Update.

diff --git a/grids.py b/grids.py
index 7752136..a115f93 100755 (executable)
--- a/grids.py
+++ b/grids.py
@@ -1126,7 +1126,7 @@ class Grids(problem.Problem):
         )
 
     def save_some_examples(self, result_dir):
-        nb, nrow = 72, 4
+        nb, nrow = 128, 4
         for t in self.all_tasks:
             print(t.__name__)
             prompts, answers = self.generate_prompts_and_answers_(nb, tasks=[t])
@@ -1155,7 +1155,7 @@ if __name__ == "__main__":
     # exit(0)
 
     # if True:
-    nb, nrow = 72, 4
+    nb, nrow = 128, 4
     # nb, nrow = 8, 2
 
     # for t in grids.all_tasks:

diff --git a/main.py b/main.py
index 4673f42..b372f12 100755 (executable)
--- a/main.py
+++ b/main.py
@@ -372,12 +372,12 @@ def one_epoch(model, quiz_machine, local_device=main_device):
 
 
 # token_logprobas are NxMxT where M is the number of models
+# def compute_valid_quizzes_(token_logprobas):
+# warnings.warn("validation with uniform constraints", RuntimeWarning)
+# l = token_logprobas.min(dim=-1).values.sort(dim=-1).values
+# return (l[:, 0] < math.log(0.1)) & (l[:, 1] > math.log(0.5))
 
-
-def compute_valid_quizzes_(token_logprobas):
-    warnings.warn("validation with uniform constraints", RuntimeWarning)
-    l = token_logprobas.min(dim=-1).values.sort(dim=-1).values
-    return (l[:, 0] < math.log(0.1)) & (l[:, 1] > math.log(0.5))
+# token_logprobas are NxMxT where M is the number of models
 
 
 def compute_valid_quizzes(token_logprobas):