paddlespeech asr ... KeyError: 'ASRExecutor' #4127
Description
paddlespeech asr --lang zh --input zh.wav
Traceback (most recent call last):
File "/v6/paddleGPU/venv/bin/paddlespeech", line 7, in
sys.exit(_execute())
~~~~~~~~^^
File "/v6/paddleGPU/venv/lib/python3.13/site-packages/paddlespeech/cli/entry.py", line 41, in _execute
com['_entry'] = locals()[cls]
~~~~~~~~^^^^^
KeyError: 'ASRExecutor'
python
Python 3.13.7 (main, Sep 9 2025, 03:40:15) [GCC 12.5.0] on linux
from paddlespeech.cli.asr.infer import ASRExecutor
... asr = ASRExecutor()
... result = asr(audio_file="zh.wav", lang="zh")
...
Traceback (most recent call last):
File "", line 3, in
result = asr(audio_file="zh.wav", lang="zh")
File "/v6/paddleGPU/venv/lib/python3.13/site-packages/paddlespeech/cli/utils.py", line 328, in _warpper
return executor_func(self, *args, **kwargs)
File "/v6/paddleGPU/venv/lib/python3.13/site-packages/paddlespeech/cli/asr/infer.py", line 502, in call
self._init_from_path(model, lang, codeswitch, sample_rate, config,
~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
decode_method, num_decoding_left_chunks, ckpt_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/v6/paddleGPU/venv/lib/python3.13/site-packages/paddlespeech/cli/asr/infer.py", line 159, in _init_from_path
self.task_resource.set_task_model(tag, version=None)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/v6/paddleGPU/venv/lib/python3.13/site-packages/paddlespeech/resource/resource.py", line 74, in set_task_model
assert model_tag in self.pretrained_models,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Can't find "conformer_u2pp_online_wenetspeech-zh-16k" in resource. Model name must be one of []
I've compile paddlepaddle from git with paddlepaddle_gpu-0.0.0-cp313-cp313-linux_x86_64.whl using
cuda_12.9.1_575.57.08_linux.run and cudnn-linux-x86_64-9.13.0.50_cuda12-archive.tar.xz
paddleocr works great. But I couldn't get paddlespeech to work.
According to the doc, it should automatically download onformer_u2pp_online_wenetspeech-zh-16k but it doesn't seem to do that.
Please help.
Thank you,
Jeff.