1 # File used to store hyperparameters.
3 # Stateless RNN, batch_size declared at fit step
8 rnn_layers: 1 # Number of hidden recurrent layers
9 rnn_units: 20 # Number of units per hidden recurrent layer
10 dense_layers: 1 # hidden dense layers AFTER recurrent layers and BEFORE final output cell
11 dense_units: 5 # number of units for hidden dense layers
12 activation: ['linear', 'linear'] # Activation type for hidden layers, dense layers respectively
14 dropout: [0.2, 0.2] # NOTE: length must match total number of layers, default is 1 hidden recurrent layer and 1 dense output layer
15 recurrent_dropout: 0.2 # Length must match number of recurrent layers
16 reset_states: True # reset hidden states after training epoch, triggers reset_states() via callbacks
17 batch_reset: 10 # reset states after given number of batches
20 clipvalue: 10.0 # gradient clipping param, gradient can't exceed this value
21 phys_initialize: False # physics initialization
23 verbose_weights: True # Prints out hashs of weights for tracking reproducibility
24 verbose_fit: False # Prints out all training epochs, makes computation much slower
25 # features_list: ['Ed', 'Ew', 'solar', 'wind', 'rain']
26 features_list: ['Ed', 'Ew', 'rain']
28 scaler: 'minmax' # One of methods in scalers dictionary in moisture_rnn.py
29 train_frac: 0.5 # percent of input data to be used in training set
30 val_frac: 0.1 # percent of input data to be used in validation set. Test set size determined from train_frac and val_frac
41 activation: ['linear', 'linear']
42 recurrent_activation: 'sigmoid'
44 dropout: [0.2, 0.2] # NOTE: length must match total number of layers, default is 1 hidden recurrent layer and 1 dense output layer
45 recurrent_dropout: 0.2 # Length must match number of recurrent layers
46 reset_states: True # reset hidden states after training epoch, triggers reset_states() via callbacks
47 batch_reset: 150 # reset states after given number of batches
50 clipvalue: 1.0 # gradient clipping param, gradient can't exceed this value
51 phys_initialize: False # physics initialization
53 verbose_weights: True # Prints out hashs of weights for tracking reproducibility
54 verbose_fit: False # Prints out all training epochs, makes computation much slower
55 features_list: ['Ed', 'Ew', 'rain']
57 scaler: 'minmax' # One of methods in scalers dictionary in moisture_rnn.py
62 # Param sets for reproducibility
73 activation: ['linear', 'linear']
76 recurrent_dropout: 0.2
82 phys_initialize: False
86 features_list: ['Ed', 'Ew', 'solar', 'wind', 'rain']
100 # dense_layers: 0 # hidden dense layers AFTER recurrent layers and BEFORE final output cell
102 # activation: ['linear', 'linear']
103 # centering: [0.0,0.0]
104 # dropout: [0.0, 0.0] # NOTE: length must match total number of layers, default is 1 hidden recurrent layer and 1 dense output layer
105 # recurrent_dropout: 0.0 # Length must match number of recurrent layers
106 # reset_states: True # reset hidden states after training epoch, triggers reset_states() via callbacks
108 # learning_rate: 0.001
109 # phys_initialize: False # physics initialization
111 # verbose_weights: True # Prints out hashs of weights for tracking reproducibility
112 # verbose_fit: False # Prints out all training epochs, makes computation much slower
113 # features_list: ['Ed', 'Ew', 'rain']
115 # scaler: 'reproducibility'