-
Notifications
You must be signed in to change notification settings - Fork 0
/
models_params_helper.py
286 lines (260 loc) · 12.4 KB
/
models_params_helper.py
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
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
from ast import literal_eval
def cnn_params(params, dict_cnn):
"""
Reads the dictionary of the detection and classification CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# CNN params
params.batchsize = literal_eval(dict_cnn['batchsize'])
params.nb_conv_layers = literal_eval(dict_cnn['nb_conv_layers']) # nb_conv+1 in total because input layer counted separately
params.nb_dense_layers = literal_eval(dict_cnn['nb_dense_layers']) # nb_dense+1 in total because last layer is not associated to a dropout layer and different nb of nodes
params.nb_filters = literal_eval(dict_cnn['nb_filters'])
params.net_type = dict_cnn['net_type']
params.filter_size = literal_eval(dict_cnn['filter_size'])
params.pool_size = literal_eval(dict_cnn['pool_size'])
params.nb_dense_nodes = literal_eval(dict_cnn['nb_dense_nodes'])
params.dropout_proba = literal_eval(dict_cnn['dropout_proba'])
# Adam
params.learn_rate_adam = literal_eval(dict_cnn['learn_rate_adam'])
params.beta_1 = literal_eval(dict_cnn['beta_1'])
params.beta_2 = literal_eval(dict_cnn['beta_2'])
params.epsilon = literal_eval(dict_cnn['epsilon'])
# early stopping
params.min_delta = literal_eval(dict_cnn['min_delta'])
params.patience = literal_eval(dict_cnn['patience'])
def cnn_params_1(params, dict_cnn):
"""
Reads the dictionary of the detection CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# CNN params
params.batchsize_1 = literal_eval(dict_cnn['batchsize'])
params.nb_conv_layers_1 = literal_eval(dict_cnn['nb_conv_layers']) # nb_conv+1 in total because input layer counted separately
params.nb_dense_layers_1 = literal_eval(dict_cnn['nb_dense_layers']) # nb_dense+1 in total because last layer is not associated to a dropout layer and different nb of nodes
params.nb_filters_1 = literal_eval(dict_cnn['nb_filters'])
params.net_type_1 = dict_cnn['net_type']
params.filter_size_1 = literal_eval(dict_cnn['filter_size'])
params.pool_size_1 = literal_eval(dict_cnn['pool_size'])
params.nb_dense_nodes_1 = literal_eval(dict_cnn['nb_dense_nodes'])
params.dropout_proba_1 = literal_eval(dict_cnn['dropout_proba'])
# Adam
params.learn_rate_adam_1 = literal_eval(dict_cnn['learn_rate_adam'])
params.beta_1_1 = literal_eval(dict_cnn['beta_1'])
params.beta_2_1 = literal_eval(dict_cnn['beta_2'])
params.epsilon_1 = literal_eval(dict_cnn['epsilon'])
# early stopping
params.min_delta_1 = literal_eval(dict_cnn['min_delta'])
params.patience_1 = literal_eval(dict_cnn['patience'])
def cnn_params_2(params, dict_cnn):
"""
Reads the dictionary of the classification CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# CNN params
params.batchsize_2 = literal_eval(dict_cnn['batchsize'])
params.nb_conv_layers_2 = literal_eval(dict_cnn['nb_conv_layers']) # nb_conv+1 in total because input layer counted separately
params.nb_dense_layers_2 = literal_eval(dict_cnn['nb_dense_layers']) # nb_dense+1 in total because last layer is not associated to a dropout layer and different nb of nodes
params.nb_filters_2 = literal_eval(dict_cnn['nb_filters'])
params.net_type_2 = dict_cnn['net_type']
params.filter_size_2 = literal_eval(dict_cnn['filter_size'])
params.pool_size_2 = literal_eval(dict_cnn['pool_size'])
params.nb_dense_nodes_2 = literal_eval(dict_cnn['nb_dense_nodes'])
params.dropout_proba_2 = literal_eval(dict_cnn['dropout_proba'])
# Adam
params.learn_rate_adam_2 = literal_eval(dict_cnn['learn_rate_adam'])
params.beta_1_2 = literal_eval(dict_cnn['beta_1'])
params.beta_2_2 = literal_eval(dict_cnn['beta_2'])
params.epsilon_2 = literal_eval(dict_cnn['epsilon'])
# early stopping
params.min_delta_2 = literal_eval(dict_cnn['min_delta'])
params.patience_2 = literal_eval(dict_cnn['patience'])
def xgboost_params(params, dict_xgboost):
"""
Read the dictionary of the XGBoost
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_xgboost : dict
The keys are the parameter names and they are associated to their values.
"""
params.max_depth = literal_eval(dict_xgboost['max_depth'])
params.eta = literal_eval(dict_xgboost['eta'])
params.min_child_weight = literal_eval(dict_xgboost['min_child_weight'])
params.n_estimators = literal_eval(dict_xgboost['n_estimators'])
params.gamma_xgb = literal_eval(dict_xgboost['gamma_xgb'])
params.subsample = literal_eval(dict_xgboost['subsample'])
params.scale_pos_weight = literal_eval(dict_xgboost['scale_pos_weight'])
def resnet_params(params, dict_resnet):
"""
Reads the dictionary of the detection and classification CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# ResNet params
params.L2_weight_decay = literal_eval(dict_resnet['L2_weight_decay'])
params.batch_norm_decay = literal_eval(dict_resnet['batch_norm_decay'])
params.batch_norm_epsilon = literal_eval(dict_resnet['batch_norm_epsilon'])
# Adam
params.learn_rate_adam = literal_eval(dict_resnet['learn_rate_adam'])
params.beta_1 = literal_eval(dict_resnet['beta_1'])
params.beta_2 = literal_eval(dict_resnet['beta_2'])
params.epsilon = literal_eval(dict_resnet['epsilon'])
params.batchsize = literal_eval(dict_resnet['batchsize'])
# early stopping
params.min_delta = literal_eval(dict_resnet['min_delta'])
params.patience = literal_eval(dict_resnet['patience'])
def resnet_params_1(params, dict_resnet):
"""
Reads the dictionary of the detection and classification CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# ResNet params
params.L2_weight_decay_1 = literal_eval(dict_resnet['L2_weight_decay'])
params.batch_norm_decay_1 = literal_eval(dict_resnet['batch_norm_decay'])
params.batch_norm_epsilon_1 = literal_eval(dict_resnet['batch_norm_epsilon'])
# Adam
params.learn_rate_adam_1 = literal_eval(dict_resnet['learn_rate_adam'])
params.beta_1_1 = literal_eval(dict_resnet['beta_1'])
params.beta_2_1 = literal_eval(dict_resnet['beta_2'])
params.epsilon_1 = literal_eval(dict_resnet['epsilon'])
params.batchsize_1 = literal_eval(dict_resnet['batchsize'])
# early stopping
params.min_delta_1 = literal_eval(dict_resnet['min_delta'])
params.patience_1 = literal_eval(dict_resnet['patience'])
def resnet_params_2(params, dict_resnet):
"""
Reads the dictionary of the detection and classification CNN
and puts the parameters in the corresponding fields of the params object.
Parameters
-----------
params : DataSetParams
Parameters of the model.
dict_cnn : dict
The keys are the parameter names and they are associated to their values.
"""
# ResNet params
params.L2_weight_decay_2 = literal_eval(dict_resnet['L2_weight_decay'])
params.batch_norm_decay_2 = literal_eval(dict_resnet['batch_norm_decay'])
params.batch_norm_epsilon_2 = literal_eval(dict_resnet['batch_norm_epsilon'])
# Adam
params.learn_rate_adam_2 = literal_eval(dict_resnet['learn_rate_adam'])
params.beta_1_2 = literal_eval(dict_resnet['beta_1'])
params.beta_2_2 = literal_eval(dict_resnet['beta_2'])
params.epsilon_2 = literal_eval(dict_resnet['epsilon'])
params.batchsize_2 = literal_eval(dict_resnet['batchsize'])
# early stopping
params.min_delta_2 = literal_eval(dict_resnet['min_delta'])
params.patience_2 = literal_eval(dict_resnet['patience'])
def params_to_dict(params):
"""
Converts the params object into a dictionary depending on the model.
Parameters
-----------
params : DataSetParams
Parameters of the model.
Returns
--------
dic : dict
The keys are the parameter names and they are associated to their values.
"""
dic = {}
if params.classification_model not in ["cnn2", "resnet8", "resnet2", "hybrid_resnet_xgboost"]:
dic["nb_conv_layers"] = params.nb_conv_layers
dic["nb_dense_layers"] = params.nb_dense_layers
dic["nb_filters"] = params.nb_filters
dic["filter_size"] = params.filter_size
dic["pool_size"] = params.pool_size
dic["nb_dense_nodes"] = params.nb_dense_nodes
dic["dropout_proba"] = params.dropout_proba
if params.classification_model not in ["cnn2", "resnet2"]:
#Adam
dic["learn_rate_adam"] = params.learn_rate_adam
dic["beta_1"] = params.beta_1
dic["beta_2"] = params.beta_2
dic["epsilon"] = params.epsilon
# early stopping
dic["min_delta"] = params.min_delta
dic["patience"] = params.patience
# fit
dic['batchsize'] = params.batchsize
if params.classification_model in ["hybrid_cnn_xgboost", "hybrid_resnet_xgboost"]:
dic["eta"] = params.eta
dic["min_child_weight"] = params.min_child_weight
dic["max_depth"] = params.max_depth
dic["n_estimators"] = params.n_estimators
dic["gamma_xgb"] = params.gamma_xgb
dic["subsample"] = params.subsample
dic["scale_pos_weight"] = params.scale_pos_weight
if params.classification_model == "cnn2":
dic["nb_conv_layers_1"] = params.nb_conv_layers_1
dic["nb_dense_layers_1"] = params.nb_dense_layers_1
dic["nb_filters_1"] = params.nb_filters_1
dic["filter_size_1"] = params.filter_size_1
dic["pool_size_1"] = params.pool_size_1
dic["nb_dense_nodes_1"] = params.nb_dense_nodes_1
dic["dropout_proba_1"] = params.dropout_proba_1
dic["nb_conv_layers_2"] = params.nb_conv_layers_2
dic["nb_dense_layers_2"] = params.nb_dense_layers_2
dic["nb_filters_2"] = params.nb_filters_2
dic["filter_size_2"] = params.filter_size_2
dic["pool_size_2"] = params.pool_size_2
dic["nb_dense_nodes_2"] = params.nb_dense_nodes_2
dic["dropout_proba_2"] = params.dropout_proba_2
if params.classification_model in ["cnn2", "resnet2"]:
dic["learn_rate_adam_1"] = params.learn_rate_adam_1
dic["beta_1_1"] = params.beta_1_1
dic["beta_2_1"] = params.beta_2_1
dic["epsilon_1"] = params.epsilon_1
dic["min_delta_1"] = params.min_delta_1
dic["patience_1"] = params.patience_1
dic['batchsize_1'] = params.batchsize_1
dic["learn_rate_adam_2"] = params.learn_rate_adam_2
dic["beta_1_2"] = params.beta_1_2
dic["beta_2_2"] = params.beta_2_2
dic["epsilon_2"] = params.epsilon_2
dic["min_delta_2"] = params.min_delta_2
dic["patience_2"] = params.patience_2
dic['batchsize_2'] = params.batchsize_2
if params.classification_model in ["resnet8", "hybrid_resnet_xgboost"]:
dic['L2_weight_decay'] = params.L2_weight_decay
dic['batch_norm_decay'] = params.batch_norm_decay
dic['batch_norm_epsilon'] = params.batch_norm_epsilon
if params.classification_model == "resnet2":
dic['L2_weight_decay_1'] = params.L2_weight_decay_1
dic['batch_norm_decay_1'] = params.batch_norm_decay_1
dic['batch_norm_epsilon_1'] = params.batch_norm_epsilon_1
dic['L2_weight_decay_2'] = params.L2_weight_decay_2
dic['batch_norm_decay_2'] = params.batch_norm_decay_2
dic['batch_norm_epsilon_2'] = params.batch_norm_epsilon_2
return dic