Skip to content

Commit 2485ed5

Browse files
committed
speedup
1 parent d639b83 commit 2485ed5

File tree

1 file changed

+3
-5
lines changed

1 file changed

+3
-5
lines changed

users/zeyer/datasets/utils/serialize.py

Lines changed: 3 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -115,7 +115,8 @@ def run(self):
115115
print("RETURNN vocab:", vocab)
116116
vocab = Vocabulary.create_vocab(**vocab)
117117
else:
118-
vocab = None
118+
assert dataset.labels[self.data_key]
119+
vocab = Vocabulary.create_vocab_from_labels(dataset.labels[self.data_key])
119120

120121
# noinspection PyBroadException
121122
try:
@@ -153,10 +154,7 @@ def run(self):
153154
f" dataset tag {dataset.get_tag(seq_idx)!r} != seq list tag {seq_list[seq_idx]!r}"
154155
)
155156
data = dataset.get_data(seq_idx, self.data_key)
156-
if vocab:
157-
s = vocab.get_seq_labels(data)
158-
else:
159-
s = dataset.serialize_data(self.data_key, data)
157+
s = vocab.get_seq_labels(data)
160158
for old, new in self.raw_replacement_list:
161159
s = s.replace(old, new)
162160
if self.raw_final_strip:

0 commit comments

Comments
 (0)