IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an email to s dot adaszewski at gmail dot com. User accounts are meant only to report issues and/or generate pull requests. This is a purpose-specific Git hosting for ADARED projects. Thank you for your understanding!
Você não pode selecionar mais de 25 tópicos Os tópicos devem começar com uma letra ou um número, podem incluir traços ('-') e podem ter até 35 caracteres.

136 linhas
4.1KB

  1. from triacontagon.util import \
  2. _clear_adjacency_matrix_except_rows, \
  3. _sparse_diag_cat, \
  4. _equal, \
  5. common_one_hot_encoding
  6. from triacontagon.model import TrainingBatch
  7. from triacontagon.decode import dedicom_decoder
  8. from triacontagon.data import Data
  9. import torch
  10. import time
  11. def test_clear_adjacency_matrix_except_rows_01():
  12. adj_mat = torch.tensor([
  13. [0, 0, 1, 0, 0],
  14. [0, 0, 0, 1, 1],
  15. [1, 0, 1, 0, 0],
  16. [1, 1, 0, 0, 0]
  17. ], dtype=torch.uint8).to_sparse()
  18. adj_mat = _sparse_diag_cat([ adj_mat, adj_mat ])
  19. res = _clear_adjacency_matrix_except_rows(adj_mat,
  20. torch.tensor([1, 3]), 4, 2)
  21. res = res.to_dense()
  22. truth = torch.tensor([
  23. [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
  24. [0, 0, 0, 1, 1, 0, 0, 0, 0, 0],
  25. [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
  26. [1, 1, 0, 0, 0, 0, 0, 0, 0, 0],
  27. [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
  28. [0, 0, 0, 0, 0, 0, 0, 0, 1, 1],
  29. [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
  30. [0, 0, 0, 0, 0, 1, 1, 0, 0, 0]
  31. ], dtype=torch.uint8)
  32. print('res:', res)
  33. assert torch.all(res == truth)
  34. def test_clear_adjacency_matrix_except_rows_02():
  35. adj_mat = torch.rand(6, 10).round().to(torch.uint8)
  36. t = time.time()
  37. res = _sparse_diag_cat([ adj_mat.to_sparse() ] * 130)
  38. print('_sparse_diag_cat() took:', time.time() - t)
  39. t = time.time()
  40. res = _clear_adjacency_matrix_except_rows(res, torch.tensor([1, 3, 5]),
  41. 6, 130)
  42. print('_clear_adjacency_matrix_except_rows() took:', time.time() - t)
  43. adj_mat[0] = adj_mat[2] = adj_mat[4] = \
  44. torch.zeros(10)
  45. truth = _sparse_diag_cat([ adj_mat.to_sparse() ] * 130)
  46. assert _equal(res, truth).all()
  47. def test_clear_adjacency_matrix_except_rows_03():
  48. adj_mat = torch.rand(6, 10).round().to(torch.uint8)
  49. t = time.time()
  50. res = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  51. print('_sparse_diag_cat() took:', time.time() - t)
  52. t = time.time()
  53. res = _clear_adjacency_matrix_except_rows(res, torch.tensor([1, 3, 5]),
  54. 6, 1300)
  55. print('_clear_adjacency_matrix_except_rows() took:', time.time() - t)
  56. adj_mat[0] = adj_mat[2] = adj_mat[4] = \
  57. torch.zeros(10)
  58. truth = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  59. assert _equal(res, truth).all()
  60. def test_clear_adjacency_matrix_except_rows_04():
  61. adj_mat = (torch.rand(2000, 2000) < 0.001).to(torch.uint8)
  62. print('adj_mat.to_sparse():', adj_mat.to_sparse())
  63. t = time.time()
  64. res = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  65. print('_sparse_diag_cat() took:', time.time() - t)
  66. t = time.time()
  67. res = _clear_adjacency_matrix_except_rows(res, torch.tensor([1, 3, 5]),
  68. 2000, 1300)
  69. print('_clear_adjacency_matrix_except_rows() took:', time.time() - t)
  70. adj_mat[0] = adj_mat[2] = adj_mat[4] = \
  71. torch.zeros(2000)
  72. adj_mat[6:] = torch.zeros(2000)
  73. truth = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  74. assert _equal(res, truth).all()
  75. def test_clear_adjacency_matrix_except_rows_05():
  76. if torch.cuda.device_count() == 0:
  77. pytest.skip('Test requires CUDA')
  78. device = torch.device('cuda:0')
  79. adj_mat = (torch.rand(2000, 2000) < 0.001).to(torch.uint8).to(device)
  80. print('adj_mat.to_sparse():', adj_mat.to_sparse())
  81. t = time.time()
  82. res = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  83. print('_sparse_diag_cat() took:', time.time() - t)
  84. rows = torch.tensor(list(range(512)), device=device)
  85. t = time.time()
  86. res = _clear_adjacency_matrix_except_rows(res, rows,
  87. 2000, 1300)
  88. print('_clear_adjacency_matrix_except_rows() took:', time.time() - t)
  89. adj_mat[512:] = torch.zeros(2000)
  90. truth = _sparse_diag_cat([ adj_mat.to_sparse() ] * 1300)
  91. assert _equal(res, truth).all()
  92. def test_common_one_hot_encoding_01():
  93. in_repr = common_one_hot_encoding([2000, 200])
  94. ref = torch.eye(2200)
  95. assert torch.all(in_repr[0].to_dense() == ref[:2000, :])
  96. assert torch.all(in_repr[1].to_dense() == ref[2000:, :])