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.

53 linhas
2.0KB

  1. from icosagon.batch import PredictionsBatch
  2. from icosagon.declayer import Predictions, \
  3. RelationPredictions, \
  4. RelationFamilyPredictions
  5. from icosagon.trainprep import prepare_training, \
  6. TrainValTest
  7. from icosagon.data import Data
  8. import torch
  9. def test_predictions_batch_01():
  10. d = Data()
  11. d.add_node_type('Dummy', 5)
  12. fam = d.add_relation_family('Dummy-Dummy', 0, 0, False)
  13. fam.add_relation_type('Dummy Rel', torch.tensor([
  14. [0, 1, 0, 0, 0],
  15. [1, 0, 0, 0, 0],
  16. [0, 0, 0, 1, 0],
  17. [0, 0, 0, 0, 1],
  18. [0, 1, 0, 0, 0]
  19. ], dtype=torch.float32))
  20. prep_d = prepare_training(d, TrainValTest(1., 0., 0.))
  21. assert len(prep_d.relation_families) == 1
  22. assert len(prep_d.relation_families[0].relation_types) == 1
  23. assert len(prep_d.relation_families[0].relation_types[0].edges_pos.train) == 5
  24. assert len(prep_d.relation_families[0].relation_types[0].edges_neg.train) == 5
  25. assert len(prep_d.relation_families[0].relation_types[0].edges_pos.val) == 0
  26. assert len(prep_d.relation_families[0].relation_types[0].edges_pos.test) == 0
  27. rel_pred = RelationPredictions(
  28. TrainValTest(torch.tensor([1, 0, 1, 0, 1], dtype=torch.float32), torch.zeros(0), torch.zeros(0)),
  29. TrainValTest(torch.tensor([1, 0, 1, 0, 1], dtype=torch.float32), torch.zeros(0), torch.zeros(0)),
  30. TrainValTest(torch.zeros(0), torch.zeros(0), torch.zeros(0)),
  31. TrainValTest(torch.zeros(0), torch.zeros(0), torch.zeros(0))
  32. )
  33. fam_pred = RelationFamilyPredictions([ rel_pred ])
  34. pred = Predictions([ fam_pred ])
  35. batch = PredictionsBatch(pred, part_type='train', batch_size=1)
  36. count = 0
  37. lst = []
  38. for (input, target) in batch:
  39. assert len(input) == 1
  40. assert len(target) == 1
  41. lst.append((input[0], target[0]))
  42. count += 1
  43. assert lst == [ (1, 1), (0, 1), (1, 1), (0, 1), (1, 1),
  44. (1, 0), (0, 0), (1, 0), (0, 0), (1, 0) ]
  45. assert count == 10