-
Notifications
You must be signed in to change notification settings - Fork 1
/
rules.py
53 lines (47 loc) · 2.18 KB
/
rules.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
import torch
import torch.nn.functional as torch_functions
class GameOfLife:
def __init__(self, device):
self.parameters = torch.zeros((2, 2, 3, 3), dtype=torch.float32, device=device)
self.parameters[1, 1, :, :] = 1
self.parameters[1, 1, 1, 1] = 9
def __call__(self, state):
next_state = torch_functions.pad(state, (1, 1, 1, 1), mode="circular")
next_state = torch_functions.conv2d(next_state, self.parameters)
next_state = ((next_state == 3) + (next_state == 11) + (next_state == 12)).to(torch.float32)
next_state[:, 0, :, :] = 1 - next_state[:, 1, :, :]
return next_state
class FallingSand:
def __init__(self, device):
self.parameters = torch.zeros((3, 3, 3, 3), dtype=torch.float32, device=device)
self.parameters[1, 1, 1, 1] = 4
self.parameters[1, 1, 2, 0] = 1
self.parameters[1, 1, 2, 1] = 1
self.parameters[1, 1, 2, 2] = 1
self.parameters[1, 2, 1, 1] = 1
self.parameters[2, 2, 0, 1] = 4
self.parameters[2, 2, 0, 0] = 3
self.parameters[2, 1, 1, 0] = 1
self.parameters[2, 2, 0, 2] = 3
self.parameters[2, 1, 1, 2] = 1
self.parameters[2, 1, 1, 1] = -12
def __call__(self, state):
next_state = torch_functions.pad(state, (1, 1, 1, 1), mode="circular")
next_state = torch_functions.conv2d(next_state, self.parameters)
next_state = (next_state > 3).to(torch.float32)
next_state[:,0,:,:] = 1 - next_state.sum(1)
return next_state
class Growth:
def __init__(self, device):
self.parameters = torch.zeros((3, 3, 3, 3), dtype=torch.float32, device=device)
self.parameters[1, 1, 1, 1] = 9
self.parameters[2, 2, :, :] = 1
self.parameters[2, 2, 1, 1] = 9
self.parameters[2, 1, 1, 1] = 8
def __call__(self, state):
next_state = torch_functions.pad(state, (1, 1, 1, 1), mode="circular")
next_state = torch_functions.conv2d(next_state, self.parameters)
next_state = (next_state > 8).to(torch.float32)
next_state[:, 1, :, :] *= 1 - next_state[:, 2, :, :]
next_state[:, 0, :, :] = 1 - next_state.sum(1)
return next_state