[Hackathon 7th] 修复 tal_cs 测试中 0D tensor to 1D

[megemini]

PR types

Bug fixes

PR changes

Others

Describe

修复 tal_cs 测试中 0D tensor to 1D

原代码中

ilen = paddle.to_tensor(feat.shape[0])

在 paddle 3.0 中是个 0D 的 tensor，而这里实际需要的是 1D 的，否则报错

> CUDA_VISIBLE_DEVICES= ./local/test_wav.sh conf/conformer.yaml conf/tuning/decode.yaml exp/conformer/checkpoints/avg_10 data/demo_01_03.wav

 File "/home/aistudio/PaddleSpeech/paddlespeech/s2t/exps/u2/bin/test_wav.py", line 130, in <module>
    main(config, args)
  File "/home/aistudio/PaddleSpeech/paddlespeech/s2t/exps/u2/bin/test_wav.py", line 122, in main
    U2Infer(config, args).run()
  File "/home/aistudio/PaddleSpeech/paddlespeech/s2t/exps/u2/bin/test_wav.py", line 86, in run
    result_transcripts = self.model.decode(
  File "/usr/local/lib/python3.8/dist-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/aistudio/.local/lib/python3.8/site-packages/paddle/base/dygraph/base.py", line 397, in _decorate_function
    return func(*args, **kwargs)
  File "/home/aistudio/PaddleSpeech/paddlespeech/s2t/models/u2/u2.py", line 821, in decode
    hyp = self.attention_rescoring(
  File "/home/aistudio/PaddleSpeech/paddlespeech/s2t/models/u2/u2.py", line 535, in attention_rescoring
    assert speech.shape[0] == speech_lengths.shape[0]
IndexError: list index out of range
Failed in evaluation!

@zxcd @Liyulingyue

[paddle-bot]

Thanks for your contribution!

[zxcd]

类似目录下的test_wav.py均有可能出现此问题，建议一起改了吧

[megemini]

一共找到三个地方，都已修改 ～ 不过，不知道哪些地方会测试用到，没法本地测试 ... ...

[zxcd]

LGTM