aishell example

游雁

2024-02-19 94de39dde2e616a01683c518023d0fab72b4e103

 examples/industrial_data_pretraining/seaco_paraformer/demo.py

@@ -5,16 +5,41 @@

from funasr import AutoModel

model = AutoModel(model="damo/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-common-vocab8404-pytorch",
                  model_revision="v2.0.2",
model = AutoModel(model="iic/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-common-vocab8404-pytorch",
                  model_revision="v2.0.4",
                  vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",
                  vad_model_revision="v2.0.2",
                  vad_model_revision="v2.0.4",
                  punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",
                  punc_model_revision="v2.0.2",
                  punc_model_revision="v2.0.4",
                  spk_model="damo/speech_campplus_sv_zh-cn_16k-common",
                  spk_model="v2.0.2",
                  spk_model_revision="v2.0.2",
                  )

res = model(input=f"{model.model_path}/example/asr_example.wav",
            hotword='达摩院 魔搭')
print(res)

# example1
res = model.generate(input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav",
                     hotword='达摩院 魔搭',
                     # preset_spk_num=2,
                     # sentence_timestamp=True,  # return sentence level information when spk_model is not given
                    )
print(res)


'''
# tensor or numpy as input
# example2
import torchaudio
import os
wav_file = os.path.join(model.model_path, "example/asr_example.wav")
input_tensor, sample_rate = torchaudio.load(wav_file)
input_tensor = input_tensor.mean(0)
res = model.generate(input=[input_tensor], batch_size_s=300, is_final=True)


# example3
import soundfile

wav_file = os.path.join(model.model_path, "example/asr_example.wav")
speech, sample_rate = soundfile.read(wav_file)
res = model.generate(input=[speech], batch_size_s=300, is_final=True)
'''