speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

support for turning off timestamps

Browse Source
This commit is contained in:
lzr265946 2023-02-10 13:46:01 +08:00
parent bd8f163ee6
commit 7aa2e885f4
8 changed files with 45 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
funasr/bin/asr_inference.py
View File

@ -453,7 +453,7 @@ def inference_modelscope(
ibest_writer["score"][key] = str(hyp.score) ibest_writer["score"][key] = str(hyp.score)
if text is not None: if text is not None:
text_postprocessed = postprocess_utils.sentence_postprocess(token) text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed} item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item) asr_result_list.append(item)
finish_count += 1 finish_count += 1

8
funasr/bin/asr_inference_paraformer.py
View File

@ -428,7 +428,11 @@ def inference_modelscope(
format="%(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s", format="%(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
) )
hotword_list_or_file = param_dict['hotword'] if param_dict is not None:
hotword_list_or_file = param_dict.get('hotword')
else:
hotword_list_or_file = None
if ngpu >= 1 and torch.cuda.is_available(): if ngpu >= 1 and torch.cuda.is_available():
device = "cuda" device = "cuda"
else: else:
@ -539,7 +543,7 @@ def inference_modelscope(
ibest_writer["rtf"][key] = rtf_cur ibest_writer["rtf"][key] = rtf_cur
if text is not None: if text is not None:
text_postprocessed = postprocess_utils.sentence_postprocess(token) text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed} item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item) asr_result_list.append(item)
finish_count += 1 finish_count += 1

2
funasr/bin/asr_inference_paraformer_timestamp.py
View File

@ -436,7 +436,7 @@ def inference(
ibest_writer["score"][key] = str(hyp.score) ibest_writer["score"][key] = str(hyp.score)
if text is not None: if text is not None:
text_postprocessed = postprocess_utils.sentence_postprocess(token) text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed} item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item) asr_result_list.append(item)
finish_count += 1 finish_count += 1

19
funasr/bin/asr_inference_paraformer_vad.py
View File

@ -242,6 +242,11 @@ def inference_modelscope(
inference=True, inference=True,
) )
if param_dict is not None:
use_timestamp = param_dict.get('use_timestamp', True)
else:
use_timestamp = True
finish_count = 0 finish_count = 0
file_count = 1 file_count = 1
lfr_factor = 6 lfr_factor = 6
@ -284,8 +289,10 @@ def inference_modelscope(
text, token, token_int = result[0], result[1], result[2] text, token, token_int = result[0], result[1], result[2]
time_stamp = None if len(result) < 4 else result[3] time_stamp = None if len(result) < 4 else result[3]
if use_timestamp and time_stamp is not None:
postprocessed_result = postprocess_utils.sentence_postprocess(token, time_stamp) postprocessed_result = postprocess_utils.sentence_postprocess(token, time_stamp)
else:
postprocessed_result = postprocess_utils.sentence_postprocess(token)
text_postprocessed = "" text_postprocessed = ""
time_stamp_postprocessed = "" time_stamp_postprocessed = ""
text_postprocessed_punc = postprocessed_result text_postprocessed_punc = postprocessed_result
@ -293,9 +300,11 @@ def inference_modelscope(
text_postprocessed, time_stamp_postprocessed, word_lists = postprocessed_result[0], \ text_postprocessed, time_stamp_postprocessed, word_lists = postprocessed_result[0], \
postprocessed_result[1], \ postprocessed_result[1], \
postprocessed_result[2] postprocessed_result[2]
text_postprocessed_punc = text_postprocessed else:
if len(word_lists) > 0 and text2punc is not None: text_postprocessed, word_lists = postprocessed_result[0], postprocessed_result[1]
text_postprocessed_punc, punc_id_list = text2punc(word_lists, 20) text_postprocessed_punc = text_postprocessed
if len(word_lists) > 0 and text2punc is not None:
text_postprocessed_punc, punc_id_list = text2punc(word_lists, 20)
item = {'key': key, 'value': text_postprocessed_punc} item = {'key': key, 'value': text_postprocessed_punc}

19
funasr/bin/asr_inference_paraformer_vad_punc.py
View File

@ -571,6 +571,11 @@ def inference_modelscope(
inference=True, inference=True,
) )
if param_dict is not None:
use_timestamp = param_dict.get('use_timestamp', True)
else:
use_timestamp = True
finish_count = 0 finish_count = 0
file_count = 1 file_count = 1
lfr_factor = 6 lfr_factor = 6
@ -613,7 +618,10 @@ def inference_modelscope(
text, token, token_int = result[0], result[1], result[2] text, token, token_int = result[0], result[1], result[2]
time_stamp = None if len(result) < 4 else result[3] time_stamp = None if len(result) < 4 else result[3]
postprocessed_result = postprocess_utils.sentence_postprocess(token, time_stamp) if use_timestamp and time_stamp is not None:
postprocessed_result = postprocess_utils.sentence_postprocess(token, time_stamp)
else:
postprocessed_result = postprocess_utils.sentence_postprocess(token)
text_postprocessed = "" text_postprocessed = ""
time_stamp_postprocessed = "" time_stamp_postprocessed = ""
text_postprocessed_punc = postprocessed_result text_postprocessed_punc = postprocessed_result
@ -621,9 +629,12 @@ def inference_modelscope(
text_postprocessed, time_stamp_postprocessed, word_lists = postprocessed_result[0], \ text_postprocessed, time_stamp_postprocessed, word_lists = postprocessed_result[0], \
postprocessed_result[1], \ postprocessed_result[1], \
postprocessed_result[2] postprocessed_result[2]
text_postprocessed_punc = text_postprocessed else:
if len(word_lists) > 0 and text2punc is not None: text_postprocessed, word_lists = postprocessed_result[0], postprocessed_result[1]
text_postprocessed_punc, punc_id_list = text2punc(word_lists, 20)
text_postprocessed_punc = text_postprocessed
if len(word_lists) > 0 and text2punc is not None:
text_postprocessed_punc, punc_id_list = text2punc(word_lists, 20)
item = {'key': key, 'value': text_postprocessed_punc} item = {'key': key, 'value': text_postprocessed_punc}
if text_postprocessed != "": if text_postprocessed != "":

2
funasr/bin/asr_inference_uniasr.py
View File

@ -492,7 +492,7 @@ def inference_modelscope(
ibest_writer["score"][key] = str(hyp.score) ibest_writer["score"][key] = str(hyp.score)
if text is not None: if text is not None:
text_postprocessed = postprocess_utils.sentence_postprocess(token) text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed} item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item) asr_result_list.append(item)
finish_count += 1 finish_count += 1

2
funasr/bin/asr_inference_uniasr_vad.py
View File

@ -492,7 +492,7 @@ def inference_modelscope(
ibest_writer["score"][key] = str(hyp.score) ibest_writer["score"][key] = str(hyp.score)
if text is not None: if text is not None:
text_postprocessed = postprocess_utils.sentence_postprocess(token) text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed} item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item) asr_result_list.append(item)
finish_count += 1 finish_count += 1

6
funasr/utils/postprocess_utils.py
View File

@ -232,5 +232,9 @@ def sentence_postprocess(words: List[Any], time_stamp: List[List] = None):
return sentence, ts_lists, real_word_lists return sentence, ts_lists, real_word_lists
else: else:
word_lists = abbr_dispose(word_lists) word_lists = abbr_dispose(word_lists)
real_word_lists = []
for ch in word_lists:
if ch != ' ':
real_word_lists.append(ch)
sentence = ''.join(word_lists).strip() sentence = ''.join(word_lists).strip()
return sentence return sentence, real_word_lists