Multi-thread call paddlespeech_client asr, server raise Exception: You need to initialize the beam_search_decoder firstly

[Scisaga]

重现步骤

1. 启动服务端

   paddlespeech_server start --config_file config.yaml

   • 配置文件：

     
     host: 0.0.0.0
     port: 5000
     protocol: 'http'
     engine_list: ['asr_python', 'text_python']

   asr_python: model: 'deepspeech2online_wenetspeech' lang: 'zh' sample_rate: 16000 cfg_path: # [optional] ckpt_path: # [optional] decode_method: 'attention_rescoring' num_decoding_left_chunks: -1 force_yes: True device: 'cpu'

   text_python: task: punc model_type: 'ernie_linear_p3_wudao' lang: 'zh' sample_rate: 16000 cfg_path: # [optional] ckpt_path: # [optional] vocab_file: # [optional] device: 'cpu' # set 'gpu:id' or 'cpu'

2. 在两个shell中同时运行

   paddlespeech_client asr --server_ip 127.0.0.1 --port 5000 --input cache/AxC9es3hCYFfiauh-0000.wav

3. shell1正确获得识别结果，shell2报错
4. 服务端异常

   [2023-07-20 23:59:38,693] [    INFO] - You need to initialize the beam_search_decoder firstly
   Traceback (most recent call last):
   File "/opt/conda/lib/python3.10/site-packages/paddlespeech/server/engine/asr/python/asr_engine.py", line 120, in run
   self.infer(self.asr_engine.config.model)
   File "/opt/conda/lib/python3.10/site-packages/decorator.py", line 232, in fun
   return caller(func, *(extras + args), **kw)
   File "/opt/conda/lib/python3.10/site-packages/paddle/fluid/dygraph/base.py", line 347, in _decorate_function
   return func(*args, **kwargs)
   File "/opt/conda/lib/python3.10/site-packages/paddlespeech/cli/asr/infer.py", line 306, in infer
   result_transcripts = self.model.decode(audio, audio_len)
   File "/opt/conda/lib/python3.10/site-packages/decorator.py", line 232, in fun
   return caller(func, *(extras + args), **kw)
   File "/opt/conda/lib/python3.10/site-packages/paddle/fluid/dygraph/base.py", line 347, in _decorate_function
   return func(*args, **kwargs)
   File "/opt/conda/lib/python3.10/site-packages/paddlespeech/s2t/models/ds2/deepspeech2.py", line 303, in decode
   self.decoder.reset_decoder(batch_size=batch_size)
   File "/opt/conda/lib/python3.10/site-packages/paddlespeech/s2t/modules/ctc.py", line 463, in reset_decoder
   raise Exception(
   Exception: You need to initialize the beam_search_decoder firstly

通过什么方式进行配置可以有效的进行多线程调用？是否与以下信息有关？

!!! Since PaddlePaddle support 0-D tensor from 2.5.0, PaddleSpeech Static model will not work for it, please re-export static model.

[zxcd]

目前不支持进行多线程调用，跟您提到的错误信息无关。

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.