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!
Du kan inte välja fler än 25 ämnen Ämnen måste starta med en bokstav eller siffra, kan innehålla bindestreck ('-') och vara max 35 tecken långa.

124 lines
3.7KB

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