tfr.keras.losses.MeanSquaredLoss

View source on GitHub

Computes mean squared loss between y_true and y_pred.

tfr.keras.losses.MeanSquaredLoss(
    reduction=tf.losses.Reduction.AUTO, name=None, ragged=False
)

loss = (y_true - y_pred)^2

Standalone usage:

y_true = [[1., 0.]]
y_pred = [[0.6, 0.8]]
loss = tfr.keras.losses.MeanSquaredLoss()
loss(y_true, y_pred).numpy()
0.4

# Using ragged tensors
y_true = tf.ragged.constant([[1., 0.], [0., 1., 0.]])
y_pred = tf.ragged.constant([[0.6, 0.8], [0.5, 0.8, 0.4]])
loss = tfr.keras.losses.MeanSquaredLoss(ragged=True)
loss(y_true, y_pred).numpy()
0.20833336

Usage with the compile() API:

model.compile(optimizer='sgd', loss=tfr.keras.losses.MeanSquaredLoss())

Definition:

\[ \mathcal{L}(\{y\}, \{s\}) = \sum_i (y_i - s_i)^{2} \]

Args
reduction	(Optional) The tf.keras.losses.Reduction to use (see tf.keras.losses.Loss).
name	(Optional) The name for the op.
ragged	(Optional) If True, this loss will accept ragged tensors. If False, this loss will accept dense tensors.

Methods

from_config

@classmethod
from_config(
    config
)

Instantiates a Loss from its config (output of get_config()).

Args
config	Output of get_config().

Returns
A Loss instance.

get_config

View source

get_config()

Returns the config dictionary for a Loss instance.

__call__

View source

__call__(
    y_true, y_pred, sample_weight=None
)

See tf.keras.losses.Loss.