tf.nn.l2_loss

tf.nn.l2_loss(
    t,
    name=None
)

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

See the guide: Neural Network > Losses

L2 Loss.

Computes half the L2 norm of a tensor without the sqrt:

output = sum(t ** 2) / 2

Args:

• t: A Tensor. Must be one of the following types: half, bfloat16, float32, float64. Typically 2-D, but may have any dimensions.
• name: A name for the operation (optional).

Returns:

A Tensor. Has the same type as t.