trisgelar / belajar-keras

Belajar Keras dari berbagai sumber

ReadMe

Belajar Keras dari Berbagai Sumber

1. Keras Succinctly link

Snippets

1. Import Keras

	from keras.models import Sequential
	from keras.layers import Dense, Activation

Catatan

1. Common keras.initializers Functions

Function	Description
Zeros()	All np.float32 0.0 values
Constant(value=0)	All a single specified np.float32 value
RandomUniform(minval=-0.05, maxval=0.05, seed=None)	Random, evenly distributed between minval and maxval
glorot_normal(seed=None)	Truncated Normal with stddev = sqrt(2 / (fan_in + fan_out))
glorot_uniform(seed=None)	Uniform random with limits sqrt(6 / (fan_in + fan_out))

2. Common Dense Layer Activation Functions

Function	Description
relu(x, alpha=0.0, max_value=None)	if x < 0 , f(x) = 0, else f(x) = x
tanh(x)	hyperbolic tangent
sigmoid(x)	f(x) = 1.0 / (1.0 + exp(-x))
linear(x)	f(x) = x
softmax(x, axis=-1)	coerces vector x values to sum to 1.0 so they can be loosely interpreted as probabilities

3. Accuracy Metrics Functions

Function	Description
binary_accuracy(y_true, y_pred)	For binary classification
categorical_accuracy(y_true, y_pred)	For multiclass classification
sparse_categorical_accuracy(y_true, y_pred)	Rarely used (see documentation))
top_k_categorical_accuracy(y_true, y_pred, k=5)	Rarely used (see documentation))
sparse_top_k_categorical_accuracy(y_true, y_pred, k=5)	Rarely used (see documentation))

4. Five Common Keras Optimizers

Optimizer	Description
SGD(lr=0.01, momentum=0.0, decay=0.0, nesterov=False)	Basic optimizer for simple neural networks
RMSprop(lr=0.001, rho=0.9, epsilon=None, decay=0.0)	Often used with recurrent neural networks, very similar to Adadelta
Adagrad(lr=0.01, epsilon=None, decay=0.0)	General purpose adaptive algorithm
Adadelta(lr=1.0, rho=0.95, epsilon=None, decay=0.0)	Advanced version of Adagrad, similar to RMSprop
Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0, amsgrad=False)	Excellent general-purpose, adaptive algorithm

5. Embedding Layer Parameters

Name	Description
input_dim	Size of the vocabulary, i.e. maximum integer index + 1
output_dim	Dimension of the dense embedding
embeddings_initializer	Initializer for the embeddings matrix
embeddings_regularizer	Regularizer function applied to the embeddings matrix
embeddings_constraint	Constraint function applied to the embeddings matrix
mask_zero	Whether or not the input value 0 is a padding value
input_length	Length of input sequences, when it is constant