = GAMES//100
     LR = random()*1e-9 + 1e-3
 

diff --git a/Othello/experiments/trainBaselinePlayerVsTraditionalOpponent.py b/Othello/experiments/trainBaselinePlayerVsTraditionalOpponent.py
@@ -71,7 +71,7 @@ def run(self, lr, silent=False):
 
     start = datetime.now()
 
-    GAMES = 1000000
+    GAMES = 3000000
     EVALUATIONS = GAMES//1000
     LR = random()*1e-9 + 1e-3  # uniform(1e-2, 1e-4)
 

diff --git a/TicTacToe/experiments/reinforced/trainBaselinePlayerVsTraditionalOpponent.py b/TicTacToe/experiments/reinforced/trainBaselinePlayerVsTraditionalOpponent.py
@@ -77,7 +77,6 @@ def run(self, lr, silent=False):
     GAMES = 3000000
     EVALUATIONS = 1000
     LR = random()*1e-9 + 1e-3  # uniform(1e-2, 1e-4)
-    WEIGHT_DECAY = 0.003
 
     PLAYER = None  # Experiment.load_player("player.pth")
     OPPONENT = None  # ExpertPlayer()
@@ -86,7 +85,7 @@ def run(self, lr, silent=False):
 
     experiment = TrainBaselinePlayerVsTraditionalOpponent(games=GAMES, evaluations=EVALUATIONS, pretrained_player=PLAYER, opponent=OPPONENT)
     try:
-        experiment.run(lr=LR, weight_decay=WEIGHT_DECAY)
+        experiment.run(lr=LR)
     except:
         experiment.save_player(experiment.player1)