def __init__(self, hyperParams=None):
self.params = hyperParams
hyperParams = {
'input': 16,
'l1_output': 4,
'l2_output': 4,
'output': 16,
}
p = hyperParams
model = Sequential()
model.add(Dense(p['input'], 5, init='uniform'))
model.add(Activation('sigmoid'))
model.add(Dropout(0.5))
# model.add(Dense(24, 4, init='uniform'))
# model.add(Activation('tanh'))
# model.add(Dropout(0.5))
model.add(Dense(5, p["output"]))
model.add(Activation('softmax'))
self.model = model
self.compile()
def __init__(self, hyperParams=None):
self.params = hyperParams
p = self.params["model"]
if(self.checkSavedModel()):
return
model = Sequential()
embed_matrix = self.params["embedding"].getWord2VecMatrix()
emb = Embedding(
embed_matrix.shape[0],
embed_matrix.shape[1],
weights=[embed_matrix],
mask_zero=True,
# learn=(self.params["learn_embedding"] == 1)
)
model.add(emb)
srnn = SimpleRNN(
input_dim=embed_matrix.shape[1],
output_dim=embed_matrix.shape[0],
activation='softmax',
init='uniform',
# inner_activation='hard_sigmoid',
return_sequences=True,
truncate_gradient=int(p.get("depth", -1)),
)
print "Done"
model.add(srnn)
# model.add(Activation('softmax'))
# model.add(LSTM(embed_matrix.shape[1], 128, activation='sigmoid', inner_activation='hard_sigmoid'))
# model.add(Dropout(0.5))
# model.add(Dense(128, 1))
# model.add(Activation('sigmoid'))
self.model = model
# import pdb; pdb.set_trace()
self.compile()
self.saveModel()
def __init__(self, hyperParams=None):
self.params = hyperParams
if(self.checkSavedModel()):
return
p = self.params["model"]
model = Sequential()
embed_matrix = self.params["embedding"].getWord2VecMatrix()
emb = Embedding(
embed_matrix.shape[0],
embed_matrix.shape[1],
weights=[embed_matrix],
mask_zero=True,
# learn=(self.params["learn_embedding"] == 1)
)
model.add(emb)
srnn = SimpleDeepRNN(
input_dim=embed_matrix.shape[1],
output_dim=embed_matrix.shape[0],
activation='softmax',
init='uniform',
inner_init='uniform',
depth = int(p.get("depth", 3)),
inner_activation='sigmoid',
return_sequences=True,
)
print "Done"
model.add(srnn)
self.model = model
self.compile()
self.saveModel()
def __init__(self, hyperParams=None):
self.params = hyperParams
if(self.checkSavedModel()):
return
p = self.params["model"]
model = Sequential()
embed_matrix = self.params["embedding"].getWord2VecMatrix()
emb = Embedding(
embed_matrix.shape[0],
embed_matrix.shape[1],
# weights=[embed_matrix],
mask_zero=True,
learn=(int(self.params["learn_embedding"]) == 1)
)
emb.W.set_value(floatX(embed_matrix))
model.add(emb)
print "Initialized Embeddings"
srnn = SimpleRNN(
input_dim=embed_matrix.shape[1],
output_dim=int(p.get("hidden_nodes", 100)),
activation='tanh',
init='uniform',
inner_init='uniform',
# inner_activation='hard_sigmoid',
return_sequences=False,
truncate_gradient=int(p.get("depth", 3)),
)
model.add(srnn)
print "Initialized Recurrent Layer"
denseL = Dense(
input_dim=int(p.get("hidden_nodes", 100)),
output_dim=int(p.get("output_nodes", 4)),
activation='softmax',
init='uniform',
)
model.add(denseL)
print "Initialized Dense Layer"
self.model = model
self.compile()
self.model.layers[0].params = [self.model.layers[0].W]
self.saveModel()
def __init__(self, embed_matrix):
model = Sequential()
# Add a mask_zero=True to the Embedding connstructor if 0 is a left-padding value in your data
emb = Embedding(embed_matrix.shape[0], embed_matrix.shape[1], weights=[embed_matrix], mask_zero=True)
model.add(emb)
model.add(LSTM(embed_matrix.shape[1], 128, activation='sigmoid', inner_activation='hard_sigmoid'))
model.add(Dropout(0.5))
model.add(Dense(128, 1))
model.add(Activation('sigmoid'))
self.model = model
self.compile()
def __init__(self, hyperParams=None):
# if hyperParams == None :
self.params = hyperParams
hyperParams = {
'l1_input': 100,
'l1_output': 64,
'l2_output': 64,
}
p = hyperParams
model = Sequential()
model.add(Dense(p['l1_input'], 64, init='uniform'))
model.add(Activation('tanh'))
model.add(Dropout(0.5))
model.add(Dense(64, 64, init='uniform'))
model.add(Activation('tanh'))
model.add(Dropout(0.5))
model.add(Dense(64, 1))
model.add(Activation('sigmoid'))
self.model = model
self.compile()
