# Written by Francois Fleuret <francois@fleuret.org>
-import math, sys, tqdm
+import math, sys, tqdm, os
import torch, torchvision
######################################################################
-colors = torch.tensor(
- [
- [255, 255, 255],
- [255, 0, 0],
- [0, 192, 0],
- [0, 0, 255],
- [255, 192, 0],
- [0, 255, 255],
- [255, 0, 255],
- [192, 255, 192],
- [255, 192, 192],
- [192, 192, 255],
- [192, 192, 192],
- ]
-)
-
-token_background = 0
-first_bird_token = 1
-nb_bird_tokens = colors.size(0) - 1
-token_forward = first_bird_token + nb_bird_tokens
-token_backward = token_forward + 1
-
-token2char = "_" + "".join([chr(ord("A") + n) for n in range(len(colors) - 1)]) + "><"
-
-
-def generate_seq(
- nb, height, width, nb_birds=3, nb_iterations=2, return_iterations=False
-):
- pairs = []
- kept_iterations = []
-
- for _ in tqdm.tqdm(range(nb), dynamic_ncols=True, desc="world generation"):
- while True:
- iterations = []
-
- f_start = torch.zeros(height, width, dtype=torch.int64)
-
- i, j, vi, vj = (
- torch.empty(nb_birds, dtype=torch.int64),
- torch.empty(nb_birds, dtype=torch.int64),
- torch.empty(nb_birds, dtype=torch.int64),
- torch.empty(nb_birds, dtype=torch.int64),
- )
-
- col = torch.randperm(colors.size(0) - 1)[:nb_birds].sort().values + 1
-
- for n in range(nb_birds):
- c = col[n]
-
- while True:
- i[n], j[n] = (
- torch.randint(height, (1,))[0],
- torch.randint(width, (1,))[0],
- )
- vm = torch.randint(4, (1,))[0]
- vi[n], vj[n] = (vm % 2) * 2 - 1, (vm // 2) * 2 - 1
- if (
- i[n] - vi[n] >= 0
- and i[n] - vi[n] < height
- and j[n] - vj[n] >= 0
- and j[n] - vj[n] < width
- and f_start[i[n], j[n]] == 0
- and f_start[i[n] - vi[n], j[n]] == 0
- and f_start[i[n], j[n] - vj[n]] == 0
- ):
- break
-
- f_start[i[n], j[n]] = c
- f_start[i[n] - vi[n], j[n]] = c
- f_start[i[n], j[n] - vj[n]] = c
-
- f_end = f_start.clone()
-
- for l in range(nb_iterations):
- iterations.append(f_end.clone())
- f_end[...] = 0
- nb_collisions = 0
- for n in range(nb_birds):
- c = col[n]
+class Problem:
+ def generate_seq(self, nb_train_samples):
+ pass
- pi, pj, pvi, pvj = (
- i[n].item(),
- j[n].item(),
- vi[n].item(),
- vj[n].item(),
- )
-
- if (i[n] == 0 and vi[n] == -1) or (
- i[n] == height - 1 and vi[n] == 1
- ):
- vi[n] = -vi[n]
- if (j[n] == 0 and vj[n] == -1) or (
- j[n] == width - 1 and vj[n] == 1
- ):
- vj[n] = -vj[n]
-
- i[n] += vi[n]
- j[n] += vj[n]
-
- if not (
- f_end[i[n], j[n]] == 0
- and f_end[i[n] - vi[n], j[n]] == 0
- and f_end[i[n], j[n] - vj[n]] == 0
- ):
- nb_collisions += 1
-
- f_end[i[n], j[n]] = c
- f_end[i[n] - vi[n], j[n]] = c
- f_end[i[n], j[n] - vj[n]] = c
-
- iterations.append(f_end.clone())
-
- if nb_collisions == 0:
- break
-
- kept_iterations.append(iterations)
- pairs.append((f_start, f_end))
-
- result = []
- for p in pairs:
- if torch.rand(1) < 0.5:
- result.append(
- torch.cat(
- [p[0].flatten(), torch.tensor([token_forward]), p[1].flatten()],
- dim=0,
- )[None, :]
- )
- else:
- result.append(
- torch.cat(
- [p[1].flatten(), torch.tensor([token_backward]), p[0].flatten()],
- dim=0,
- )[None, :]
- )
-
- if return_iterations:
- # iterations = torch.cat([ torch.cat([ x[None, None] for x in l], dim = 1) for l in kept_iterations ], dim=0)
- return torch.cat(result, dim=0), kept_iterations
- else:
- return torch.cat(result, dim=0)
+ def save_quizzes(self, input, result_dir, filename_prefix, logger):
+ pass
+ def direction_tokens(self):
+ pass
-######################################################################
+class Sky:
+ colors = torch.tensor(
+ [
+ [255, 255, 255],
+ [255, 0, 0],
+ [0, 192, 0],
+ [0, 0, 255],
+ [255, 192, 0],
+ [0, 255, 255],
+ [255, 0, 255],
+ [192, 255, 192],
+ [255, 192, 192],
+ [192, 192, 255],
+ [192, 192, 192],
+ ]
+ )
+
+ token_background = 0
+ first_bird_token = 1
+ nb_bird_tokens = colors.size(0) - 1
+ token_forward = first_bird_token + nb_bird_tokens
+ token_backward = token_forward + 1
+
+ token2char = (
+ "_" + "".join([chr(ord("A") + n) for n in range(len(colors) - 1)]) + "><"
+ )
+
+ def __init__(self, height=6, width=8, nb_birds=3, nb_iterations=2):
+ self.height = height
+ self.width = width
+ self.nb_birds = nb_birds
+ self.nb_iterations = nb_iterations
+
+ def direction_tokens(self):
+ return self.token_forward, self.token_backward
+
+ def generate_seq(self, nb, return_iterations=False):
+ pairs = []
+ kept_iterations = []
+
+ for _ in tqdm.tqdm(range(nb), dynamic_ncols=True, desc="world generation"):
+ while True:
+ iterations = []
+
+ f_start = torch.zeros(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),
+ torch.empty(self.nb_birds, dtype=torch.int64),
+ torch.empty(self.nb_birds, dtype=torch.int64),
+ )
+
+ col = (
+ torch.randperm(self.colors.size(0) - 1)[: self.nb_birds]
+ .sort()
+ .values
+ + 1
+ )
+
+ for n in range(self.nb_birds):
+ c = col[n]
-def generate_seq_old(
- nb,
- height,
- width,
- nb_birds=3,
- nb_iterations=2,
-):
- pairs = []
-
- for n in tqdm.tqdm(range(nb), dynamic_ncols=True, desc="world generation"):
- f_start = torch.zeros(height, width, dtype=torch.int64)
- f_end = torch.zeros(height, width, dtype=torch.int64)
- n = torch.arange(f_start.size(0))
-
- for c in (
- (torch.randperm(nb_bird_tokens) + first_bird_token)[:nb_birds].sort().values
- ):
- i, j = (
- torch.randint(height - 2, (1,))[0] + 1,
- torch.randint(width - 2, (1,))[0] + 1,
- )
- vm = torch.randint(4, (1,))[0]
- vi, vj = (vm // 2) * (2 * (vm % 2) - 1), (1 - vm // 2) * (2 * (vm % 2) - 1)
-
- f_start[i, j] = c
- f_start[i - vi, j - vj] = c
- f_start[i + vj, j - vi] = c
- f_start[i - vj, j + vi] = c
-
- for l in range(nb_iterations):
- i += vi
- j += vj
- if i < 0 or i >= height or j < 0 or j >= width:
- i -= vi
- j -= vj
- vi, vj = -vi, -vj
+ while True:
+ i[n], j[n] = (
+ torch.randint(self.height, (1,))[0],
+ torch.randint(self.width, (1,))[0],
+ )
+ vm = torch.randint(4, (1,))[0]
+ 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
+ and f_start[i[n], j[n]] == 0
+ and f_start[i[n] - vi[n], j[n]] == 0
+ and f_start[i[n], j[n] - vj[n]] == 0
+ ):
+ break
+
+ f_start[i[n], j[n]] = c
+ f_start[i[n] - vi[n], j[n]] = c
+ f_start[i[n], j[n] - vj[n]] = c
+
+ f_end = f_start.clone()
+
+ for l in range(self.nb_iterations):
+ iterations.append(f_end.clone())
+ f_end[...] = 0
+ nb_collisions = 0
+ for n in range(self.nb_birds):
+ c = col[n]
+
+ pi, pj, pvi, pvj = (
+ i[n].item(),
+ j[n].item(),
+ vi[n].item(),
+ vj[n].item(),
+ )
+
+ 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]
+
+ i[n] += vi[n]
+ j[n] += vj[n]
+
+ if not (
+ f_end[i[n], j[n]] == 0
+ and f_end[i[n] - vi[n], j[n]] == 0
+ and f_end[i[n], j[n] - vj[n]] == 0
+ ):
+ nb_collisions += 1
+
+ f_end[i[n], j[n]] = c
+ f_end[i[n] - vi[n], j[n]] = c
+ f_end[i[n], j[n] - vj[n]] = c
+
+ iterations.append(f_end.clone())
+
+ if nb_collisions == 0:
+ break
+
+ kept_iterations.append(iterations)
+ pairs.append((f_start, f_end))
+
+ result = []
+ for p in pairs:
+ if torch.rand(1) < 0.5:
+ result.append(
+ torch.cat(
+ [
+ p[0].flatten(),
+ torch.tensor([self.token_forward]),
+ p[1].flatten(),
+ ],
+ dim=0,
+ )[None, :]
+ )
+ else:
+ result.append(
+ torch.cat(
+ [
+ p[1].flatten(),
+ torch.tensor([self.token_backward]),
+ p[0].flatten(),
+ ],
+ dim=0,
+ )[None, :]
+ )
+
+ if return_iterations:
+ # iterations = torch.cat([ torch.cat([ x[None, None] for x in l], dim = 1) for l in kept_iterations ], dim=0)
+ return torch.cat(result, dim=0), kept_iterations
+ else:
+ return torch.cat(result, dim=0)
+
+ ######################################################################
+
+ def generate_seq_old(
+ self,
+ nb,
+ ):
+ pairs = []
+
+ for n in tqdm.tqdm(range(nb), dynamic_ncols=True, desc="world generation"):
+ f_start = torch.zeros(self.height, self.width, dtype=torch.int64)
+ f_end = torch.zeros(self.height, self.width, dtype=torch.int64)
+ n = torch.arange(f_start.size(0))
+
+ for c in (
+ (torch.randperm(self.nb_bird_tokens) + self.first_bird_token)[
+ : self.nb_birds
+ ]
+ .sort()
+ .values
+ ):
+ i, j = (
+ torch.randint(self.height - 2, (1,))[0] + 1,
+ torch.randint(self.width - 2, (1,))[0] + 1,
+ )
+ vm = torch.randint(4, (1,))[0]
+ vi, vj = (vm // 2) * (2 * (vm % 2) - 1), (1 - vm // 2) * (
+ 2 * (vm % 2) - 1
+ )
+
+ f_start[i, j] = c
+ f_start[i - vi, j - vj] = c
+ f_start[i + vj, j - vi] = c
+ f_start[i - vj, j + vi] = c
+
+ for l in range(self.nb_iterations):
i += vi
j += vj
+ if i < 0 or i >= self.height or j < 0 or j >= self.width:
+ i -= vi
+ j -= vj
+ vi, vj = -vi, -vj
+ i += vi
+ j += vj
+
+ f_end[i, j] = c
+ f_end[i - vi, j - vj] = c
+ f_end[i + vj, j - vi] = c
+ f_end[i - vj, j + vi] = c
+
+ pairs.append((f_start, f_end))
+
+ result = []
+ for p in pairs:
+ if torch.rand(1) < 0.5:
+ result.append(
+ torch.cat(
+ [
+ p[0].flatten(),
+ torch.tensor([self.token_forward]),
+ p[1].flatten(),
+ ],
+ dim=0,
+ )[None, :]
+ )
+ else:
+ result.append(
+ torch.cat(
+ [
+ p[1].flatten(),
+ torch.tensor([self.token_backward]),
+ p[0].flatten(),
+ ],
+ dim=0,
+ )[None, :]
+ )
- f_end[i, j] = c
- f_end[i - vi, j - vj] = c
- f_end[i + vj, j - vi] = c
- f_end[i - vj, j + vi] = c
-
- pairs.append((f_start, f_end))
-
- result = []
- for p in pairs:
- if torch.rand(1) < 0.5:
- result.append(
- torch.cat(
- [p[0].flatten(), torch.tensor([token_forward]), p[1].flatten()],
- dim=0,
- )[None, :]
- )
- else:
- result.append(
- torch.cat(
- [p[1].flatten(), torch.tensor([token_backward]), p[0].flatten()],
- dim=0,
- )[None, :]
- )
-
- return torch.cat(result, dim=0)
-
-
-def frame2img(x, height, width, upscale=15):
- x = x.reshape(-1, height, width)
- m = torch.logical_and(x >= 0, x < first_bird_token + nb_bird_tokens).long()
- x = colors[x * m].permute(0, 3, 1, 2)
- s = x.shape
- x = x[:, :, :, None, :, None].expand(-1, -1, -1, upscale, -1, upscale)
- x = x.reshape(s[0], s[1], s[2] * upscale, s[3] * upscale)
-
- x[:, :, :, torch.arange(0, x.size(3), upscale)] = 0
- x[:, :, torch.arange(0, x.size(2), upscale), :] = 0
- x = x[:, :, 1:, 1:]
-
- for n in range(m.size(0)):
- for i in range(m.size(1)):
- for j in range(m.size(2)):
- if m[n, i, j] == 0:
- for k in range(2, upscale - 2):
- x[n, :, i * upscale + k, j * upscale + k] = 0
- x[n, :, i * upscale + upscale - 1 - k, j * upscale + k] = 0
-
- return x
-
-
-def seq2img(seq, height, width, upscale=15):
- f_first = seq[:, : height * width].reshape(-1, height, width)
- f_second = seq[:, height * width + 1 :].reshape(-1, height, width)
- direction = seq[:, height * width]
-
- direction_symbol = torch.full((direction.size(0), height * upscale - 1, upscale), 0)
- direction_symbol = colors[direction_symbol].permute(0, 3, 1, 2)
- separator = torch.full((direction.size(0), 3, height * upscale - 1, 1), 0)
-
- for n in range(direction_symbol.size(0)):
- if direction[n] == token_forward:
- for k in range(upscale):
- direction_symbol[
- n,
- :,
- (height * upscale) // 2 - upscale // 2 + k,
- 3 + upscale // 2 - abs(k - upscale // 2),
- ] = 0
- elif direction[n] == token_backward:
- for k in range(upscale):
- direction_symbol[
- n,
- :,
- (height * upscale) // 2 - upscale // 2 + k,
- 3 + abs(k - upscale // 2),
- ] = 0
- else:
- for k in range(2, upscale - 2):
- direction_symbol[
- n, :, (height * upscale) // 2 - upscale // 2 + k, k
- ] = 0
- direction_symbol[
- n, :, (height * upscale) // 2 - upscale // 2 + k, upscale - 1 - k
- ] = 0
-
- return torch.cat(
- [
- frame2img(f_first, height, width, upscale),
- separator,
- direction_symbol,
- separator,
- frame2img(f_second, height, width, upscale),
- ],
- dim=3,
- )
+ return torch.cat(result, dim=0)
+ def frame2img(self, x, upscale=15):
+ x = x.reshape(-1, self.height, self.width)
+ m = torch.logical_and(
+ x >= 0, x < self.first_bird_token + self.nb_bird_tokens
+ ).long()
+ x = self.colors[x * m].permute(0, 3, 1, 2)
+ s = x.shape
+ x = x[:, :, :, None, :, None].expand(-1, -1, -1, upscale, -1, upscale)
+ x = x.reshape(s[0], s[1], s[2] * upscale, s[3] * upscale)
+
+ x[:, :, :, torch.arange(0, x.size(3), upscale)] = 0
+ x[:, :, torch.arange(0, x.size(2), upscale), :] = 0
+ x = x[:, :, 1:, 1:]
+
+ for n in range(m.size(0)):
+ for i in range(m.size(1)):
+ for j in range(m.size(2)):
+ if m[n, i, j] == 0:
+ for k in range(2, upscale - 2):
+ x[n, :, i * upscale + k, j * upscale + k] = 0
+ x[n, :, i * upscale + upscale - 1 - k, j * upscale + k] = 0
+
+ return x
+
+ def seq2img(self, seq, upscale=15):
+ f_first = seq[:, : self.height * self.width].reshape(
+ -1, self.height, self.width
+ )
+ f_second = seq[:, self.height * self.width + 1 :].reshape(
+ -1, self.height, self.width
+ )
+ direction = seq[:, self.height * self.width]
+
+ direction_symbol = torch.full(
+ (direction.size(0), self.height * upscale - 1, upscale), 0
+ )
+ direction_symbol = self.colors[direction_symbol].permute(0, 3, 1, 2)
+ separator = torch.full((direction.size(0), 3, self.height * upscale - 1, 1), 0)
+
+ for n in range(direction_symbol.size(0)):
+ if direction[n] == self.token_forward:
+ for k in range(upscale):
+ direction_symbol[
+ n,
+ :,
+ (self.height * upscale) // 2 - upscale // 2 + k,
+ 3 + upscale // 2 - abs(k - upscale // 2),
+ ] = 0
+ elif direction[n] == self.token_backward:
+ for k in range(upscale):
+ direction_symbol[
+ n,
+ :,
+ (self.height * upscale) // 2 - upscale // 2 + k,
+ 3 + abs(k - upscale // 2),
+ ] = 0
+ else:
+ for k in range(2, upscale - 2):
+ direction_symbol[
+ n, :, (self.height * upscale) // 2 - upscale // 2 + k, k
+ ] = 0
+ direction_symbol[
+ n,
+ :,
+ (self.height * upscale) // 2 - upscale // 2 + k,
+ upscale - 1 - k,
+ ] = 0
+
+ return torch.cat(
+ [
+ self.frame2img(f_first, upscale),
+ separator,
+ direction_symbol,
+ separator,
+ self.frame2img(f_second, upscale),
+ ],
+ dim=3,
+ )
-def seq2str(seq):
- result = []
- for s in seq:
- result.append("".join([token2char[v] for v in s]))
- return result
+ def seq2str(self, seq):
+ result = []
+ for s in seq:
+ result.append("".join([self.token2char[v] for v in s]))
+ return result
+
+ def save_image(self, input, result_dir, filename, logger):
+ img = self.seq2img(input.to("cpu"))
+ 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}")
+
+ def save_quizzes(self, input, result_dir, filename_prefix, logger):
+ self.save_image(input, result_dir, filename_prefix + ".png", logger)
######################################################################
if __name__ == "__main__":
import time
- height, width = 6, 8
+ sky = Sky(height=6, width=8, nb_iterations=100)
+
start_time = time.perf_counter()
- seq, it = generate_seq(
- nb=64, height=height, width=width, nb_iterations=100, return_iterations=True
- )
+ seq, it = sky.generate_seq(nb=64, return_iterations=True)
delay = time.perf_counter() - start_time
print(f"{seq.size(0)/delay:02f} samples/s")
- print(seq2str(seq[:4]))
+ print(sky.seq2str(seq[:4]))
for t in range(len(it[0])):
- img = torch.cat([frame2img(f[t], height, width) for f in it], dim=0)
+ img = torch.cat([sky.frame2img(f[t]) for f in it], dim=0)
torchvision.utils.save_image(
img.float() / 255.0,
f"/tmp/frame_{t:03d}.png",
# m = (torch.rand(seq.size()) < 0.05).long()
# seq = (1 - m) * seq + m * 23
- img = seq2img(seq, height, width)
+ img = sky.seq2img(seq)
print(img.size())
torchvision.utils.save_image(