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

update paraformer streaming recipe

Browse Source
This commit is contained in:
仁迷 2023-03-17 23:47:07 +08:00
parent f8c7f71ba1
commit ffd4aac427
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
egs_modelscope/asr/paraformer/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online/infer.py
View File

@ -13,7 +13,7 @@ inference_pipeline = pipeline(
model='damo/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online',
model_revision='v1.0.2')
waveform, sample_rate = torchaudio.load("asr_example_zh.wav")
waveform, sample_rate = torchaudio.load("waihu.wav")
speech_length = waveform.shape[1]
speech = waveform[0]
@ -28,7 +28,7 @@ speech_buffer = speech
speech_cache = []
final_result = ""
while len(speech_buffer) > 0:
while len(speech_buffer) >= 960:
if first_chunk:
if len(speech_buffer) >= 14400:
rec_result = inference_pipeline(audio_in=speech_buffer[0:14400], param_dict=param_dict)
@ -51,7 +51,7 @@ while len(speech_buffer) > 0:
cache_en["pad_right"] = 0
rec_result = inference_pipeline(audio_in=speech_buffer, param_dict=param_dict)
speech_buffer = []
if rec_result['text'] != "sil":
if len(rec_result) !=0 and rec_result['text'] != "sil":
final_result += rec_result['text']
print(rec_result)
print(final_result)