Added missing keyword definitions, fixed broken model files, made args consistent.

Author: jmohdyusof

Pilot2/P2B1/p2b1.py

@@ -31,18 +31,27 @@
 {'name':'train_bool', 'type':candle.str2bool,'default':True,'help':'Invoke training'},
 {'name':'eval_bool', 'type':candle.str2bool,'default':False,'help':'Use model for inference'},
 {'name':'home_dir','help':'Home Directory','type':str,'default':'.'},
-{'name':'config_file','help':'Config File','type':str,'default':os.path.join(file_path, 'p2b1_default_model.txt')},
+#{'name':'config_file','help':'Config File','type':str,'default':os.path.join(file_path, 'p2b1_default_model.txt')},
 {'name':'weight_path','help':'Trained Model Pickle File','type':str,'default':None},
 {'name':'base_memo','help':'Memo','type':str,'default':None},
-{'name':'seed', 'type':candle.str2bool,'default':False,'help':'Random Seed'},
+#{'name':'seed_bool', 'type':candle.str2bool,'default':False,'help':'Random Seed'},
 {'name':'case','help':'[Full, Center, CenterZ]','type':str,'default':'Full'},
 {'name':'fig_bool', 'type':candle.str2bool,'default':False,'help':'Generate Prediction Figure'},
 {'name':'set_sel','help':'[3k_Disordered, 3k_Ordered, 3k_Ordered_and_gel, 6k_Disordered, 6k_Ordered, 6k_Ordered_and_gel]','type':str,'default':'3k_Disordered'},
 {'name':'conv_bool', 'type':candle.str2bool, 'default':True, 'help':'Invoke training using 1D Convs for inner AE'},
 {'name':'full_conv_bool', 'type':candle.str2bool, 'default':False, 'help':'Invoke training using fully convolutional NN for inner AE'},
 {'name':'type_bool', 'type':candle.str2bool, 'default':True, 'help':'Include molecule type information in desining AE'},
 {'name':'nbr_type', 'type':str, 'default':'relative', 'help':'Defines the type of neighborhood data to use. [relative, invariant]'},
-{'name':'backend', 'help':'Keras Backend', 'type':str, 'default':'tensorflow'}
+{'name':'backend', 'help':'Keras Backend', 'type':str, 'default':'tensorflow'},
+{'name':'cool', 'help':'Boolean: cool learning rate', 'type':candle.str2bool, 'default':False},
+{'name':'data_set', 'help':'Data set for training', 'type':str, 'default':None},
+{'name':'l2_reg', 'help':'Regularization parameter', 'type':float, 'default':None},
+{'name':'molecular_nbrs', 'help':'Data dimension for molecular autoencoder', 'type':int, 'default':None},
+{'name':'molecular_nonlinearity', 'help':'Activation for molecular netowrk', 'type':str, 'default':None},
+{'name':'molecular_num_hidden', 'nargs':'+', 'help':'Layer sizes for molecular network', 'type':int, 'default':None},
+{'name':'noise_factor', 'help':'Noise factor', 'type':float, 'default':None},
+{'name':'num_hidden', 'nargs':'+', 'help':'Dense layer specification', 'type':int, 'default':None},
+{'name':'sampling_density', 'help':'Sampling density', 'type':float, 'default':None}
 ]
 
 required = [
@@ -61,7 +70,7 @@
     'molecular_num_hidden',
     'molecular_nonlinearity',
     'molecular_nbrs',
-    'drop_prob',
+    'dropout',
     'l2_reg',
     'sampling_density',
     'save_path'

Pilot2/P2B1/p2b1_baseline_keras2.py

@@ -12,7 +12,7 @@
         from importlib import reload  # Python 3.4+
     except ImportError:
         from imp import reload  # Python 3.0 - 3.3
-        
+
 TIMEOUT=3600 # in sec; set this to -1 for no timeout
 file_path = os.path.dirname(os.path.realpath(__file__))
 #lib_path = os.path.abspath(os.path.join(file_path, '..', 'common'))
@@ -22,7 +22,7 @@
 
 from keras import backend as K
 
-import p2b1 
+import p2b1
 import candle
 
 import p2b1_AE_models as AE_models
@@ -53,7 +53,7 @@ def initialize_parameters(default_model = 'p2b1_default_model.txt'):
     print ('\nTraining parameters:')
     for key in sorted(GP):
         print ("\t%s: %s" % (key, GP[key]))
-        
+
     # print json.dumps(GP, indent=4, skipkeys=True, sort_keys=True)
 
     if GP['backend'] != 'theano' and GP['backend'] != 'tensorflow':
@@ -79,8 +79,8 @@ def initialize_parameters(default_model = 'p2b1_default_model.txt'):
 def run(GP):
 
     # set the seed
-    if GP['seed']:
-        np.random.seed(GP['seed'])
+    if GP['rng_seed']:
+        np.random.seed(GP['rng_seed'])
     else:
         np.random.seed(np.random.randint(10000))
 
@@ -211,7 +211,7 @@ def run(GP):
                                                                            nonlinearity=molecular_nonlinearity,
                                                                            hidden_layers=molecular_hidden_layers,
                                                                            l2_reg=GP['l2_reg'],
-                                                                           drop=float(GP['drop_prob']))
+                                                                           drop=float(GP['dropout']))
     elif full_conv_bool:
         molecular_model, molecular_encoder = AE_models.full_conv_mol_auto(bead_k_size=bead_kernel_size,
                                                                           mol_k_size=mol_kernel_size,
@@ -220,14 +220,14 @@ def run(GP):
                                                                           nonlinearity=molecular_nonlinearity,
                                                                           hidden_layers=molecular_hidden_layers,
                                                                           l2_reg=GP['l2_reg'],
-                                                                          drop=float(GP['drop_prob']))
+                                                                          drop=float(GP['dropout']))
 
     else:
         molecular_model, molecular_encoder = AE_models.dense_auto(weights_path=None, input_shape=(molecular_input_dim,),
                                                                   nonlinearity=molecular_nonlinearity,
                                                                   hidden_layers=molecular_hidden_layers,
                                                                   l2_reg=GP['l2_reg'],
-                                                                  drop=float(GP['drop_prob']))
+                                                                  drop=float(GP['dropout']))
 
     if GP['loss'] == 'mse':
         loss_func = 'mse'
@@ -238,7 +238,7 @@ def run(GP):
     print ('\nModel Summary: \n')
     molecular_model.summary()
     ##### set up callbacks and cooling for the molecular_model ##########
-    drop = 0.5
+    drop = GP['dropout']
     mb_epochs = GP['epochs']
     initial_lrate = GP['learning_rate']
     epochs_drop = 1+int(np.floor(mb_epochs/3))

Pilot2/P2B1/p2b1_default_model.txt

@@ -13,7 +13,7 @@ molecular_nonlinearity='elu'
 molecular_num_hidden=[256, 128, 64, 32, 16, 8]
 molecular_nbrs = 200
 base_memo='p2b1'
-drop_prob = 0.5
+dropout = 0.5
 data_set='3k_Ordered'
 sampling_density = 0.15
 save_path='.'

Pilot2/P2B1/p2b1_medium_model.txt

@@ -4,6 +4,15 @@ batch_size=32
 learning_rate=0.01
 epochs=10
 cool=True
-weight_decay=0.0005
+optimizer='adam'
+loss='custom'
+activation='relu'
+molecular_nonlinearity='elu'
+molecular_num_hidden=[256, 128, 64, 32, 16, 8]
+molecular_nbrs = 200
 noise_factor=0
 base_memo='p2b1'
+dropout = 0.5
+l2_reg=0.01
+sampling_density = 0.15
+save_path='.'

Pilot2/P2B1/p2b1_small_model.txt

@@ -4,13 +4,15 @@ batch_size=32
 learning_rate=0.01
 epochs=10
 cool=False
-weight_decay=0.0005
 noise_factor=0.0
 optimizer='adam'
 loss='mse'
 activation='relu'
-molecular_nonlinearity=elu
+molecular_nonlinearity='elu'
 molecular_num_hidden=[54,12]
-molecular_epochs=1
 molecular_nbrs=10
 base_memo='p2b1'
+dropout = 0.5
+l2_reg=0.01
+sampling_density = 0.15
+save_path='.'