aboutsummaryrefslogtreecommitdiff
path: root/swr2_asr/utils/tokenizer.py
diff options
context:
space:
mode:
Diffstat (limited to 'swr2_asr/utils/tokenizer.py')
-rw-r--r--swr2_asr/utils/tokenizer.py8
1 files changed, 3 insertions, 5 deletions
diff --git a/swr2_asr/utils/tokenizer.py b/swr2_asr/utils/tokenizer.py
index 5482bbe..22569eb 100644
--- a/swr2_asr/utils/tokenizer.py
+++ b/swr2_asr/utils/tokenizer.py
@@ -1,8 +1,6 @@
"""Tokenizer for Multilingual Librispeech datasets"""
-
-
-from datetime import datetime
import os
+from datetime import datetime
from tqdm.autonotebook import tqdm
@@ -119,8 +117,8 @@ class CharTokenizer:
line = line.strip()
if line:
char, index = line.split()
- tokenizer.char_map[char] = int(index)
- tokenizer.index_map[int(index)] = char
+ load_tokenizer.char_map[char] = int(index)
+ load_tokenizer.index_map[int(index)] = char
return load_tokenizer