Watch talks from the 2019 TensorFlow Dev Summit Watch now

tf.nn.relu_layer

tf.nn.relu_layer(
    x,
    weights,
    biases,
    name=None
)

Defined in tensorflow/python/ops/nn_impl.py.

Computes Relu(x * weight + biases).

Args:

  • x: a 2D tensor. Dimensions typically: batch, in_units
  • weights: a 2D tensor. Dimensions typically: in_units, out_units
  • biases: a 1D tensor. Dimensions: out_units
  • name: A name for the operation (optional). If not specified "nn_relu_layer" is used.

Returns:

A 2-D Tensor computing relu(matmul(x, weights) + biases). Dimensions typically: batch, out_units.