tf.compat.v1.nn.relu_layer

View source on GitHub

Computes Relu(x * weight + biases).

tf.compat.v1.nn.relu_layer(
    x, weights, biases, name=None
)

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.