![]() |
Class Huber
Computes the Huber loss between y_true
and y_pred
.
Aliases:
- Class
tf.compat.v1.keras.losses.Huber
- Class
tf.compat.v2.keras.losses.Huber
- Class
tf.compat.v2.losses.Huber
For each value x in error = y_true - y_pred
:
loss = 0.5 * x^2 if |x| <= d
loss = 0.5 * d^2 + d * (|x| - d) if |x| > d
where d is delta
. See: https://en.wikipedia.org/wiki/Huber_loss
Usage:
l = tf.keras.losses.Huber()
loss = l([0., 1., 1.], [1., 0., 1.])
print('Loss: ', loss.numpy()) # Loss: 0.333
Usage with the compile
API:
model = tf.keras.Model(inputs, outputs)
model.compile('sgd', loss=tf.keras.losses.Huber())
Args:
delta
: A float, the point where the Huber loss function changes from a quadratic to linear.reduction
: (Optional) Type oftf.keras.losses.Reduction
to apply to loss. Default value isAUTO
.AUTO
indicates that the reduction option will be determined by the usage context. For almost all cases this defaults toSUM_OVER_BATCH_SIZE
. When used withtf.distribute.Strategy
, outside of built-in training loops such astf.keras
compile
andfit
, usingAUTO
orSUM_OVER_BATCH_SIZE
will raise an error. Please see https://www.tensorflow.org/alpha/tutorials/distribute/training_loops for more details on this.name
: Optional name for the op.
__init__
__init__(
delta=1.0,
reduction=losses_utils.ReductionV2.AUTO,
name='huber_loss'
)
Initialize self. See help(type(self)) for accurate signature.
Methods
tf.keras.losses.Huber.__call__
__call__(
y_true,
y_pred,
sample_weight=None
)
Invokes the Loss
instance.
Args:
y_true
: Ground truth values. shape =[batch_size, d0, .. dN]
y_pred
: The predicted values. shape =[batch_size, d0, .. dN]
sample_weight
: Optionalsample_weight
acts as a coefficient for the loss. If a scalar is provided, then the loss is simply scaled by the given value. Ifsample_weight
is a tensor of size[batch_size]
, then the total loss for each sample of the batch is rescaled by the corresponding element in thesample_weight
vector. If the shape ofsample_weight
is[batch_size, d0, .. dN-1]
(or can be broadcasted to this shape), then each loss element ofy_pred
is scaled by the corresponding value ofsample_weight
. (Note ondN-1
: all loss functions reduce by 1 dimension, usually axis=-1.)
Returns:
Weighted loss float Tensor
. If reduction
is NONE
, this has
shape [batch_size, d0, .. dN-1]
; otherwise, it is scalar. (Note dN-1
because all loss functions reduce by 1 dimension, usually axis=-1.)
Raises:
ValueError
: If the shape ofsample_weight
is invalid.
tf.keras.losses.Huber.from_config
from_config(
cls,
config
)
Instantiates a Loss
from its config (output of get_config()
).
Args:
config
: Output ofget_config()
.
Returns:
A Loss
instance.
tf.keras.losses.Huber.get_config
get_config()