from modelscope.pipelines import pipeline
|
from modelscope.utils.constant import Tasks
|
|
if __name__ == '__main__':
|
audio_in = 'https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/vad_example.wav'
|
output_dir = "./results"
|
inference_pipeline = pipeline(
|
task=Tasks.auto_speech_recognition,
|
model='damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch',
|
vad_model='damo/speech_fsmn_vad_zh-cn-16k-common-pytorch',
|
punc_model='damo/punc_ct-transformer_cn-en-common-vocab471067-large',
|
output_dir=output_dir,
|
)
|
rec_result = inference_pipeline(audio_in=audio_in, batch_size_token=5000, batch_size_token_threshold_s=40)
|
print(rec_result)
|
