Update.
authorFrançois Fleuret <francois@fleuret.org>
Sun, 22 Oct 2023 16:47:36 +0000 (18:47 +0200)
committerFrançois Fleuret <francois@fleuret.org>
Sun, 22 Oct 2023 16:47:36 +0000 (18:47 +0200)
problems.py

index 675a47a..4059856 100755 (executable)
@@ -287,7 +287,7 @@ class ProblemAddition(Problem):
 
 
 class ProblemMixing(Problem):
-    def __init__(self, height=3, width=3, nb_time_steps=12, hard=False):
+    def __init__(self, height=4, width=4, nb_time_steps=9, hard=False):
         self.height = height
         self.width = width
         self.nb_time_steps = nb_time_steps
@@ -296,7 +296,15 @@ class ProblemMixing(Problem):
     def start_random(self, nb):
         y = torch.arange(self.height * self.width).reshape(1, -1).expand(nb, -1)
 
-        m = (torch.rand(y.size()).sort(dim=-1).indices < y.size(1) // 2).long()
+        # m = (torch.rand(y.size()).sort(dim=-1).indices < y.size(1) // 2).long()
+
+        i = torch.arange(self.height).reshape(1,-1,1).expand(nb,self.height,self.width)
+        j = torch.arange(self.width).reshape(1,1,-1).expand(nb,self.height,self.width)
+
+        ri = torch.randint(self.height, (nb,)).reshape(nb,1,1)
+        rj = torch.randint(self.width, (nb,)).reshape(nb,1,1)
+
+        m = 1 - torch.logical_or(i==ri,j==rj).long().flatten(1)
 
         y = (y * m + self.height * self.width * (1 - m)).reshape(
             nb, self.height, self.width
@@ -305,12 +313,18 @@ class ProblemMixing(Problem):
         return y
 
     def start_error(self, x):
+        i = torch.arange(self.height).reshape(1,-1,1).expand_as(x)
+        j = torch.arange(self.width).reshape(1,1,-1).expand_as(x)
+
+        ri = (x == self.height * self.width).long().sum(dim=-1).argmax(-1).view(-1,1,1)
+        rj = (x == self.height * self.width).long().sum(dim=-2).argmax(-1).view(-1,1,1)
+
+        m = 1 - torch.logical_or(i==ri,j==rj).long().flatten(1)
+
         x = x.flatten(1)
-        u = torch.arange(self.height * self.width).reshape(1, -1)
-        m = ((x - u).abs() == 0).long()
-        d = (x - (m * u + (1 - m) * self.height * self.width)).abs().sum(-1) + (
-            m.sum(dim=-1) != self.height * self.width // 2
-        ).long()
+        u = torch.arange(self.height * self.width, device = x.device).reshape(1, -1)
+
+        d = (x - (m * u + (1 - m) * self.height * self.width)).abs().sum(-1)
         return d
 
     def moves(self, x):
@@ -376,7 +390,7 @@ class ProblemMixing(Problem):
         return " | ".join(
             [
                 " ".join(
-                    ["-".join([f"{x:02d}" for x in s]) for s in r.split(self.width)]
+                    ["-".join([f"{x:02d}" if x < self.height * self.width else "**" for x in s]) for s in r.split(self.width)]
                 )
                 for r in seq.split(self.height * self.width)
             ]