Update.
authorFrançois Fleuret <francois@fleuret.org>
Wed, 21 Jun 2023 18:10:29 +0000 (20:10 +0200)
committerFrançois Fleuret <francois@fleuret.org>
Wed, 21 Jun 2023 18:10:29 +0000 (20:10 +0200)
main.py
snake.py [new file with mode: 0755]

diff --git a/main.py b/main.py
index 43d2900..9679236 100755 (executable)
--- a/main.py
+++ b/main.py
@@ -39,7 +39,7 @@ parser.add_argument("--result_dir", type=str, default="results_default")
 
 parser.add_argument("--seed", type=int, default=0)
 
-parser.add_argument("--nb_epochs", type=int, default=25)
+parser.add_argument("--nb_epochs", type=int, default=None)
 
 parser.add_argument("--batch_size", type=int, default=None)
 
@@ -100,7 +100,7 @@ parser.add_argument("--snake_height", type=int, default=6)
 
 parser.add_argument("--snake_width", type=int, default=8)
 
-parser.add_argument("--snake_nb_colors", type=int, default=3)
+parser.add_argument("--snake_nb_colors", type=int, default=5)
 
 parser.add_argument("--snake_length", type=int, default=400)
 
@@ -131,15 +131,19 @@ if args.seed >= 0:
 
 default_args = {
     "picoclvr": {
+        "nb_epochs": 25,
         "batch_size": 25,
     },
     "mnist": {
+        "nb_epochs": 25,
         "batch_size": 10,
     },
     "maze": {
+        "nb_epochs": 25,
         "batch_size": 25,
     },
     "snake": {
+        "nb_epochs": 25,
         "batch_size": 20,
     },
 }
@@ -663,106 +667,7 @@ class TaskMaze(Task):
 ######################################################################
 
 
-def generate_snake_sequences(
-    nb, height, width, nb_colors, length, prompt_length, device=torch.device("cpu")
-):
-    worlds = torch.randint(nb_colors, (nb, height, width), device=device)
-    nb_prior_visits = torch.zeros(nb, height, width, device=device)
-
-    # nb x 2
-    snake_position = torch.cat(
-        (
-            torch.randint(height, (nb, 1), device=device),
-            torch.randint(width, (nb, 1), device=device),
-        ),
-        1,
-    )
-    snake_direction = torch.randint(4, (nb,), device=device)
-    sequences = torch.empty(nb, 2 * length, device=device, dtype=torch.int64)
-    sequences_prior_visits = torch.zeros(
-        nb, 2 * length, device=device, dtype=torch.int64
-    )
-    i = torch.arange(nb, device=device)  # [:,None]
-
-    for l in range(length):
-        # nb x 3
-        snake_next_direction = torch.cat(
-            (
-                (snake_direction[:, None] - 1) % 4,
-                snake_direction[:, None],
-                (snake_direction[:, None] + 1) % 4,
-            ),
-            1,
-        )
-
-        # nb x 3
-        vh = (snake_next_direction + 1) % 2 * (snake_next_direction - 1)
-        vw = snake_next_direction % 2 * (snake_next_direction - 2)
-
-        # nb x 3 x 2
-        snake_next_speed = torch.cat((vh[:, :, None], vw[:, :, None]), 2)
-        snake_next_position = snake_position[:, None, :] + snake_next_speed
-
-        # nb x 3
-        val = torch.logical_and(
-            torch.logical_and(
-                snake_next_position[:, :, 0] >= 0, snake_next_position[:, :, 0] < height
-            ),
-            torch.logical_and(
-                snake_next_position[:, :, 1] >= 0, snake_next_position[:, :, 1] < width
-            ),
-        ).float()
-        val = (
-            # The multiplicative factors bias toward moving forward
-            torch.rand_like(val)
-            * val
-            * torch.tensor([[1.0, 2.0, 1.0]], device=device)
-        )
-
-        # nb
-        j = val.argmax(1)
-        snake_direction = snake_next_direction[i, j]
-
-        sequences[:, 2 * l] = worlds[i, snake_position[:, 0], snake_position[:, 1]] + 4
-        sequences_prior_visits[:, 2 * l] = nb_prior_visits[
-            i, snake_position[:, 0], snake_position[:, 1]
-        ]
-        if l < prompt_length:
-            nb_prior_visits[i, snake_position[:, 0], snake_position[:, 1]] += 1
-        sequences[:, 2 * l + 1] = snake_direction
-
-        # nb x 2
-        snake_position = snake_next_position[i, j]
-
-    return sequences, sequences_prior_visits
-
-
-# generate_snake_sequences(nb=1, height=4, width=6, nb_colors=3, length=20)
-# exit(0)
-
-
-def snake_solver(input, ar_mask):
-    for n in range(input.size(0)):
-        i, j, memory = 0, 0, {}
-        # print(input[n])
-        # print(ar_mask[n])
-        for l in range(input.size(1) // 2):
-            if ar_mask[n, 2 * l] == 1:
-                if memory.get((i, j)) is None:
-                    input[n, 2 * l] = -1
-                else:
-                    input[n, 2 * l] = memory[(i, j)]
-            else:
-                # print(f'@3 {memory=}')
-                if memory.get((i, j)) is None:
-                    memory[(i, j)] = input[n, 2 * l]
-                else:
-                    assert memory[(i, j)] == input[n, 2 * l], f"n={n} l={l}"
-            # print(f'@1 {i=} {j=}')
-            d = input[n, 2 * l + 1].item()
-            i += (d + 1) % 2 * (d - 1)
-            j += d % 2 * (d - 2)
-            # print(f'@2 {i=} {j=}')
+import snake
 
 
 class TaskSnake(Task):
@@ -784,7 +689,7 @@ class TaskSnake(Task):
         self.device = device
         self.prompt_length = prompt_length
 
-        self.train_input, self.train_prior_visits = generate_snake_sequences(
+        self.train_input, self.train_prior_visits = snake.generate_sequences(
             nb_train_samples,
             height,
             width,
@@ -793,7 +698,7 @@ class TaskSnake(Task):
             prompt_length,
             self.device,
         )
-        self.test_input, self.test_prior_visits = generate_snake_sequences(
+        self.test_input, self.test_prior_visits = snake.generate_sequences(
             nb_test_samples,
             height,
             width,
@@ -835,7 +740,7 @@ class TaskSnake(Task):
                 )
                 result *= 1 - ar_mask
 
-                # snake_solver(result,ar_mask)
+                # snake.solver(result,ar_mask)
 
                 masked_inplace_autoregression(
                     model, self.batch_size, result, ar_mask, device=self.device
diff --git a/snake.py b/snake.py
new file mode 100755 (executable)
index 0000000..eb46a07
--- /dev/null
+++ b/snake.py
@@ -0,0 +1,145 @@
+#!/usr/bin/env python
+
+# Any copyright is dedicated to the Public Domain.
+# https://creativecommons.org/publicdomain/zero/1.0/
+
+# Written by Francois Fleuret <francois@fleuret.org>
+
+import torch, torchvision
+import torch.nn.functional as F
+
+
+def generate_sequences(
+    nb, height, width, nb_colors, length, prompt_length, device=torch.device("cpu")
+):
+    worlds = torch.randint(nb_colors, (nb, height, width), device=device)
+    nb_prior_visits = torch.zeros(nb, height, width, device=device)
+
+    # nb x 2
+    snake_position = torch.cat(
+        (
+            torch.randint(height, (nb, 1), device=device),
+            torch.randint(width, (nb, 1), device=device),
+        ),
+        1,
+    )
+    snake_direction = torch.randint(4, (nb,), device=device)
+    sequences = torch.empty(nb, 2 * length, device=device, dtype=torch.int64)
+    sequences_prior_visits = torch.zeros(
+        nb, 2 * length, device=device, dtype=torch.int64
+    )
+    i = torch.arange(nb, device=device)  # [:,None]
+
+    for l in range(length):
+        # nb x 3
+        snake_next_direction = torch.cat(
+            (
+                (snake_direction[:, None] - 1) % 4,
+                snake_direction[:, None],
+                (snake_direction[:, None] + 1) % 4,
+            ),
+            1,
+        )
+
+        # nb x 3
+        vh = (snake_next_direction + 1) % 2 * (snake_next_direction - 1)
+        vw = snake_next_direction % 2 * (snake_next_direction - 2)
+
+        # nb x 3 x 2
+        snake_next_speed = torch.cat((vh[:, :, None], vw[:, :, None]), 2)
+        snake_next_position = snake_position[:, None, :] + snake_next_speed
+
+        # nb x 3
+        val = torch.logical_and(
+            torch.logical_and(
+                snake_next_position[:, :, 0] >= 0, snake_next_position[:, :, 0] < height
+            ),
+            torch.logical_and(
+                snake_next_position[:, :, 1] >= 0, snake_next_position[:, :, 1] < width
+            ),
+        ).float()
+        val = (
+            # The multiplicative factors bias toward moving forward
+            torch.rand_like(val)
+            * val
+            * torch.tensor([[1.0, 2.0, 1.0]], device=device)
+        )
+
+        # nb
+        j = val.argmax(1)
+        snake_direction = snake_next_direction[i, j]
+
+        sequences[:, 2 * l] = worlds[i, snake_position[:, 0], snake_position[:, 1]] + 4
+        sequences_prior_visits[:, 2 * l] = nb_prior_visits[
+            i, snake_position[:, 0], snake_position[:, 1]
+        ]
+        if l < prompt_length:
+            nb_prior_visits[i, snake_position[:, 0], snake_position[:, 1]] += 1
+        sequences[:, 2 * l + 1] = snake_direction
+
+        # nb x 2
+        snake_position = snake_next_position[i, j]
+
+    return sequences, sequences_prior_visits
+
+
+# generate_snake_sequences(nb=1, height=4, width=6, nb_colors=3, length=20)
+# exit(0)
+
+
+def solver(input, ar_mask):
+    for n in range(input.size(0)):
+        i, j, memory = 0, 0, {}
+        # print(input[n])
+        # print(ar_mask[n])
+        for l in range(input.size(1) // 2):
+            if ar_mask[n, 2 * l] == 1:
+                if memory.get((i, j)) is None:
+                    input[n, 2 * l] = -1
+                else:
+                    input[n, 2 * l] = memory[(i, j)]
+            else:
+                # print(f'@3 {memory=}')
+                if memory.get((i, j)) is None:
+                    memory[(i, j)] = input[n, 2 * l]
+                else:
+                    assert memory[(i, j)] == input[n, 2 * l], f"n={n} l={l}"
+            # print(f'@1 {i=} {j=}')
+            d = input[n, 2 * l + 1].item()
+            i += (d + 1) % 2 * (d - 1)
+            j += d % 2 * (d - 2)
+            # print(f'@2 {i=} {j=}')
+
+
+######################################################################
+
+if __name__ == "__main__":
+    for n in range(16):
+        descr = generate(nb=1, height=12, width=16)
+
+        print(nb_properties(descr, height=12, width=16))
+
+        with open(f"picoclvr_example_{n:02d}.txt", "w") as f:
+            for d in descr:
+                f.write(f"{d}\n\n")
+
+        img = descr2img(descr, height=12, width=16)
+        if img.size(0) == 1:
+            img = F.pad(img, (1, 1, 1, 1), value=64)
+
+        torchvision.utils.save_image(
+            img / 255.0,
+            f"picoclvr_example_{n:02d}.png",
+            padding=1,
+            nrow=4,
+            pad_value=0.8,
+        )
+
+    import time
+
+    start_time = time.perf_counter()
+    descr = generate(nb=1000, height=12, width=16)
+    end_time = time.perf_counter()
+    print(f"{len(descr) / (end_time - start_time):.02f} samples per second")
+
+######################################################################