Update.

[culture.git] / tasks.py

diff --git a/tasks.py b/tasks.py
index 1254323..80ffdbb 100755 (executable)
--- a/tasks.py
+++ b/tasks.py
@@ -22,6 +22,8 @@ def masked_inplace_autoregression(
     batch_size,
     input,
     ar_mask,
+    summed_logits,
+    temperature,
     deterministic_synthesis,
     forbidden_tokens=None,
     logit_biases=None,
@@ -46,11 +48,13 @@ def masked_inplace_autoregression(
 
         for input, ar_mask in batches:
             model.masked_inplace_autoregression(
-                input,
-                ar_mask,
-                deterministic_synthesis,
-                forbidden_tokens,
-                logit_biases,
+                input=input,
+                ar_mask=ar_mask,
+                summed_logits=summed_logits,
+                temperature=temperature,
+                deterministic_synthesis=deterministic_synthesis,
+                forbidden_tokens=forbidden_tokens,
+                forced_biases=logit_biases,
             )
 
         model.train(t)
@@ -79,7 +83,7 @@ import world
 
 class World(Task):
     def save_image(self, input, result_dir, filename, logger):
-        img = world.sample2img(input.to("cpu"), self.height, self.width)
+        img = world.seq2img(input.to("cpu"), self.height, self.width)
         image_name = os.path.join(result_dir, filename)
         torchvision.utils.save_image(img.float() / 255.0, image_name, nrow=6, padding=4)
         logger(f"wrote {image_name}")
@@ -101,8 +105,8 @@ class World(Task):
 
         self.batch_size = batch_size
         self.device = device
-        self.height = 7
-        self.width = 9
+        self.height = 6
+        self.width = 8
 
         self.train_input = world.generate_seq(
             nb_train_samples, height=self.height, width=self.width
@@ -112,13 +116,6 @@ class World(Task):
             nb_test_samples, height=self.height, width=self.width
         ).to(device)
 
-        # print()
-        # for a in world.seq2str(self.train_input):
-        # print(a)
-        # for a in world.seq2str(self.test_input):
-        # print(a)
-        # exit(0)
-
         self.nb_codes = max(self.train_input.max(), self.test_input.max()) + 1
 
         self.train_quizzes = []
@@ -155,6 +152,9 @@ class World(Task):
             self.nb_batch_samples_world = input.size(0)
             self.nb_batch_samples_quizzes = 0
 
+        # Shuffle
+        input = input[torch.randperm(input.size(0))]
+
         if desc is None:
             desc = f"epoch-{split}"
         for batch in tqdm.tqdm(
@@ -174,11 +174,13 @@ class World(Task):
             result = input.clone() * (1 - ar_mask)
 
             masked_inplace_autoregression(
-                model,
-                self.batch_size,
-                result,
-                ar_mask,
-                deterministic_synthesis,
+                model=model,
+                batch_size=self.batch_size,
+                input=result,
+                ar_mask=ar_mask,
+                summed_logits=None,
+                temperature=1.0,
+                deterministic_synthesis=deterministic_synthesis,
                 progress_bar_desc=None,
                 device=self.device,
             )
@@ -212,11 +214,13 @@ class World(Task):
         result = input.clone() * (1 - ar_mask)
 
         masked_inplace_autoregression(
-            model,
-            self.batch_size,
-            result,
-            ar_mask,
-            deterministic_synthesis,
+            model=model,
+            batch_size=self.batch_size,
+            input=result,
+            ar_mask=ar_mask,
+            summed_logits=None,
+            temperature=1.0,
+            deterministic_synthesis=deterministic_synthesis,
             progress_bar_desc=None,
             device=self.device,
         )
@@ -230,6 +234,14 @@ class World(Task):
 
         return main_test_accuracy
 
+    def renew_samples(self, nb, for_train=True):
+        input = self.train_input if for_train else self.test_input
+        nb = min(nb, input.size(0))
+        input[:-nb] = input[nb:].clone()
+        input[-nb:] = world.generate_seq(nb, height=self.height, width=self.width).to(
+            self.device
+        )
+
     def store_new_quizzes(self, new_quizzes, for_train=True):
         if for_train:
             self.train_quizzes.append(new_quizzes)
@@ -244,6 +256,7 @@ class World(Task):
         nb,
         model,
         other_models,
+        desired_average_logits=None,
     ):
         ###############################################################
         # Generate quizzes with model
@@ -251,17 +264,48 @@ class World(Task):
         quizzes = torch.empty(
             nb, self.height * self.width * 2 + 1, device=self.device, dtype=torch.int64
         )
+
         ar_mask = torch.full(quizzes.size(), 1, device=self.device)
+        summed_logits = torch.empty(nb, device=self.device)
 
-        masked_inplace_autoregression(
-            model,
-            self.batch_size,
-            quizzes,
-            ar_mask,
-            deterministic_synthesis=False,
-            progress_bar_desc="creating quizzes",
-            device=self.device,
-        )
+        temperature = 1
+        d_temperature = 1
+
+        while True:
+            summed_logits[...] = 0
+
+            masked_inplace_autoregression(
+                model=model,
+                batch_size=self.batch_size,
+                input=quizzes,
+                ar_mask=ar_mask,
+                summed_logits=summed_logits,
+                temperature=temperature,
+                deterministic_synthesis=False,
+                progress_bar_desc="creating quizzes",
+                device=self.device,
+            )
+
+            average_logits = summed_logits.mean()
+
+            logger(f"{average_logits=} {desired_average_logits=}")
+
+            if desired_average_logits is None:
+                break
+
+            # Oh man that's ugly
+            if average_logits < desired_average_logits * 1.1:
+                if d_temperature > 0:
+                    d_temperature *= -0.5
+                temperature += d_temperature
+            elif average_logits > desired_average_logits:
+                if d_temperature < 0:
+                    d_temperature *= -0.5
+                temperature += d_temperature
+            else:
+                break
+
+            logger(f"changing temperature to {temperature}")
 
         ###############################################################
         # Create the reverse quizzes
@@ -287,10 +331,12 @@ class World(Task):
             result = quizzes.clone()
 
             masked_inplace_autoregression(
-                m,
-                self.batch_size,
-                result,
-                ar_mask,
+                model=m,
+                batch_size=self.batch_size,
+                input=result,
+                ar_mask=ar_mask,
+                summed_logits=None,
+                temperature=1.0,
                 deterministic_synthesis=True,
                 progress_bar_desc="solving quizzes",
                 device=self.device,
@@ -301,10 +347,12 @@ class World(Task):
             reverse_result = reverse_quizzes.clone()
 
             masked_inplace_autoregression(
-                m,
-                self.batch_size,
-                reverse_result,
-                ar_mask,
+                model=m,
+                batch_size=self.batch_size,
+                input=reverse_result,
+                ar_mask=ar_mask,
+                summed_logits=None,
+                temperature=1.0,
                 deterministic_synthesis=True,
                 progress_bar_desc="solving reversed quizzes",
                 device=self.device,
@@ -318,9 +366,9 @@ class World(Task):
 
         nb_correct = torch.cat(nb_correct, dim=0)
 
-        filename = os.path.join(result_dir, "correct_{n_epoch:04d}.dat")
-        with open(filename, "w") as f:
-            for k in nb_correct:
-                f.write(f"{k}\n")
+        # filename = os.path.join(result_dir, "correct_{n_epoch:04d}.dat")
+        # with open(filename, "w") as f:
+        # for k in nb_correct:
+        # f.write(f"{k}\n")
 
-        return quizzes, nb_correct.sum(dim=0)
+        return quizzes, nb_correct.sum(dim=0), summed_logits.mean()