From c6574bf4f4390a9c0dfdd5aa2c4f39eb85e1557e Mon Sep 17 00:00:00 2001
From: Shi Xian <40013335+R1ckShi@users.noreply.github.com>
Date: 星期二, 09 四月 2024 10:41:11 +0800
Subject: [PATCH] Merge pull request #1596 from wanghuii1/main
---
funasr/datasets/large_datasets/utils/tokenize.py | 12 +++++++++++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/funasr/datasets/large_datasets/utils/tokenize.py b/funasr/datasets/large_datasets/utils/tokenize.py
index a7eb6d2..34a97c1 100644
--- a/funasr/datasets/large_datasets/utils/tokenize.py
+++ b/funasr/datasets/large_datasets/utils/tokenize.py
@@ -54,7 +54,17 @@
length = len(text)
if 'hw_tag' in data:
- hotword_indxs = sample_hotword(length, **hw_config)
+ pre_index = None
+ if hw_config['pre_hwlist'] is not None and hw_config['pre_prob'] > 0:
+ # enable preset hotword detect in sampling
+ for hw in hw_config['pre_hwlist']:
+ hw = " ".join(seg_tokenize(hw, seg_dict))
+ _find = " ".join(text).find(hw)
+ if _find != -1:
+ # _find = text[:_find].count(" ") # bpe sometimes
+ pre_index = [_find, _find + max(hw.count(" "), 1)]
+ break
+ hotword_indxs = sample_hotword(length, **hw_config, pre_index=pre_index)
data['hotword_indxs'] = hotword_indxs
del data['hw_tag']
for i in range(length):
--
Gitblit v1.9.1