interpretable-graph-classification/config.yml at master · tsKenneth/interpretable-graph-classification · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
general:
  data_autobalance: False
  print_dataset_features: True
  batch_size: 1
  extract_features: False
run:
  num_epochs : 50
  learning_rate : 0.0001
  seed: 1800
  k_fold: 5
GNN_models:
  DGCNN:
    convolution_layers_size : 32-32-32-1
    sortpooling_k : 0.6
    n_hidden : 128
    convolution_dropout: 0.5
    pred_dropout : 0.5
    FP_len : 0
  GCN:
    convolution_layers_size : 128-256-512
    dropout: 0.5
  GCND:
    convolution_layers_size: 128-256-512
    dropout: 0.5
  DiffPool:
    convolution_layers_size : 64-64-64
    pred_hidden_layers: 50-50-50
    assign_ratio: 0.25
    number_of_pooling: 1
    concat_tensors: False
  DiffPoolD:
    convolution_layers_size : 64-64-64
    pred_hidden_layers: 50-50-50
    assign_ratio: 0.25
    number_of_pooling: 1
    concat_tensors: False
interpretability_methods:
  DeepLIFT:
    enabled: False
    compare_with_zero_tensor: True
    compare_with_isomorphic_samples: True
    number_of_zero_tensor_samples: 3
    sample_ids:
    number_of_isomorphic_sample_pairs: 5
  saliency:
    enabled: False
    number_of_samples: 3
    sample_ids:
  LayerGradCAM:
    enabled: True
    number_of_samples: 3
    sample_ids:
    layer: 0
    assign_attribution: hard
metrics:
  fidelity:
    enabled: True
    importance_range: 0.5,1
  contrastivity:
    enabled: True
    importance_range: 0.5,1
  sparsity:
    enabled: True
    importance_range: 0.5,1
custom_visualisation_options:
  GNN_models:
    DiffPool:
      cluster_nodes: True
  dataset:
    MUTAG:
      custom_mapping: {"0": "C","1": "N","2": "O","3": "F","4": "I","5": "Cl","6": "Br"}