malar
/
speech-scoring
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

using a Bi-LSTM layer as the first layer

master
Malar Kannan 2017-11-17 14:17:12 +05:30
parent 6ff052be9b
commit c682962c8f
1 changed files with 3 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
speech_model.py
View File

@ -3,7 +3,7 @@ from __future__ import print_function
import numpy as np import numpy as np
from speech_data import read_siamese_tfrecords_generator from speech_data import read_siamese_tfrecords_generator
from keras.models import Model,load_model,model_from_yaml from keras.models import Model,load_model,model_from_yaml
from keras.layers import Input, Dense, Dropout, LSTM, Lambda, Concatenate from keras.layers import Input, Dense, Dropout, LSTM, Lambda, Concatenate, Bidirectional
from keras.losses import categorical_crossentropy from keras.losses import categorical_crossentropy
from keras.utils import to_categorical from keras.utils import to_categorical
from keras.optimizers import RMSprop from keras.optimizers import RMSprop
@ -17,7 +17,7 @@ def create_base_rnn_network(input_dim):
''' '''
inp = Input(shape=input_dim) inp = Input(shape=input_dim)
# ls0 = LSTM(512, return_sequences=True)(inp) # ls0 = LSTM(512, return_sequences=True)(inp)
ls1 = LSTM(256, return_sequences=True)(inp) ls1 = Bidirectional(LSTM(256, return_sequences=True))(inp)
ls2 = LSTM(128, return_sequences=True)(ls1) ls2 = LSTM(128, return_sequences=True)(ls1)
# ls3 = LSTM(32, return_sequences=True)(ls2) # ls3 = LSTM(32, return_sequences=True)(ls2)
ls4 = LSTM(64)(ls2) ls4 = LSTM(64)(ls2)
@ -55,10 +55,6 @@ def siamese_model(input_dim):
processed_b = base_network(input_b) processed_b = base_network(input_b)
final_output = dense_classifier([processed_a,processed_b]) final_output = dense_classifier([processed_a,processed_b])
model = Model([input_a, input_b], final_output) model = Model([input_a, input_b], final_output)
# distance = Lambda(
# euclidean_distance,
# output_shape=eucl_dist_output_shape)([processed_a, processed_b])
# model = Model([input_a, input_b], distance)
return model return model
def write_model_arch(mod,mod_file): def write_model_arch(mod,mod_file):
@ -73,7 +69,7 @@ def load_model_arch(mod_file):
return mod return mod
def train_siamese(audio_group = 'audio'): def train_siamese(audio_group = 'audio'):
batch_size = 256 batch_size = 128
model_dir = './models/'+audio_group model_dir = './models/'+audio_group
create_dir(model_dir) create_dir(model_dir)
log_dir = './logs/'+audio_group log_dir = './logs/'+audio_group