GRUCell

source

GRUCell(
   input_layer, n_units, previous_state = None, activation = tf.tanh,
   gate_activation = tf.sigmoid, w_init = tf.initializers.glorot_uniform(),
   u_init = tf.initializers.orthogonal(), bias_init = tf.initializers.zeros(),
   u_dropconnect = None, w_dropconnect = None, x_dropout = None, r_dropout = None,
   y_dropout = None, dropout_locked = True, regularized = False,
   share_state_with = None, name = 'gru_cell'
)

---

Gated Recurrent Unit Cell.

Performs a single step with a gated recurrent unit where. These units have two gates: The first defines how much do we use the values from the recurrent connection to predict the current state The second

Methods:

.init_state

source

.init_state()

.compute

source

.compute(
   input_layer, *previous_state
)

.reuse_with

source

.reuse_with(
   input_layer, *previous_state, regularized = None, name = None
)