lkrsnik/stress_asignment
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Added multiple results and error analysis

Browse Source
This commit is contained in:
lkrsnik
2017-09-07 19:32:36 +02:00
parent 7180550b0d
commit dbd0b90f92
14 changed files with 653 additions and 62 deletions
Download Patch File Download Diff File Expand all files Collapse all files
workbench.py
+21 -13
View File
@@ -31,30 +31,40 @@ from prepare_data import *
# data = Data('l', save_generated_data=False, number_of_syllables=True)
# syllabled letters
data = Data('l', save_generated_data=False, accent_classification=True)
data = Data('s', save_generated_data=False, accent_classification=True)
data.generate_data('letters_word_accetuation_train',
'letters_word_accetuation_test',
'letters_word_accetuation_validate', content_name='SlovarIJS_BESEDE_utf8.lex',
content_shuffle_vector='content_shuffle_vector', shuffle_vector='shuffle_vector',
inputs_location='', content_location='')
# concatenate test and train data
# data.x_train = np.concatenate((data.x_train, data.x_test), axis=0)
# data.x_other_features_train = np.concatenate((data.x_other_features_train, data.x_other_features_test), axis=0)
# data.y_train = np.concatenate((data.y_train, data.y_test), axis=0)
# concatenate all data
data.x_train = np.concatenate((data.x_train, data.x_test, data.x_validate), axis=0)
data.x_other_features_train = np.concatenate((data.x_other_features_train, data.x_other_features_test, data.x_other_features_validate), axis=0)
data.y_train = np.concatenate((data.y_train, data.y_test, data.y_validate), axis=0)
num_examples = len(data.x_train) # training set size
nn_output_dim = 13
nn_hdim = 516
batch_size = 16
# actual_epoch = 1
actual_epoch = 40
actual_epoch = 20
# num_fake_epoch = 2
num_fake_epoch = 20
# letters
conv_input_shape=(23, 36)
# conv_input_shape=(23, 36)
# syllabled letters
# conv_input_shape=(10, 5168)
# conv_input_shape=(10, 252)
# syllables
conv_input_shape=(10, 5168)
# othr_input = (140, )
@@ -62,11 +72,11 @@ othr_input = (150, )
conv_input = Input(shape=conv_input_shape, name='conv_input')
# letters
x_conv = Conv1D(115, (3), padding='same', activation='relu')(conv_input)
x_conv = Conv1D(46, (3), padding='same', activation='relu')(x_conv)
# x_conv = Conv1D(115, (3), padding='same', activation='relu')(conv_input)
# x_conv = Conv1D(46, (3), padding='same', activation='relu')(x_conv)
# syllabled letters
# x_conv = Conv1D(200, (2), padding='same', activation='relu')(conv_input)
x_conv = Conv1D(200, (2), padding='same', activation='relu')(conv_input)
x_conv = MaxPooling1D(pool_size=2)(x_conv)
x_conv = Flatten()(x_conv)
@@ -76,9 +86,9 @@ x = concatenate([x_conv, othr_input])
# x = Dense(1024, input_dim=(516 + 256), activation='relu')(x)
x = Dense(256, activation='relu')(x)
x = Dropout(0.3)(x)
x = Dense(512, activation='relu')(x)
x = Dense(256, activation='relu')(x)
x = Dropout(0.3)(x)
x = Dense(512, activation='relu')(x)
x = Dense(256, activation='relu')(x)
x = Dropout(0.3)(x)
x = Dense(nn_output_dim, activation='sigmoid')(x)
@@ -94,8 +104,6 @@ model.compile(loss='binary_crossentropy', optimizer=opt, metrics=[actual_accurac
history = model.fit_generator(data.generator('train', batch_size, content_name='SlovarIJS_BESEDE_utf8.lex', content_location=''),
data.x_train.shape[0]/(batch_size * num_fake_epoch),
epochs=actual_epoch*num_fake_epoch,
validation_data=data.generator('test', batch_size, content_name='SlovarIJS_BESEDE_utf8.lex', content_location=''),
validation_steps=data.x_test.shape[0]/(batch_size * num_fake_epoch),
verbose=2
)