Please enable JavaScript.
Coggle requires JavaScript to display documents.
Keras: Sequential Models Layers - Coggle Diagram
Keras: Sequential Models Layers
Introduction
what is
sequential model
classic layer to layer ANN
1 layer has 1 tensor input and 1 tensor output
why is it important
introductory ANN in tensorflow
choose if:
good if model has single input and single output
single layers input and output
no layer sharing
linear toplology
Dense
what is it
very basic ANN perceptron/neuron layer
neurons implements the
base
equation
when to use
dense
every
sequential
training, at least have 1 layer of Dense.
Flatten
what is it
a
reshaping layer
transform tensor to 1d tensor
similar to numpy.ravel
the output is x * y tensor if from 2D
when to use it
supplying data to
Dense
Dropout
what is it
regularization layer
randomly set input units to 0
when to use it
when you want to avoid overfitting
COMING SOON Activation