The English model is not working properly

[treya-lin]

🐛 Bug

Hi I am trying the english ASR model paraformer-en , but the output is not in English at all.

To Reproduce

Load the paraformer-en and transcribe an English audio file.

from funasr import AutoModel
# paraformer-zh is a multi-functional asr model
# use vad, punc, spk or not as you need
model = AutoModel(model="paraformer-en",  
                  vad_model="fsmn-vad", 
                  punc_model="ct-punc", 
# spk_model="cam++"
                  )
res = model.generate(input=f"https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_en.wav", 
            batch_size_s=300)
print(res)

Output: (The correct transcription should be "He tried to think how it could be." )

[{'key': 'asr_example_en', 'text': ' Procedure telegram, impatience phuket, sleeping danielot pro.'}]

Code sample

Expected behavior

I hope the English model could transcribe properly.

Environment

• OS (e.g., Linux):
• FunASR Version (e.g., 1.0.0): 1.1.6
• ModelScope Version (e.g., 1.11.0): 1.14.0
• PyTorch Version (e.g., 2.0.0): 2.1.0+cu118
• How you installed funasr (pip, source): pip
• Python version: python 3.10
• GPU (e.g., V100M32): 3090 Ti
• CUDA/cuDNN version (e.g., cuda11.7): 11.8
• Docker version (e.g., funasr-runtime-sdk-cpu-0.4.1): Not using docker
• Any other relevant information:

Additional context

[pvtoan]

Hi,

I have exactly the same problem when input a speech: "How are you doing today" and I got the result:

"sleeping war suspense mm sentosa"

Could you please explain what caused this bug?

[liguiyuan]

@pvtoan Hi, I have the same problem, have you solved it? thanks