[S2T]Absurd basic bug

[Imbernoulli]

(cad) (deepspeech-venv) root@8ikb0p7c24u18-0:/bjzhyai03/bohan/PaddleSpeech# paddlespeech asr --lang zh --input zh.wav
/root/miniconda3/envs/cad/lib/python3.8/site-packages/setuptools/sandbox.py:14: DeprecationWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html
  import pkg_resources
/root/miniconda3/envs/cad/lib/python3.8/site-packages/pkg_resources/__init__.py:2832: DeprecationWarning: Deprecated call to `pkg_resources.declare_namespace('mpl_toolkits')`.
Implementing implicit namespace packages (as specified in PEP 420) is preferred to `pkg_resources.declare_namespace`. See https://setuptools.pypa.io/en/latest/references/keywords.html#keyword-namespace-packages
  declare_namespace(pkg)
/root/miniconda3/envs/cad/lib/python3.8/site-packages/pkg_resources/__init__.py:2832: DeprecationWarning: Deprecated call to `pkg_resources.declare_namespace('google')`.
Implementing implicit namespace packages (as specified in PEP 420) is preferred to `pkg_resources.declare_namespace`. See https://setuptools.pypa.io/en/latest/references/keywords.html#keyword-namespace-packages
  declare_namespace(pkg)
/root/miniconda3/envs/cad/lib/python3.8/site-packages/librosa/core/constantq.py:1059: DeprecationWarning: `np.complex` is a deprecated alias for the builtin `complex`. To silence this warning, use `complex` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.complex128` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  dtype=np.complex,
/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddle/fluid/dygraph/math_op_patch.py:275: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.float32, but right dtype is paddle.int64, the right dtype will convert to paddle.float32
  warnings.warn(
[2024-03-21 14:01:24,841] [   ERROR] - (InvalidArgument) Broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [1, 1, 0, 498] and the shape of Y = [1, 123, 123]. Received [498] in X is not equal to [123] in Y at i:3.
  [Hint: Expected x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1 == true, but received x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1:0 != true:1.] (at /paddle/paddle/phi/kernels/funcs/common_shape.h:84)
Traceback (most recent call last):
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/cli/asr/infer.py", line 314, in infer
    result_transcripts = self.model.decode(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddle/fluid/dygraph/base.py", line 375, in _decorate_function
    return func(*args, **kwargs)
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/models/u2/u2.py", line 818, in decode
    hyp = self.attention_rescoring(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/models/u2/u2.py", line 543, in attention_rescoring
    hyps, encoder_out = self._ctc_prefix_beam_search(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/models/u2/u2.py", line 424, in _ctc_prefix_beam_search
    encoder_out, encoder_mask = self._forward_encoder(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/models/u2/u2.py", line 229, in _forward_encoder
    encoder_out, encoder_mask = self.encoder(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddle/fluid/dygraph/layers.py", line 948, in __call__
    return self.forward(*inputs, **kwargs)
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/modules/encoder.py", line 184, in forward
    chunk_masks = add_optional_chunk_mask(
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddlespeech/s2t/modules/mask.py", line 202, in add_optional_chunk_mask
    chunk_masks = masks.logical_and(chunk_masks)  # (B, L, L)
  File "/root/miniconda3/envs/cad/lib/python3.8/site-packages/paddle/tensor/logic.py", line 122, in logical_and
    return _C_ops.logical_and(x, y)
ValueError: (InvalidArgument) Broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [1, 1, 0, 498] and the shape of Y = [1, 123, 123]. Received [498] in X is not equal to [123] in Y at i:3.
  [Hint: Expected x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1 == true, but received x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1:0 != true:1.] (at /paddle/paddle/phi/kernels/funcs/common_shape.h:84)

[Ray961123]

开发者你好，感谢关注 PaddleSpeech 开源项目，抱歉给你带来了不好的开发体验，目前开源项目维护人力有限，建议参考：#3692