mohdelgaar commited on
Commit
9f22f23
1 Parent(s): 435bd3d

add missing file

Browse files
Files changed (1) hide show
  1. utils.py +16 -0
utils.py ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from transformers import T5Tokenizer, T5EncoderModel
2
+ import torch
3
+ def sequence_mask(lengths, max_len=None):
4
+ """
5
+ Creates a boolean mask from sequence lengths.
6
+ :param lengths: 1d tensor [batch_size]
7
+ :param max_len: int
8
+ """
9
+ batch_size = lengths.numel()
10
+ max_len = max_len or lengths.max()
11
+ return (torch.arange(0, max_len, device=lengths.device)
12
+ .type_as(lengths)
13
+ .repeat(batch_size, 1)
14
+ .lt(lengths.unsqueeze(1))
15
+ .long())
16
+