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 oracle num for asr with spk

Browse Source
This commit is contained in:
shixian.shi 2024-01-11 19:16:51 +08:00
parent f6c82b1c3e
commit d72a4497a5
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
funasr/bin/inference.py
View File

@ -137,6 +137,9 @@ class AutoModel:
if spk_mode not in ["default", "vad_segment", "punc_segment"]:
logging.error("spk_mode should be one of default, vad_segment and punc_segment.")
self.spk_mode = spk_mode
self.preset_spk_num = kwargs.get("preset_spk_num", None)
if self.preset_spk_num:
logging.warning("Using preset speaker number: {}".format(self.preset_spk_num))
logging.warning("Many to print when using speaker model...")
self.kwargs = kwargs
@ -397,7 +400,7 @@ class AutoModel:
if self.spk_model is not None:
all_segments = sorted(all_segments, key=lambda x: x[0])
spk_embedding = result['spk_embedding']
labels = self.cb_model(spk_embedding)
labels = self.cb_model(spk_embedding, oracle_num=self.preset_spk_num)
del result['spk_embedding']
sv_output = postprocess(all_segments, None, labels, spk_embedding)
if self.spk_mode == 'vad_segment':