"filters": 32,
"kernel_size": 3,
"choice_pooling": {
"pooling": True,
"pool_size": 2
},
"neurones": 100,
"batch_size": 50,
"optimizer": "adam",
"activation": "sigmoid",
"nc": args.nc,
"dropout": 0,
"bi": args.bi,
"cost": args.cost
}
create_model(params, rootw=rootw, wn=args.initw, typem=args.typem)
"""
print(X_train.shape, y_train.shape)
print(y_train[::40], np.mean(y_train, axis=0))
print(X_train.dtype, y_train.dtype)
# for yi in y_train:
# print(yi)
# , validation_data=(X_val, y_val[::, 0], y_val[::, 1])
if args.incweightT is not None:
print(np.mean(y_train, axis=0))
y_train[y_train[::, 0] == 0, 1] *= args.incweightT
print(np.mean(y_train, axis=0))
print("Accuracy: %.2f%%" % (scores))
"""
def model(typem=1, base=False, nc=1, window_length=None):
init = 1
if base:
init = 5
if typem == 1:
if window_length is None:
lenv = 200
else:
lenv = window_length
"""
model = Sequential()
# model.add(Embedding(top_words, embedding_vecor_length,
input_length=max_review_length))
model.add(Conv1D(filters=32, kernel_size=3, padding='same',
activation='relu', input_shape=(lenv, init)))
model.add(MaxPooling1D(pool_size=2))
model.add(LSTM(100))
model.add(Dense(1, activation='linear'))
model.compile(loss='mse', optimizer='adam') # , metrics=['accuracy'])
"""
params = {
"filters": 32,
"kernel_size": 3,
"choice_pooling": {
"pooling": True,
"pool_size": 2
},
"neurones": 100,
"batch_size": 50,
"optimizer": "adam",
"activation": "sigmoid",
"nc": nc,
"dropout": 0,
"bi": False
}
ntwk = create_model(params, create_only=True)
if typem == 7:
model = Sequential()
# model.add(Embedding(top_words, embedding_vecor_length, input_length=max_review_length))
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu',
input_shape=(96, init)))
"""
model.add(MaxPooling1D(pool_size=4)) # 16
model.add(Conv1D(filters=64, kernel_size=5, padding='same',
activation='relu'))
model.add(MaxPooling1D(pool_size=4)) #4
model.add(Conv1D(filters=64, kernel_size=5, padding='same',
activation='relu'))
#model.add(LSTM(100))
#model.add(Dense(1, activation='linear'))
"""
model.add(MaxPooling1D(pool_size=4))
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu'))
model.add(MaxPooling1D(pool_size=4))
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu'))
model.add(TimeDistributed(Dense(1, activation='sigmoid')))
model.add(AveragePooling1D(pool_size=6))
model.add(Flatten())
ntwk = model
lenv = 96
ntwk.summary()
return ntwk
def model(typem=1,
window_length=None,
base=False,
idu=False,
activation="linear"):
init = 1
if base:
init = 5
print(init)
if typem in [1, 3]:
if window_length is None:
lenv = 200
else:
lenv = window_length
model = Sequential()
model.add(
Conv1D(filters=32,
kernel_size=3,
padding='same',
activation='relu',
input_shape=(lenv, init)))
model.add(MaxPooling1D(pool_size=2))
model.add(LSTM(100))
model.add(Dense(1, activation=activation))
model.compile(loss='mse', optimizer='adam') # , metrics=['accuracy'])
ntwk = model
if idu:
params = {
"filters": 32,
"kernel_size": 3,
"choice_pooling": {
"pooling": True,
"pool_size": 2
},
"neurones": 100,
"batch_size": 50,
"optimizer": "adam",
"activation": "sigmoid",
"nc": 2,
"dropout": 0,
"bi": False
}
ntwk = create_model(params, create_only=True, typem=args.typem)
lenv = 160
if typem == 7:
model = Sequential()
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu',
input_shape=(96, init)))
"""
model.add(MaxPooling1D(pool_size=4)) # 16
model.add(Conv1D(filters=64, kernel_size=5, padding='same',
activation='relu'))
model.add(MaxPooling1D(pool_size=4)) #4
model.add(Conv1D(filters=64, kernel_size=5, padding='same',
activation='relu'))
# model.add(LSTM(100))
# model.add(Dense(1, activation='linear'))
"""
model.add(MaxPooling1D(pool_size=4))
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu'))
model.add(MaxPooling1D(pool_size=4))
model.add(
Conv1D(filters=32,
kernel_size=5,
padding='same',
activation='relu'))
model.add(TimeDistributed(Dense(1, activation='sigmoid')))
model.add(AveragePooling1D(pool_size=6))
model.add(Flatten())
ntwk = model
lenv = 96
print(ntwk.summary())
return ntwk, lenv