-
Notifications
You must be signed in to change notification settings - Fork 4
/
utils.py
38 lines (28 loc) · 1.05 KB
/
utils.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
import torch
import shutil
import os
def data_mask(x, pad_idx):
mask = (x != pad_idx).unsqueeze(1)
return mask
def target_mask(x, pad_idx):
mask = (x != pad_idx).unsqueeze(-2)
b, l = x.size()
mask = mask & torch.tril(torch.ones(1, l, l), 0).byte().type_as(mask.data)
return mask
def save_checkpoint(state, is_best, epoch, save_path):
filename = f'checkpoint_e{epoch:02d}.pth.tar'
save_name = os.path.join(save_path, 'model', filename)
torch.save(state, save_name)
if is_best:
shutil.move(save_name, os.path.join(save_path, 'model_best.pth.tar'))
def __test():
dummy_data = [1, 2, 14, 6, 3, 0, 0, 0, 0, 0]
dummy_target = [1, 345, 33, 21, 4, 5, 5, 3, 0, 0]
dummy_data = torch.tensor(dummy_data).unsqueeze(0)
dummy_target = torch.tensor(dummy_target).unsqueeze(0).long()
dummy_data_mask = data_mask(dummy_data, 0)
dummy_target_mask = target_mask(dummy_target, 0)
print(dummy_data_mask)
print(dummy_target_mask)
if __name__ == '__main__':
__test()