zzy34407230
/
mindspore2022
Not watched
1
0
Fork 0
Code
Releases 22 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
Tree: cdd3e50e98
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'cdd3e50e98'
${ noResults }
mindspore2022/tests/ut/python/dataset/test_melscale_fbanks.py

149 lines
7.5 kB
Raw Blame History 

  1. # Copyright 2022 Huawei Technologies Co., Ltd

  2. #

  3. # Licensed under the Apache License, Version 2.0 (the "License");

  4. # you may not use this file except in compliance with the License.

  5. # You may obtain a copy of the License at

  6. #

  7. # http://www.apache.org/licenses/LICENSE-2.0

  8. #

  9. # Unless required by applicable law or agreed to in writing, software

  10. # distributed under the License is distributed on an "AS IS" BASIS,

  11. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  12. # See the License for the specific language governing permissions and

  13. # limitations under the License.

  14. # ==============================================================================

  15. import numpy as np

  16. import pytest

  17. 

  18. import mindspore.dataset.audio.utils as audio

  19. from mindspore import log as logger

  20. 

  21. 

  22. def count_unequal_element(data_expected, data_me, rtol, atol):

  23.     assert data_expected.shape == data_me.shape

  24.     total_count = len(data_expected.flatten())

  25.     error = np.abs(data_expected - data_me)

  26.     greater = np.greater(error, atol + np.abs(data_expected) * rtol)

  27.     loss_count = np.count_nonzero(greater)

  28.     assert (loss_count / total_count) < rtol, \

  29.         "\ndata_expected_std:{0}\ndata_me_error:{1}\nloss:{2}". \

  30.         format(data_expected[greater], data_me[greater], error[greater])

  31. 

  32. 

  33. def test_melscale_fbanks_normal():

  34.     """

  35.     Feature: melscale_fbanks.

  36.     Description: Test normal operation with NormType.NONE and MelType.HTK.

  37.     Expectation: The output data is the same as the result of torchaudio.functional.melscale_fbanks.

  38.     """

  39.     expect = np.array([[0.0000, 0.0000, 0.0000, 0.0000],

  40.                        [0.5502, 0.0000, 0.0000, 0.0000],

  41.                        [0.6898, 0.3102, 0.0000, 0.0000],

  42.                        [0.0000, 0.9366, 0.0634, 0.0000],

  43.                        [0.0000, 0.1924, 0.8076, 0.0000],

  44.                        [0.0000, 0.0000, 0.4555, 0.5445],

  45.                        [0.0000, 0.0000, 0.0000, 0.7247],

  46.                        [0.0000, 0.0000, 0.0000, 0.0000]], dtype=np.float64)

  47.     output = audio.melscale_fbanks(8, 2, 50, 4, 100, audio.NormType.NONE, audio.MelType.HTK)

  48.     count_unequal_element(expect, output, 0.0001, 0.0001)

  49. 

  50. 

  51. def test_melscale_fbanks_none_slaney():

  52.     """

  53.     Feature: melscale_fbanks.

  54.     Description: Test normal operation with NormType.NONE and MelType.SLANEY.

  55.     Expectation: The output data is the same as the result of torchaudio.functional.melscale_fbanks.

  56.     """

  57.     expect = np.array([[0.0000, 0.0000, 0.0000, 0.0000],

  58.                        [0.5357, 0.0000, 0.0000, 0.0000],

  59.                        [0.7202, 0.2798, 0.0000, 0.0000],

  60.                        [0.0000, 0.9762, 0.0238, 0.0000],

  61.                        [0.0000, 0.2321, 0.7679, 0.0000],

  62.                        [0.0000, 0.0000, 0.4881, 0.5119],

  63.                        [0.0000, 0.0000, 0.0000, 0.7440],

  64.                        [0.0000, 0.0000, 0.0000, 0.0000]], dtype=np.float64)

  65.     output = audio.melscale_fbanks(8, 2, 50, 4, 100, audio.NormType.NONE, audio.MelType.SLANEY)

  66.     count_unequal_element(expect, output, 0.0001, 0.0001)

  67. 

  68. 

  69. def test_melscale_fbanks_with_slaney_htk():

  70.     """

  71.     Feature: melscale_fbanks.

  72.     Description: Test normal operation with NormType.SLANEY and MelType.HTK.

  73.     Expectation: The output data is the same as the result of torchaudio.functional.melscale_fbanks.

  74.     """

  75.     output = audio.melscale_fbanks(10, 0, 50, 5, 100, audio.NormType.SLANEY, audio.MelType.HTK)

  76.     expect = np.array([[0.0000, 0.0000, 0.0000, 0.0000, 0.0000],

  77.                        [0.0843, 0.0000, 0.0000, 0.0000, 0.0000],

  78.                        [0.0776, 0.0447, 0.0000, 0.0000, 0.0000],

  79.                        [0.0000, 0.1158, 0.0055, 0.0000, 0.0000],

  80.                        [0.0000, 0.0344, 0.0860, 0.0000, 0.0000],

  81.                        [0.0000, 0.0000, 0.0741, 0.0454, 0.0000],

  82.                        [0.0000, 0.0000, 0.0000, 0.1133, 0.0053],

  83.                        [0.0000, 0.0000, 0.0000, 0.0355, 0.0822],

  84.                        [0.0000, 0.0000, 0.0000, 0.0000, 0.0760],

  85.                        [0.0000, 0.0000, 0.0000, 0.0000, 0.0000]], dtype=np.float64)

  86.     count_unequal_element(expect, output, 0.0001, 0.0001)

  87. 

  88. 

  89. def test_melscale_fbanks_with_slaney_slaney():

  90.     """

  91.     Feature: melscale_fbanks.

  92.     Description: Test normal operation with NormType.SLANEY and MelType.SLANEY.

  93.     Expectation: The output data is the same as the result of torchaudio.functional.melscale_fbanks.

  94.     """

  95.     output = audio.melscale_fbanks(8, 2, 50, 4, 100, audio.NormType.SLANEY, audio.MelType.SLANEY)

  96.     expect = np.array([[0.0000, 0.0000, 0.0000, 0.0000],

  97.                        [0.0558, 0.0000, 0.0000, 0.0000],

  98.                        [0.0750, 0.0291, 0.0000, 0.0000],

  99.                        [0.0000, 0.1017, 0.0025, 0.0000],

  100.                        [0.0000, 0.0242, 0.0800, 0.0000],

  101.                        [0.0000, 0.0000, 0.0508, 0.0533],

  102.                        [0.0000, 0.0000, 0.0000, 0.0775],

  103.                        [0.0000, 0.0000, 0.0000, 0.0000]], dtype=np.float64)

  104.     count_unequal_element(expect, output, 0.0001, 0.0001)

  105. 

  106. 

  107. def test_melscale_fbanks_invalid_input():

  108.     """

  109.     Feature: melscale_fbanks.

  110.     Description: Test operation with invalid input.

  111.     Expectation: Throw exception as expected.

  112.     """

  113. 

  114.     def test_invalid_input(test_name, n_freqs, f_min, f_max, n_mels, sample_rate, norm, mel_type, error, error_msg):

  115.         logger.info("Test melscale_fbanks with bad input: {0}".format(test_name))

  116.         with pytest.raises(error) as error_info:

  117.             audio.melscale_fbanks(n_freqs, f_min, f_max, n_mels, sample_rate, norm, mel_type)

  118.         print(error_info)

  119.         assert error_msg in str(error_info.value)

  120. 

  121.     test_invalid_input("invalid n_freqs parameter Value", 99999999999, 0, 50, 5, 100, audio.NormType.NONE,

  122.                        audio.MelType.HTK, ValueError, "n_freqs")

  123.     test_invalid_input("invalid n_freqs parameter type", 10.5, 0, 50, 5, 100, audio.NormType.NONE, audio.MelType.HTK,

  124.                        TypeError, "n_freqs")

  125.     test_invalid_input("invalid f_min parameter type", 10, None, 50, 5, 100, audio.NormType.NONE, audio.MelType.HTK,

  126.                        TypeError, "f_min")

  127.     test_invalid_input("invalid f_max parameter type", 10, 0, None, 5, 100, audio.NormType.NONE, audio.MelType.HTK,

  128.                        TypeError, "f_max")

  129.     test_invalid_input("invalid n_mels parameter type", 10, 0, 50, 10.1, 100, audio.NormType.NONE, audio.MelType.HTK,

  130.                        TypeError, "n_mels")

  131.     test_invalid_input("invalid n_mels parameter Value", 20, 0, 50, 999999999999, 100, audio.NormType.NONE,

  132.                        audio.MelType.HTK, ValueError, "n_mels")

  133.     test_invalid_input("invalid sample_rate parameter type", 10, 0, 50, 5, 100.1, audio.NormType.NONE,

  134.                        audio.MelType.HTK, TypeError, "sample_rate")

  135.     test_invalid_input("invalid sample_rate parameter Value", 20, 0, 50, 5, 999999999999, audio.NormType.NONE,

  136.                        audio.MelType.HTK, ValueError, "sample_rate")

  137.     test_invalid_input("invalid norm parameter type", 10, 0, 50, 5, 100, None, audio.MelType.HTK,

  138.                        TypeError, "norm")

  139.     test_invalid_input("invalid norm parameter type", 10, 0, 50, 5, 100, audio.NormType.SLANEY, None,

  140.                        TypeError, "mel_type")

  141. 

  142. 

  143. if __name__ == "__main__":

  144.     test_melscale_fbanks_normal()

  145.     test_melscale_fbanks_none_slaney()

  146.     test_melscale_fbanks_with_slaney_htk()

  147.     test_melscale_fbanks_with_slaney_slaney()

  148.     test_melscale_fbanks_invalid_input()