diff --git a/egs/gigaspeech2/SSL/local/preprocess_gigaspeech2.py b/egs/gigaspeech2/SSL/local/preprocess_gigaspeech2.py index 7383a5308..3477cc1e5 100755 --- a/egs/gigaspeech2/SSL/local/preprocess_gigaspeech2.py +++ b/egs/gigaspeech2/SSL/local/preprocess_gigaspeech2.py @@ -63,11 +63,16 @@ def normalize_text( text = re.sub(r"\u0039", r"\u0E59", text) # Currency symbols mapping - text = re.sub(r"\u0024", r"\u0E14\u0E2D\u0E25\u0E25\u0E32\u0E23\u0E4C", text) # $ -> ดอลลาร์ - text = re.sub(r"\u00A3", r"\u0E1B\u0E2D\u0E19\u0E14\u0E4C", text) # £ -> ปอนด์ - text = re.sub(r"\u00A5", r"\u0E2E\u0E22\u0E27\u0E31\u0E19", text) # ¥ -> หยวน - text = re.sub(r"\u20AC", r"\u0E22\u0E39\u0E42\u0E23", text) # € -> ยูโร - text = re.sub(r"\u0E3F", r"\u0E1A\u0E32\u0E17", text) # ฿ -> บาท + text = re.sub(r"\u0024", "ดอลลาร์", text) # $ + text = re.sub(r"\u00A3", "ปอนด์", text) # £ + text = re.sub(r"\u00A5", "หยวน", text) # ¥ + text = re.sub(r"\u20AC", "ยูโร", text) # € + text = re.sub(r"\u0E3F", "บาท", text) # ฿ + + # Temperature/Angle symbols mapping + text = re.sub(r"\u00B0\u0043", "องศาเซลเซียส", text) # °C + text = re.sub(r"\u00B0\u0046", "องศาฟาเรนไฮต์", text) # °F + text = re.sub(r"\u00B0", "องศา", text) # ° # Remove blank symbols text = re.sub(r"\s", "", utt)