Update.

diff --git a/main.py b/main.py
index d7fb3d1..1565499 100755 (executable)
--- a/main.py
+++ b/main.py
@@ -343,7 +343,6 @@ def run_tests(model, quizz_machine, deterministic_synthesis):
             n_epoch=n_epoch,
             model=model,
             result_dir=args.result_dir,
-            logger=log_string,
             deterministic_synthesis=deterministic_synthesis,
         )
 
@@ -397,7 +396,6 @@ def create_c_quizzes(
             min_ave_seq_logproba=min_ave_seq_logproba,
             n_epoch=n_epoch,
             result_dir=args.result_dir,
-            logger=log_string,
         )
 
         sum_logits += new_c_quizzes.size(0) * ave_seq_logproba
@@ -487,7 +485,8 @@ for n_epoch in range(args.nb_epochs):
 
     a = [(model.id, float(model.main_test_accuracy)) for model in models]
     a.sort(key=lambda p: p[0])
-    log_string(f"current accuracies {a}")
+    s = " ".join([f"{p[1]*100:.02f}%" for p in a])
+    log_string(f"current accuracies {s}")
 
     # select the model with lowest accuracy
     models.sort(key=lambda model: model.main_test_accuracy)

diff --git a/quizz_machine.py b/quizz_machine.py
index bf36d0b..49e7835 100755 (executable)
--- a/quizz_machine.py
+++ b/quizz_machine.py
@@ -161,8 +161,8 @@ class QuizzMachine:
         nb_train_samples,
         nb_test_samples,
         batch_size,
-        result_dir=None,
-        logger=None,
+        result_dir,
+        logger,
         device=torch.device("cpu"),
     ):
         super().__init__()
@@ -170,6 +170,7 @@ class QuizzMachine:
         self.problem = problem
         self.batch_size = batch_size
         self.device = device
+        self.logger = logger
 
         self.train_w_quizzes = self.problem.generate_token_sequences(
             nb_train_samples
@@ -231,9 +232,9 @@ class QuizzMachine:
         return self.nb_codes
 
     def produce_results(
-        self, n_epoch, model, result_dir, logger, deterministic_synthesis, nmax=1000
+        self, n_epoch, model, result_dir, deterministic_synthesis, nmax=1000
     ):
-        def compute_accuracy(input, logger=None):
+        def compute_accuracy(input):
             input = input[:nmax]
             ar_mask = self.make_ar_mask(input)
             result = input.clone() * (1 - ar_mask)
@@ -260,18 +261,18 @@ class QuizzMachine:
 
         train_nb_total, train_nb_correct = compute_accuracy(self.train_w_quizzes)
 
-        logger(
+        self.logger(
             f"accuracy_train {n_epoch} nb_total {train_nb_total} nb_correct {train_nb_correct} accuracy {(100.0*train_nb_correct)/train_nb_total:.02f}%"
         )
 
-        test_nb_total, test_nb_correct = compute_accuracy(self.test_w_quizzes, logger)
+        test_nb_total, test_nb_correct = compute_accuracy(self.test_w_quizzes)
 
-        logger(
+        self.logger(
             f"accuracy_test {n_epoch} nb_total {test_nb_total} nb_correct {test_nb_correct} accuracy {(100.0*test_nb_correct)/test_nb_total:.02f}%"
         )
 
         main_test_accuracy = test_nb_correct / test_nb_total
-        logger(f"main_test_accuracy {n_epoch} {main_test_accuracy}")
+        self.logger(f"main_test_accuracy {n_epoch} {main_test_accuracy}")
 
         ##############################
 
@@ -418,7 +419,7 @@ class QuizzMachine:
             else:
                 break
 
-            logger(f"changing temperature to {temperature}")
+            self.logger(f"changing temperature to {temperature}")
 
         return c_quizzes, seq_logproba.mean()
 
@@ -432,7 +433,6 @@ class QuizzMachine:
         min_ave_seq_logproba,
         n_epoch,
         result_dir,
-        logger,
     ):
         c_quizzes, ave_seq_logproba = self.generate_quizzes(
             nb, model_for_generation, min_ave_seq_logproba
@@ -453,7 +453,6 @@ class QuizzMachine:
         min_ave_seq_logproba,
         n_epoch,
         result_dir,
-        logger,
     ):
         model_for_generation = Gang(models, nb_models_for_generation, mode)
         models_for_validation = models
@@ -464,5 +463,4 @@ class QuizzMachine:
             min_ave_seq_logproba,
             n_epoch,
             result_dir,
-            logger,
         )

diff --git a/sky.py b/sky.py
index e93c88a..abcd394 100755 (executable)
--- a/sky.py
+++ b/sky.py
@@ -44,12 +44,21 @@ class Sky(problem.Problem):
         "_" + "".join([chr(ord("A") + n) for n in range(len(colors) - 1)]) + "><"
     )
 
-    def __init__(self, height=6, width=8, nb_birds=3, speed=1, nb_iterations=4):
+    def __init__(
+        self,
+        height=6,
+        width=8,
+        nb_birds=3,
+        speed=2,
+        nb_iterations=2,
+        avoid_collision=True,
+    ):
         self.height = height
         self.width = width
         self.nb_birds = nb_birds
         self.speed = speed
         self.nb_iterations = nb_iterations
+        self.avoid_collision = avoid_collision
 
     def direction_tokens(self):
         return self.token_forward, self.token_backward
@@ -58,10 +67,6 @@ class Sky(problem.Problem):
         frame_sequences = []
 
         for _ in tqdm.tqdm(range(nb), dynamic_ncols=True, desc="world generation"):
-            result = torch.zeros(
-                self.nb_iterations, self.height, self.width, dtype=torch.int64
-            )
-
             i, j, vi, vj = (
                 torch.empty(self.nb_birds, dtype=torch.int64),
                 torch.empty(self.nb_birds, dtype=torch.int64),
@@ -69,49 +74,77 @@ class Sky(problem.Problem):
                 torch.empty(self.nb_birds, dtype=torch.int64),
             )
 
+            def collision_okay():
+                if not self.avoid_collision:
+                    return True
+
+                count = torch.zeros(self.height, self.width, dtype=torch.int64)
+
+                for n in range(self.nb_birds):
+                    count[i[n], j[n]] += 1
+                    count[i[n] - vi[n], j[n]] += 1
+                    count[i[n], j[n] - vj[n]] += 1
+
+                return count.max() <= 1
+
             col = (
                 torch.randperm(self.colors.size(0) - 1)[: self.nb_birds].sort().values
                 + 1
             )
 
-            for n in range(self.nb_birds):
+            while True:
                 while True:
-                    i[n] = torch.randint(self.height, (1,))
-                    j[n] = torch.randint(self.width, (1,))
-                    vm = torch.randint(4, (1,))
-                    vi[n], vj[n] = (vm % 2) * 2 - 1, (vm // 2) * 2 - 1
-                    if (
-                        i[n] - vi[n] >= 0
-                        and i[n] - vi[n] < self.height
-                        and j[n] - vj[n] >= 0
-                        and j[n] - vj[n] < self.width
-                    ):
+                    for n in range(self.nb_birds):
+                        while True:
+                            i[n] = torch.randint(self.height, (1,))
+                            j[n] = torch.randint(self.width, (1,))
+                            vm = torch.randint(4, (1,))
+                            vi[n], vj[n] = (vm % 2) * 2 - 1, (vm // 2) * 2 - 1
+                            if (
+                                i[n] - vi[n] >= 0
+                                and i[n] - vi[n] < self.height
+                                and j[n] - vj[n] >= 0
+                                and j[n] - vj[n] < self.width
+                            ):
+                                break
+
+                    if collision_okay():
                         break
 
-            for l in range(self.nb_iterations):
-                for n in range(self.nb_birds):
-                    c = col[n]
-                    result[l, i[n], j[n]] = c
-                    result[l, i[n] - vi[n], j[n]] = c
-                    result[l, i[n], j[n] - vj[n]] = c
+                result = torch.zeros(
+                    self.nb_iterations, self.height, self.width, dtype=torch.int64
+                )
+
+                for l in range(self.nb_iterations):
+                    fine = collision_okay()
+                    for n in range(self.nb_birds):
+                        c = col[n]
+                        result[l, i[n], j[n]] = c
+                        result[l, i[n] - vi[n], j[n]] = c
+                        result[l, i[n], j[n] - vj[n]] = c
 
-                    if (i[n] == 0 and vi[n] == -1) or (
-                        i[n] == self.height - 1 and vi[n] == 1
-                    ):
-                        vi[n] = -vi[n]
+                        if (i[n] == 0 and vi[n] == -1) or (
+                            i[n] == self.height - 1 and vi[n] == 1
+                        ):
+                            vi[n] = -vi[n]
 
-                    if (j[n] == 0 and vj[n] == -1) or (
-                        j[n] == self.width - 1 and vj[n] == 1
-                    ):
-                        vj[n] = -vj[n]
+                        if (j[n] == 0 and vj[n] == -1) or (
+                            j[n] == self.width - 1 and vj[n] == 1
+                        ):
+                            vj[n] = -vj[n]
 
-                    i[n] += vi[n]
-                    j[n] += vj[n]
+                        i[n] += vi[n]
+                        j[n] += vj[n]
+
+                if fine:
+                    break
 
             frame_sequences.append(result)
 
         return frame_sequences
 
+    ######################################################################
+
     def generate_token_sequences(self, nb):
         frame_sequences = self.generate_frame_sequences(nb)
 
@@ -256,12 +289,12 @@ class Sky(problem.Problem):
 if __name__ == "__main__":
     import time
 
-    sky = Sky(height=6, width=8, speed=1, nb_iterations=4)
+    sky = Sky(height=6, width=8, speed=2, nb_iterations=2)
 
     start_time = time.perf_counter()
-    seq = sky.generate_frame_sequences(nb=64)
+    token_sequences = sky.generate_token_sequences(nb=64)
     delay = time.perf_counter() - start_time
-    print(f"{seq.size(0)/delay:02f} seq/s")
+    print(f"{token_sequences.size(0)/delay:02f} seq/s")
 
     # print(sky.seq2str(seq[:4]))
 
@@ -278,9 +311,9 @@ if __name__ == "__main__":
     # m = (torch.rand(seq.size()) < 0.05).long()
     # seq = (1 - m) * seq + m * 23
 
-    print(seq.size())
-    img = sky.seq2img(seq)
-    print(img.size())
+    # print(seq.size())
+    img = sky.seq2img(token_sequences)
+    # print(img.size())
 
     torchvision.utils.save_image(
         img.float() / 255.0, "/tmp/world.png", nrow=6, padding=6, pad_value=0