In [2]:

    

from keras.models import Sequential


    

In [34]:

    

import numpy as np
X_train = np.random.rand(1000,100)
Y_train = [np.random.randint(0,5) for i in xrange(0,1000)]


    

In [35]:

    

model = Sequential()


    

In [36]:

    

from keras.layers.core import Dense,Activation

model.add(Dense(output_dim=64, input_dim=100, init="glorot_uniform"))
model.add(Activation("relu"))
model.add(Dense(output_dim=10, init="glorot_uniform"))
model.add(Activation("softmax"))


    

In [37]:

    

model.compile(loss="categorical_crossentropy", optimizer="sgd")


    

In [38]:

    

from keras.optimizers import SGD
model.compile(loss="categorical_crossentropy", optimizer=SGD(lr=0.01, momentum=0.9, nesterov=True))


    

In [39]:

    

model.fit(X_train, Y_train, nb_epoch=5, batch_size=32)


    

    




Epoch 1/5
1000/1000 [==============================] - 0s - loss: 46.9767     
Epoch 2/5
1000/1000 [==============================] - 0s - loss: 46.4905     
Epoch 3/5
1000/1000 [==============================] - 0s - loss: 46.4733     
Epoch 4/5
1000/1000 [==============================] - 0s - loss: 46.4694     
Epoch 5/5
1000/1000 [==============================] - 0s - loss: 46.4680     






    
Out[39]:





<keras.callbacks.History at 0x7fa2cc208950>

In [41]:

    

objective_score = model.evaluate(X_train, Y_train, batch_size=32)


    

    




1000/1000 [==============================] - 0s     

In [43]:

    

objective_score


    

    
Out[43]:





46.467602020263669

In [44]:

    

classes = model.predict_classes(X_train, batch_size=32)


    

    




1000/1000 [==============================] - 0s     

In [46]:

    

proba = model.predict_proba(X_train, batch_size=32)


    

    




1000/1000 [==============================] - 0s     

In [ ]: