python 3.5 - how to frame model for deep learning with vectors as input? -
how can create deep learning model multiclass classification, if feature set contain 5 vectors of length 8 each? feature set looks features=[[0.34, 0.67, 0.87, .. ][0.67, 0.87, 0.88 ....][][][],....]
model = sequential() model.add(dense(8, input_dim=40, activation='relu')) model.add(dense(33, activation='softmax')) model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
just flatten data tf.contrib.layers.flatten
, , have 40 features in 1d, that's it.
however if there specific 2d arrangement might consider convolutions exploit fact.
Comments
Post a Comment