From 41481008e2e51eb930799e92fb75cce603637988 Mon Sep 17 00:00:00 2001
From: shixian.shi <shixian.shi@alibaba-inc.com>
Date: 星期二, 27 六月 2023 19:16:14 +0800
Subject: [PATCH] update
---
funasr/datasets/large_datasets/utils/tokenize.py | 13 ++++++++++++-
1 files changed, 12 insertions(+), 1 deletions(-)
diff --git a/funasr/datasets/large_datasets/utils/tokenize.py b/funasr/datasets/large_datasets/utils/tokenize.py
index a7eb6d2..61a1ad0 100644
--- a/funasr/datasets/large_datasets/utils/tokenize.py
+++ b/funasr/datasets/large_datasets/utils/tokenize.py
@@ -54,7 +54,18 @@
length = len(text)
if 'hw_tag' in data:
- hotword_indxs = sample_hotword(length, **hw_config)
+ if hw_config['pre_hwlist'] is not None and hw_config['pre_prob'] > 0:
+ # enable preset hotword detect in sampling
+ pre_index = None
+ 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)]
+ import pdb; pdb.set_trace()
+ 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