执行wekws/examples/speechcommand_v1/s0/run.sh 报错：KeyError: 'num_layers'

[taorui-plus]

Describe the bug

执行wekws/examples/speechcommand_v1/s0/run.sh
报错：

Traceback (most recent call last):
  File "wekws/bin/export_onnx.py", line 105, in <module>
    main()
  File "wekws/bin/export_onnx.py", line 45, in main
    num_layers = configs['model']['backbone']['num_layers']
KeyError: 'num_layers'

查看代码需要在配置文件中读取这个参数
查看配置文件没有num_layers这个参数

To Reproduce

在配置文件中设置num_layers=16，生成了ONNX文件，但是报错：

Export model successfully, see exp/mdtc_epoch10/avg_10.zip
/dev/shm/taorui/taorui/wekws/wekws/model/mdtc.py:247: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if in_cache.size(0) > 0:
/dev/shm/taorui/taorui/wekws/wekws/model/mdtc.py:106: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if cache.size(0) == 0:
/dev/shm/taorui/taorui/wekws/wekws/model/mdtc.py:110: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  assert outputs.size(2) > self.padding
/dev/shm/taorui/taorui/wekws/wekws/model/mdtc.py:257: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if in_cache.size(0) > 0:
/dev/shm/taorui/taorui/wekws/wekws/model/mdtc.py:187: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
  if in_cache.size(0) > 0:
Receptive Fields: 244
some model information
Export to onnx succeed, but pytorch/onnx have different
                 outputs when given the same input, please check!!!....'

[mlxu995]

@taorui-plus 你可以试下设置 num_layers=4，应该能解决导出的问题。
这个 error 是因为 mdtc 和 tcn 中表示层数的变量名称不一致导致的，tcn 里的 num_layers 实际上对应于 mdtc 中的 num_stack

[zhuleia]

在配置文件中加了num_layers=4,但在执行wekws/examples/speechcommand_v1/s0/run.sh的时候，报以下错误：
/home/zl/work/face/SpeechRecognition/wekws/wekws/model/mdtc.py:247: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if in_cache.size(0) > 0:
/home/zl/work/face/SpeechRecognition/wekws/wekws/model/mdtc.py:106: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if cache.size(0) == 0:
/home/zl/work/face/SpeechRecognition/wekws/wekws/model/mdtc.py:110: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert outputs.size(2) > self.padding
/home/zl/work/face/SpeechRecognition/wekws/wekws/model/mdtc.py:257: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if in_cache.size(0) > 0:
/home/zl/work/face/SpeechRecognition/wekws/wekws/model/mdtc.py:187: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if in_cache.size(0) > 0:
Export to onnx succeed, but pytorch/onnx have different
outputs when given the same input, please check!!!