tensorflow/python/keras/losses.py
# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
# pylint: disable=g-classes-have-attributes
"""Built-in loss functions."""
import abc
import functools
from tensorflow.python.autograph.core import ag_ctx
from tensorflow.python.autograph.impl import api as autograph
from tensorflow.python.distribute import distribute_lib
from tensorflow.python.eager import context
from tensorflow.python.framework import constant_op
from tensorflow.python.framework import ops
from tensorflow.python.framework import smart_cond
from tensorflow.python.framework import tensor_conversion
from tensorflow.python.framework import tensor_spec
from tensorflow.python.framework import tensor_util
from tensorflow.python.keras import backend
from tensorflow.python.keras.utils import losses_utils
from tensorflow.python.keras.utils import tf_utils
from tensorflow.python.keras.utils.generic_utils import deserialize_keras_object
from tensorflow.python.keras.utils.generic_utils import serialize_keras_object
from tensorflow.python.ops import array_ops
from tensorflow.python.ops import cond
from tensorflow.python.ops import math_ops
from tensorflow.python.ops import nn
from tensorflow.python.ops.losses import losses_impl
from tensorflow.python.ops.ragged import ragged_map_ops
from tensorflow.python.ops.ragged import ragged_tensor
from tensorflow.python.ops.ragged import ragged_util
from tensorflow.python.util import dispatch
from tensorflow.tools.docs import doc_controls
class Loss:
"""Loss base class.
To be implemented by subclasses:
* `call()`: Contains the logic for loss calculation using `y_true`, `y_pred`.
Example subclass implementation:
```python
class MeanSquaredError(Loss):
def call(self, y_true, y_pred):
y_pred = tf.convert_to_tensor_v2(y_pred)
y_true = tf.cast(y_true, y_pred.dtype)
return tf.reduce_mean(math_ops.square(y_pred - y_true), axis=-1)
```
When used with `tf.distribute.Strategy`, outside of built-in training loops
such as `tf.keras` `compile` and `fit`, please use 'SUM' or 'NONE' reduction
types, and reduce losses explicitly in your training loop. Using 'AUTO' or
'SUM_OVER_BATCH_SIZE' will raise an error.
Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for more
details on this.
You can implement 'SUM_OVER_BATCH_SIZE' using global batch size like:
```python
with strategy.scope():
loss_obj = tf.keras.losses.CategoricalCrossentropy(
reduction=tf.keras.losses.Reduction.NONE)
....
loss = (tf.reduce_sum(loss_obj(labels, predictions)) *
(1. / global_batch_size))
```
"""
def __init__(self, reduction=losses_utils.ReductionV2.AUTO, name=None):
"""Initializes `Loss` class.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance.
"""
losses_utils.ReductionV2.validate(reduction)
self.reduction = reduction
self.name = name
# SUM_OVER_BATCH is only allowed in losses managed by `fit` or
# CannedEstimators.
self._allow_sum_over_batch_size = False
self._set_name_scope()
def _set_name_scope(self):
"""Creates a valid `name_scope` name."""
if self.name is None:
self._name_scope = self.__class__.__name__
elif self.name == '<lambda>':
self._name_scope = 'lambda'
else:
# E.g. '_my_loss' => 'my_loss'
self._name_scope = self.name.strip('_')
def __call__(self, y_true, y_pred, sample_weight=None):
"""Invokes the `Loss` instance.
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`, except
sparse loss functions such as sparse categorical crossentropy where
shape = `[batch_size, d0, .. dN-1]`
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`
sample_weight: Optional `sample_weight` acts as a coefficient for the
loss. If a scalar is provided, then the loss is simply scaled by the
given value. If `sample_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 the `sample_weight` vector. If the shape of
`sample_weight` is `[batch_size, d0, .. dN-1]` (or can be broadcasted to
this shape), then each loss element of `y_pred` is scaled
by the corresponding value of `sample_weight`. (Note on`dN-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 of `sample_weight` is invalid.
"""
# If we are wrapping a lambda function strip '<>' from the name as it is not
# accepted in scope name.
graph_ctx = tf_utils.graph_context_for_symbolic_tensors(
y_true, y_pred, sample_weight)
with backend.name_scope(self._name_scope), graph_ctx:
if context.executing_eagerly():
call_fn = self.call
else:
call_fn = autograph.tf_convert(self.call, ag_ctx.control_status_ctx())
losses = call_fn(y_true, y_pred)
return losses_utils.compute_weighted_loss(
losses, sample_weight, reduction=self._get_reduction())
@classmethod
def from_config(cls, config):
"""Instantiates a `Loss` from its config (output of `get_config()`).
Args:
config: Output of `get_config()`.
Returns:
A `Loss` instance.
"""
return cls(**config)
def get_config(self):
"""Returns the config dictionary for a `Loss` instance."""
return {'reduction': self.reduction, 'name': self.name}
@abc.abstractmethod
@doc_controls.for_subclass_implementers
def call(self, y_true, y_pred):
"""Invokes the `Loss` instance.
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`, except
sparse loss functions such as sparse categorical crossentropy where
shape = `[batch_size, d0, .. dN-1]`
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`
Returns:
Loss values with the shape `[batch_size, d0, .. dN-1]`.
"""
raise NotImplementedError('Must be implemented in subclasses.')
def _get_reduction(self):
"""Handles `AUTO` reduction cases and returns the reduction value."""
if (not self._allow_sum_over_batch_size and
distribute_lib.has_strategy() and
(self.reduction == losses_utils.ReductionV2.AUTO or
self.reduction == losses_utils.ReductionV2.SUM_OVER_BATCH_SIZE)):
raise ValueError(
'Please use `tf.keras.losses.Reduction.SUM` or '
'`tf.keras.losses.Reduction.NONE` for loss reduction when losses are '
'used with `tf.distribute.Strategy` outside of the built-in training '
'loops. You can implement '
'`tf.keras.losses.Reduction.SUM_OVER_BATCH_SIZE` using global batch '
'size like:\n```\nwith strategy.scope():\n'
' loss_obj = tf.keras.losses.CategoricalCrossentropy('
'reduction=tf.keras.losses.Reduction.NONE)\n....\n'
' loss = tf.reduce_sum(loss_obj(labels, predictions)) * '
'(1. / global_batch_size)\n```\nPlease see '
'https://www.tensorflow.org/tutorials/distribute/custom_training'
' for more details.')
if self.reduction == losses_utils.ReductionV2.AUTO:
return losses_utils.ReductionV2.SUM_OVER_BATCH_SIZE
return self.reduction
class LossFunctionWrapper(Loss):
"""Wraps a loss function in the `Loss` class."""
def __init__(self,
fn,
reduction=losses_utils.ReductionV2.AUTO,
name=None,
**kwargs):
"""Initializes `LossFunctionWrapper` class.
Args:
fn: The loss function to wrap, with signature `fn(y_true, y_pred,
**kwargs)`.
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance.
**kwargs: The keyword arguments that are passed on to `fn`.
"""
super().__init__(reduction=reduction, name=name)
self.fn = fn
self._fn_kwargs = kwargs
def call(self, y_true, y_pred):
"""Invokes the `LossFunctionWrapper` instance.
Args:
y_true: Ground truth values.
y_pred: The predicted values.
Returns:
Loss values per sample.
"""
if tensor_util.is_tf_type(y_pred) and tensor_util.is_tf_type(y_true):
y_pred, y_true = losses_utils.squeeze_or_expand_dimensions(y_pred, y_true)
ag_fn = autograph.tf_convert(self.fn, ag_ctx.control_status_ctx())
return ag_fn(y_true, y_pred, **self._fn_kwargs)
def get_config(self):
config = {}
for k, v in self._fn_kwargs.items():
config[k] = backend.eval(v) if tf_utils.is_tensor_or_variable(v) else v
base_config = super().get_config()
return dict(list(base_config.items()) + list(config.items()))
class MeanSquaredError(LossFunctionWrapper):
"""Computes the mean of squares of errors between labels and predictions.
`loss = square(y_true - y_pred)`
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[1., 1.], [1., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> mse = tf.keras.losses.MeanSquaredError()
>>> mse(y_true, y_pred).numpy()
0.5
>>> # Calling with 'sample_weight'.
>>> mse(y_true, y_pred, sample_weight=[0.7, 0.3]).numpy()
0.25
>>> # Using 'sum' reduction type.
>>> mse = tf.keras.losses.MeanSquaredError(
... reduction=tf.keras.losses.Reduction.SUM)
>>> mse(y_true, y_pred).numpy()
1.0
>>> # Using 'none' reduction type.
>>> mse = tf.keras.losses.MeanSquaredError(
... reduction=tf.keras.losses.Reduction.NONE)
>>> mse(y_true, y_pred).numpy()
array([0.5, 0.5], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.MeanSquaredError())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='mean_squared_error'):
"""Initializes `MeanSquaredError` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'mean_squared_error'.
"""
super().__init__(mean_squared_error, name=name, reduction=reduction)
class MeanAbsoluteError(LossFunctionWrapper):
"""Computes the mean of absolute difference between labels and predictions.
`loss = abs(y_true - y_pred)`
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[1., 1.], [1., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> mae = tf.keras.losses.MeanAbsoluteError()
>>> mae(y_true, y_pred).numpy()
0.5
>>> # Calling with 'sample_weight'.
>>> mae(y_true, y_pred, sample_weight=[0.7, 0.3]).numpy()
0.25
>>> # Using 'sum' reduction type.
>>> mae = tf.keras.losses.MeanAbsoluteError(
... reduction=tf.keras.losses.Reduction.SUM)
>>> mae(y_true, y_pred).numpy()
1.0
>>> # Using 'none' reduction type.
>>> mae = tf.keras.losses.MeanAbsoluteError(
... reduction=tf.keras.losses.Reduction.NONE)
>>> mae(y_true, y_pred).numpy()
array([0.5, 0.5], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.MeanAbsoluteError())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='mean_absolute_error'):
"""Initializes `MeanAbsoluteError` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'mean_absolute_error'.
"""
super().__init__(mean_absolute_error, name=name, reduction=reduction)
class MeanAbsolutePercentageError(LossFunctionWrapper):
"""Computes the mean absolute percentage error between `y_true` and `y_pred`.
`loss = 100 * abs(y_true - y_pred) / y_true`
Standalone usage:
>>> y_true = [[2., 1.], [2., 3.]]
>>> y_pred = [[1., 1.], [1., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> mape = tf.keras.losses.MeanAbsolutePercentageError()
>>> mape(y_true, y_pred).numpy()
50.
>>> # Calling with 'sample_weight'.
>>> mape(y_true, y_pred, sample_weight=[0.7, 0.3]).numpy()
20.
>>> # Using 'sum' reduction type.
>>> mape = tf.keras.losses.MeanAbsolutePercentageError(
... reduction=tf.keras.losses.Reduction.SUM)
>>> mape(y_true, y_pred).numpy()
100.
>>> # Using 'none' reduction type.
>>> mape = tf.keras.losses.MeanAbsolutePercentageError(
... reduction=tf.keras.losses.Reduction.NONE)
>>> mape(y_true, y_pred).numpy()
array([25., 75.], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd',
loss=tf.keras.losses.MeanAbsolutePercentageError())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='mean_absolute_percentage_error'):
"""Initializes `MeanAbsolutePercentageError` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to
'mean_absolute_percentage_error'.
"""
super().__init__(
mean_absolute_percentage_error, name=name, reduction=reduction)
class MeanSquaredLogarithmicError(LossFunctionWrapper):
"""Computes the mean squared logarithmic error between `y_true` and `y_pred`.
`loss = square(log(y_true + 1.) - log(y_pred + 1.))`
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[1., 1.], [1., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> msle = tf.keras.losses.MeanSquaredLogarithmicError()
>>> msle(y_true, y_pred).numpy()
0.240
>>> # Calling with 'sample_weight'.
>>> msle(y_true, y_pred, sample_weight=[0.7, 0.3]).numpy()
0.120
>>> # Using 'sum' reduction type.
>>> msle = tf.keras.losses.MeanSquaredLogarithmicError(
... reduction=tf.keras.losses.Reduction.SUM)
>>> msle(y_true, y_pred).numpy()
0.480
>>> # Using 'none' reduction type.
>>> msle = tf.keras.losses.MeanSquaredLogarithmicError(
... reduction=tf.keras.losses.Reduction.NONE)
>>> msle(y_true, y_pred).numpy()
array([0.240, 0.240], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd',
loss=tf.keras.losses.MeanSquaredLogarithmicError())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='mean_squared_logarithmic_error'):
"""Initializes `MeanSquaredLogarithmicError` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to
'mean_squared_logarithmic_error'.
"""
super().__init__(
mean_squared_logarithmic_error, name=name, reduction=reduction)
class BinaryCrossentropy(LossFunctionWrapper):
"""Computes the cross-entropy loss between true labels and predicted labels.
Use this cross-entropy loss for binary (0 or 1) classification applications.
The loss function requires the following inputs:
- `y_true` (true label): This is either 0 or 1.
- `y_pred` (predicted value): This is the model's prediction, i.e, a single
floating-point value which either represents a
[logit](https://en.wikipedia.org/wiki/Logit), (i.e, value in [-inf, inf]
when `from_logits=True`) or a probability (i.e, value in [0., 1.] when
`from_logits=False`).
**Recommended Usage:** (set `from_logits=True`)
With `tf.keras` API:
```python
model.compile(
loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
....
)
```
As a standalone function:
>>> # Example 1: (batch_size = 1, number of samples = 4)
>>> y_true = [0, 1, 0, 0]
>>> y_pred = [-18.6, 0.51, 2.94, -12.8]
>>> bce = tf.keras.losses.BinaryCrossentropy(from_logits=True)
>>> bce(y_true, y_pred).numpy()
0.865
>>> # Example 2: (batch_size = 2, number of samples = 4)
>>> y_true = [[0, 1], [0, 0]]
>>> y_pred = [[-18.6, 0.51], [2.94, -12.8]]
>>> # Using default 'auto'/'sum_over_batch_size' reduction type.
>>> bce = tf.keras.losses.BinaryCrossentropy(from_logits=True)
>>> bce(y_true, y_pred).numpy()
0.865
>>> # Using 'sample_weight' attribute
>>> bce(y_true, y_pred, sample_weight=[0.8, 0.2]).numpy()
0.243
>>> # Using 'sum' reduction` type.
>>> bce = tf.keras.losses.BinaryCrossentropy(from_logits=True,
... reduction=tf.keras.losses.Reduction.SUM)
>>> bce(y_true, y_pred).numpy()
1.730
>>> # Using 'none' reduction type.
>>> bce = tf.keras.losses.BinaryCrossentropy(from_logits=True,
... reduction=tf.keras.losses.Reduction.NONE)
>>> bce(y_true, y_pred).numpy()
array([0.235, 1.496], dtype=float32)
**Default Usage:** (set `from_logits=False`)
>>> # Make the following updates to the above "Recommended Usage" section
>>> # 1. Set `from_logits=False`
>>> tf.keras.losses.BinaryCrossentropy() # OR ...('from_logits=False')
>>> # 2. Update `y_pred` to use probabilities instead of logits
>>> y_pred = [0.6, 0.3, 0.2, 0.8] # OR [[0.6, 0.3], [0.2, 0.8]]
"""
def __init__(self,
from_logits=False,
label_smoothing=0,
axis=-1,
reduction=losses_utils.ReductionV2.AUTO,
name='binary_crossentropy'):
"""Initializes `BinaryCrossentropy` instance.
Args:
from_logits: Whether to interpret `y_pred` as a tensor of
[logit](https://en.wikipedia.org/wiki/Logit) values. By default, we
assume that `y_pred` contains probabilities (i.e., values in [0, 1]).
label_smoothing: Float in [0, 1]. When 0, no smoothing occurs. When > 0,
we compute the loss between the predicted labels and a smoothed version
of the true labels, where the smoothing squeezes the labels towards 0.5.
Larger values of `label_smoothing` correspond to heavier smoothing.
axis: The axis along which to compute crossentropy (the features axis).
Defaults to -1.
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Name for the op. Defaults to 'binary_crossentropy'.
"""
super().__init__(
binary_crossentropy,
name=name,
reduction=reduction,
from_logits=from_logits,
label_smoothing=label_smoothing,
axis=axis)
self.from_logits = from_logits
class CategoricalCrossentropy(LossFunctionWrapper):
"""Computes the crossentropy loss between the labels and predictions.
Use this crossentropy loss function when there are two or more label classes.
We expect labels to be provided in a `one_hot` representation. If you want to
provide labels as integers, please use `SparseCategoricalCrossentropy` loss.
There should be `# classes` floating point values per feature.
In the snippet below, there is `# classes` floating pointing values per
example. The shape of both `y_pred` and `y_true` are
`[batch_size, num_classes]`.
Standalone usage:
>>> y_true = [[0, 1, 0], [0, 0, 1]]
>>> y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> cce = tf.keras.losses.CategoricalCrossentropy()
>>> cce(y_true, y_pred).numpy()
1.177
>>> # Calling with 'sample_weight'.
>>> cce(y_true, y_pred, sample_weight=tf.constant([0.3, 0.7])).numpy()
0.814
>>> # Using 'sum' reduction type.
>>> cce = tf.keras.losses.CategoricalCrossentropy(
... reduction=tf.keras.losses.Reduction.SUM)
>>> cce(y_true, y_pred).numpy()
2.354
>>> # Using 'none' reduction type.
>>> cce = tf.keras.losses.CategoricalCrossentropy(
... reduction=tf.keras.losses.Reduction.NONE)
>>> cce(y_true, y_pred).numpy()
array([0.0513, 2.303], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.CategoricalCrossentropy())
```
"""
def __init__(self,
from_logits=False,
label_smoothing=0,
axis=-1,
reduction=losses_utils.ReductionV2.AUTO,
name='categorical_crossentropy'):
"""Initializes `CategoricalCrossentropy` instance.
Args:
from_logits: Whether `y_pred` is expected to be a logits tensor. By
default, we assume that `y_pred` encodes a probability distribution.
label_smoothing: Float in [0, 1]. When > 0, label values are smoothed,
meaning the confidence on label values are relaxed. For example, if
`0.1`, use `0.1 / num_classes` for non-target labels and
`0.9 + 0.1 / num_classes` for target labels.
axis: The axis along which to compute crossentropy (the features axis).
Defaults to -1.
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance.
Defaults to 'categorical_crossentropy'.
"""
super().__init__(
categorical_crossentropy,
name=name,
reduction=reduction,
from_logits=from_logits,
label_smoothing=label_smoothing,
axis=axis)
class SparseCategoricalCrossentropy(LossFunctionWrapper):
"""Computes the crossentropy loss between the labels and predictions.
Use this crossentropy loss function when there are two or more label classes.
We expect labels to be provided as integers. If you want to provide labels
using `one-hot` representation, please use `CategoricalCrossentropy` loss.
There should be `# classes` floating point values per feature for `y_pred`
and a single floating point value per feature for `y_true`.
In the snippet below, there is a single floating point value per example for
`y_true` and `# classes` floating pointing values per example for `y_pred`.
The shape of `y_true` is `[batch_size]` and the shape of `y_pred` is
`[batch_size, num_classes]`.
Standalone usage:
>>> y_true = [1, 2]
>>> y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> scce = tf.keras.losses.SparseCategoricalCrossentropy()
>>> scce(y_true, y_pred).numpy()
1.177
>>> # Calling with 'sample_weight'.
>>> scce(y_true, y_pred, sample_weight=tf.constant([0.3, 0.7])).numpy()
0.814
>>> # Using 'sum' reduction type.
>>> scce = tf.keras.losses.SparseCategoricalCrossentropy(
... reduction=tf.keras.losses.Reduction.SUM)
>>> scce(y_true, y_pred).numpy()
2.354
>>> # Using 'none' reduction type.
>>> scce = tf.keras.losses.SparseCategoricalCrossentropy(
... reduction=tf.keras.losses.Reduction.NONE)
>>> scce(y_true, y_pred).numpy()
array([0.0513, 2.303], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd',
loss=tf.keras.losses.SparseCategoricalCrossentropy())
```
"""
def __init__(self,
from_logits=False,
reduction=losses_utils.ReductionV2.AUTO,
name='sparse_categorical_crossentropy'):
"""Initializes `SparseCategoricalCrossentropy` instance.
Args:
from_logits: Whether `y_pred` is expected to be a logits tensor. By
default, we assume that `y_pred` encodes a probability distribution.
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to
'sparse_categorical_crossentropy'.
"""
super().__init__(
sparse_categorical_crossentropy,
name=name,
reduction=reduction,
from_logits=from_logits)
class Hinge(LossFunctionWrapper):
"""Computes the hinge loss between `y_true` and `y_pred`.
`loss = maximum(1 - y_true * y_pred, 0)`
`y_true` values are expected to be -1 or 1. If binary (0 or 1) labels are
provided we will convert them to -1 or 1.
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> h = tf.keras.losses.Hinge()
>>> h(y_true, y_pred).numpy()
1.3
>>> # Calling with 'sample_weight'.
>>> h(y_true, y_pred, sample_weight=[1, 0]).numpy()
0.55
>>> # Using 'sum' reduction type.
>>> h = tf.keras.losses.Hinge(
... reduction=tf.keras.losses.Reduction.SUM)
>>> h(y_true, y_pred).numpy()
2.6
>>> # Using 'none' reduction type.
>>> h = tf.keras.losses.Hinge(
... reduction=tf.keras.losses.Reduction.NONE)
>>> h(y_true, y_pred).numpy()
array([1.1, 1.5], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.Hinge())
```
"""
def __init__(self, reduction=losses_utils.ReductionV2.AUTO, name='hinge'):
"""Initializes `Hinge` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'hinge'.
"""
super().__init__(hinge, name=name, reduction=reduction)
class SquaredHinge(LossFunctionWrapper):
"""Computes the squared hinge loss between `y_true` and `y_pred`.
`loss = square(maximum(1 - y_true * y_pred, 0))`
`y_true` values are expected to be -1 or 1. If binary (0 or 1) labels are
provided we will convert them to -1 or 1.
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> h = tf.keras.losses.SquaredHinge()
>>> h(y_true, y_pred).numpy()
1.86
>>> # Calling with 'sample_weight'.
>>> h(y_true, y_pred, sample_weight=[1, 0]).numpy()
0.73
>>> # Using 'sum' reduction type.
>>> h = tf.keras.losses.SquaredHinge(
... reduction=tf.keras.losses.Reduction.SUM)
>>> h(y_true, y_pred).numpy()
3.72
>>> # Using 'none' reduction type.
>>> h = tf.keras.losses.SquaredHinge(
... reduction=tf.keras.losses.Reduction.NONE)
>>> h(y_true, y_pred).numpy()
array([1.46, 2.26], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.SquaredHinge())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='squared_hinge'):
"""Initializes `SquaredHinge` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'squared_hinge'.
"""
super().__init__(squared_hinge, name=name, reduction=reduction)
class CategoricalHinge(LossFunctionWrapper):
"""Computes the categorical hinge loss between `y_true` and `y_pred`.
`loss = maximum(neg - pos + 1, 0)`
where `neg=maximum((1-y_true)*y_pred) and pos=sum(y_true*y_pred)`
Standalone usage:
>>> y_true = [[0, 1], [0, 0]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> h = tf.keras.losses.CategoricalHinge()
>>> h(y_true, y_pred).numpy()
1.4
>>> # Calling with 'sample_weight'.
>>> h(y_true, y_pred, sample_weight=[1, 0]).numpy()
0.6
>>> # Using 'sum' reduction type.
>>> h = tf.keras.losses.CategoricalHinge(
... reduction=tf.keras.losses.Reduction.SUM)
>>> h(y_true, y_pred).numpy()
2.8
>>> # Using 'none' reduction type.
>>> h = tf.keras.losses.CategoricalHinge(
... reduction=tf.keras.losses.Reduction.NONE)
>>> h(y_true, y_pred).numpy()
array([1.2, 1.6], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.CategoricalHinge())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='categorical_hinge'):
"""Initializes `CategoricalHinge` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'categorical_hinge'.
"""
super().__init__(categorical_hinge, name=name, reduction=reduction)
class Poisson(LossFunctionWrapper):
"""Computes the Poisson loss between `y_true` and `y_pred`.
`loss = y_pred - y_true * log(y_pred)`
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[1., 1.], [0., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> p = tf.keras.losses.Poisson()
>>> p(y_true, y_pred).numpy()
0.5
>>> # Calling with 'sample_weight'.
>>> p(y_true, y_pred, sample_weight=[0.8, 0.2]).numpy()
0.4
>>> # Using 'sum' reduction type.
>>> p = tf.keras.losses.Poisson(
... reduction=tf.keras.losses.Reduction.SUM)
>>> p(y_true, y_pred).numpy()
0.999
>>> # Using 'none' reduction type.
>>> p = tf.keras.losses.Poisson(
... reduction=tf.keras.losses.Reduction.NONE)
>>> p(y_true, y_pred).numpy()
array([0.999, 0.], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.Poisson())
```
"""
def __init__(self, reduction=losses_utils.ReductionV2.AUTO, name='poisson'):
"""Initializes `Poisson` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'poisson'.
"""
super().__init__(poisson, name=name, reduction=reduction)
class LogCosh(LossFunctionWrapper):
"""Computes the logarithm of the hyperbolic cosine of the prediction error.
`logcosh = log((exp(x) + exp(-x))/2)`,
where x is the error `y_pred - y_true`.
Standalone usage:
>>> y_true = [[0., 1.], [0., 0.]]
>>> y_pred = [[1., 1.], [0., 0.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> l = tf.keras.losses.LogCosh()
>>> l(y_true, y_pred).numpy()
0.108
>>> # Calling with 'sample_weight'.
>>> l(y_true, y_pred, sample_weight=[0.8, 0.2]).numpy()
0.087
>>> # Using 'sum' reduction type.
>>> l = tf.keras.losses.LogCosh(
... reduction=tf.keras.losses.Reduction.SUM)
>>> l(y_true, y_pred).numpy()
0.217
>>> # Using 'none' reduction type.
>>> l = tf.keras.losses.LogCosh(
... reduction=tf.keras.losses.Reduction.NONE)
>>> l(y_true, y_pred).numpy()
array([0.217, 0.], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.LogCosh())
```
"""
def __init__(self, reduction=losses_utils.ReductionV2.AUTO, name='log_cosh'):
"""Initializes `LogCosh` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'log_cosh'.
"""
super().__init__(log_cosh, name=name, reduction=reduction)
class KLDivergence(LossFunctionWrapper):
"""Computes Kullback-Leibler divergence loss between `y_true` and `y_pred`.
`loss = y_true * log(y_true / y_pred)`
See: https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence
Standalone usage:
>>> y_true = [[0, 1], [0, 0]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> kl = tf.keras.losses.KLDivergence()
>>> kl(y_true, y_pred).numpy()
0.458
>>> # Calling with 'sample_weight'.
>>> kl(y_true, y_pred, sample_weight=[0.8, 0.2]).numpy()
0.366
>>> # Using 'sum' reduction type.
>>> kl = tf.keras.losses.KLDivergence(
... reduction=tf.keras.losses.Reduction.SUM)
>>> kl(y_true, y_pred).numpy()
0.916
>>> # Using 'none' reduction type.
>>> kl = tf.keras.losses.KLDivergence(
... reduction=tf.keras.losses.Reduction.NONE)
>>> kl(y_true, y_pred).numpy()
array([0.916, -3.08e-06], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.KLDivergence())
```
"""
def __init__(self,
reduction=losses_utils.ReductionV2.AUTO,
name='kl_divergence'):
"""Initializes `KLDivergence` instance.
Args:
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'kl_divergence'.
"""
super().__init__(kl_divergence, name=name, reduction=reduction)
class Huber(LossFunctionWrapper):
"""Computes the Huber loss between `y_true` and `y_pred`.
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
Standalone usage:
>>> y_true = [[0, 1], [0, 0]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> h = tf.keras.losses.Huber()
>>> h(y_true, y_pred).numpy()
0.155
>>> # Calling with 'sample_weight'.
>>> h(y_true, y_pred, sample_weight=[1, 0]).numpy()
0.09
>>> # Using 'sum' reduction type.
>>> h = tf.keras.losses.Huber(
... reduction=tf.keras.losses.Reduction.SUM)
>>> h(y_true, y_pred).numpy()
0.31
>>> # Using 'none' reduction type.
>>> h = tf.keras.losses.Huber(
... reduction=tf.keras.losses.Reduction.NONE)
>>> h(y_true, y_pred).numpy()
array([0.18, 0.13], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.Huber())
```
"""
def __init__(self,
delta=1.0,
reduction=losses_utils.ReductionV2.AUTO,
name='huber_loss'):
"""Initializes `Huber` instance.
Args:
delta: A float, the point where the Huber loss function changes from a
quadratic to linear.
reduction: Type of `tf.keras.losses.Reduction` to apply to
loss. Default value is `AUTO`. `AUTO` indicates that the reduction
option will be determined by the usage context. For almost all cases
this defaults to `SUM_OVER_BATCH_SIZE`. When used with
`tf.distribute.Strategy`, outside of built-in training loops such as
`tf.keras` `compile` and `fit`, using `AUTO` or `SUM_OVER_BATCH_SIZE`
will raise an error. Please see this custom training [tutorial](
https://www.tensorflow.org/tutorials/distribute/custom_training) for
more details.
name: Optional name for the instance. Defaults to 'huber_loss'.
"""
super().__init__(huber, name=name, reduction=reduction, delta=delta)
@dispatch.add_dispatch_support
def mean_squared_error(y_true, y_pred):
"""Computes the mean squared error between labels and predictions.
After computing the squared distance between the inputs, the mean value over
the last dimension is returned.
`loss = mean(square(y_true - y_pred), axis=-1)`
Standalone usage:
>>> y_true = np.random.randint(0, 2, size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.mean_squared_error(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> assert np.array_equal(
... loss.numpy(), np.mean(np.square(y_true - y_pred), axis=-1))
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Mean squared error values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
return backend.mean(math_ops.squared_difference(y_pred, y_true), axis=-1)
def _ragged_tensor_apply_loss(loss_fn, y_true, y_pred, y_pred_extra_dim=False):
"""Apply a loss function on a per batch basis.
Args:
loss_fn: The loss function
y_true: truth values (RaggedTensor)
y_pred: predicted values (RaggedTensor)
y_pred_extra_dim: whether y_pred has an additional dimension compared to
y_true
Returns:
Loss-function result. A dense tensor if the output has a single dimension
(per-batch loss value); a ragged tensor otherwise.
"""
def rt_is_equiv_dense(rt):
"""Returns true if this RaggedTensor has the same row_lenghts across
all ragged dimensions and thus can be converted to a dense tensor
without loss of information.
Args:
rt: RaggedTensor.
"""
return math_ops.reduce_all([
math_ops.equal(
math_ops.reduce_variance(math_ops.cast(row_lens, backend.floatx())),
constant_op.constant([0.])) for row_lens in rt.nested_row_lengths()
])
def _convert_to_dense(inputs):
return tuple(
rt.to_tensor() if isinstance(rt, ragged_tensor.RaggedTensor) else rt
for rt in inputs)
def _call_loss(inputs, ragged_output):
""" Adapt the result to ragged or dense tensor according to the expected
output type. This is done so that all the return values of the map
operation have the same type.
"""
r = loss_fn(*inputs)
if ragged_output and not isinstance(r, ragged_tensor.RaggedTensor):
r = ragged_tensor.RaggedTensor.from_tensor(r)
elif not ragged_output and isinstance(r, ragged_tensor.RaggedTensor):
r = r.to_tensor()
return r
def _wrapper(inputs, ragged_output):
_, y_pred = inputs
if isinstance(y_pred, ragged_tensor.RaggedTensor):
return cond.cond(
rt_is_equiv_dense(y_pred),
lambda: _call_loss(_convert_to_dense(inputs), ragged_output),
lambda: _call_loss(inputs, ragged_output))
return loss_fn(*inputs)
if not isinstance(y_true, ragged_tensor.RaggedTensor):
return loss_fn(y_true, y_pred.to_tensor())
lshape = y_pred.shape.as_list()[1:-1]
if len(lshape) > 0:
spec = ragged_tensor.RaggedTensorSpec(shape=lshape, dtype=y_pred.dtype)
else:
spec = tensor_spec.TensorSpec(shape=[], dtype=y_pred.dtype)
nested_splits_list = [rt.nested_row_splits for rt in (y_true, y_pred)]
if y_pred_extra_dim:
# The last dimension of a categorical prediction may be ragged or not.
rdims = [len(slist) for slist in nested_splits_list]
if rdims[0] == rdims[1] - 1:
nested_splits_list[1] = nested_splits_list[1][:-1]
map_fn = functools.partial(_wrapper, ragged_output=len(lshape) > 1)
assertion_list = ragged_util.assert_splits_match(nested_splits_list)
with ops.control_dependencies(assertion_list):
return ragged_map_ops.map_fn(map_fn, elems=(y_true, y_pred), dtype=spec)
@dispatch.dispatch_for_types(mean_squared_error, ragged_tensor.RaggedTensor)
def _ragged_tensor_mse(y_true, y_pred):
"""Implements support for handling RaggedTensors.
Args:
y_true: RaggedTensor truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: RaggedTensor predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Mean squared error values. shape = `[batch_size, d0, .. dN-1]`.
When the number of dimensions of the batch feature vector [d0, .. dN] is
greater than one the return value is a RaggedTensor. Otherwise a Dense
tensor with dimensions [batch_size] is returned.
"""
return _ragged_tensor_apply_loss(mean_squared_error, y_true, y_pred)
@dispatch.add_dispatch_support
def mean_absolute_error(y_true, y_pred):
"""Computes the mean absolute error between labels and predictions.
`loss = mean(abs(y_true - y_pred), axis=-1)`
Standalone usage:
>>> y_true = np.random.randint(0, 2, size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.mean_absolute_error(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> assert np.array_equal(
... loss.numpy(), np.mean(np.abs(y_true - y_pred), axis=-1))
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Mean absolute error values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
return backend.mean(math_ops.abs(y_pred - y_true), axis=-1)
@dispatch.dispatch_for_types(mean_absolute_error, ragged_tensor.RaggedTensor)
def _ragged_tensor_mae(y_true, y_pred):
"""RaggedTensor adapter for mean_absolute_error."""
return _ragged_tensor_apply_loss(mean_absolute_error, y_true, y_pred)
@dispatch.add_dispatch_support
def mean_absolute_percentage_error(y_true, y_pred):
"""Computes the mean absolute percentage error between `y_true` and `y_pred`.
`loss = 100 * mean(abs((y_true - y_pred) / y_true), axis=-1)`
Standalone usage:
>>> y_true = np.random.random(size=(2, 3))
>>> y_true = np.maximum(y_true, 1e-7) # Prevent division by zero
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.mean_absolute_percentage_error(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> assert np.array_equal(
... loss.numpy(),
... 100. * np.mean(np.abs((y_true - y_pred) / y_true), axis=-1))
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Mean absolute percentage error values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
diff = math_ops.abs(
(y_true - y_pred) / backend.maximum(math_ops.abs(y_true),
backend.epsilon()))
return 100. * backend.mean(diff, axis=-1)
@dispatch.dispatch_for_types(mean_absolute_percentage_error,
ragged_tensor.RaggedTensor)
def _ragged_tensor_mape(y_true, y_pred):
"""Support RaggedTensors."""
return _ragged_tensor_apply_loss(mean_absolute_percentage_error, y_true,
y_pred)
@dispatch.add_dispatch_support
def mean_squared_logarithmic_error(y_true, y_pred):
"""Computes the mean squared logarithmic error between `y_true` and `y_pred`.
`loss = mean(square(log(y_true + 1) - log(y_pred + 1)), axis=-1)`
Standalone usage:
>>> y_true = np.random.randint(0, 2, size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.mean_squared_logarithmic_error(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> y_true = np.maximum(y_true, 1e-7)
>>> y_pred = np.maximum(y_pred, 1e-7)
>>> assert np.allclose(
... loss.numpy(),
... np.mean(
... np.square(np.log(y_true + 1.) - np.log(y_pred + 1.)), axis=-1))
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Mean squared logarithmic error values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
first_log = math_ops.log(backend.maximum(y_pred, backend.epsilon()) + 1.)
second_log = math_ops.log(backend.maximum(y_true, backend.epsilon()) + 1.)
return backend.mean(
math_ops.squared_difference(first_log, second_log), axis=-1)
@dispatch.dispatch_for_types(mean_squared_logarithmic_error,
ragged_tensor.RaggedTensor)
def _ragged_tensor_msle(y_true, y_pred):
"""Implements support for handling RaggedTensors."""
return _ragged_tensor_apply_loss(mean_squared_logarithmic_error, y_true,
y_pred)
def _maybe_convert_labels(y_true):
"""Converts binary labels into -1/1."""
are_zeros = math_ops.equal(y_true, 0)
are_ones = math_ops.equal(y_true, 1)
is_binary = math_ops.reduce_all(math_ops.logical_or(are_zeros, are_ones))
def _convert_binary_labels():
# Convert the binary labels to -1 or 1.
return 2. * y_true - 1.
updated_y_true = smart_cond.smart_cond(is_binary, _convert_binary_labels,
lambda: y_true)
return updated_y_true
@dispatch.add_dispatch_support
def squared_hinge(y_true, y_pred):
"""Computes the squared hinge loss between `y_true` and `y_pred`.
`loss = mean(square(maximum(1 - y_true * y_pred, 0)), axis=-1)`
Standalone usage:
>>> y_true = np.random.choice([-1, 1], size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.squared_hinge(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> assert np.array_equal(
... loss.numpy(),
... np.mean(np.square(np.maximum(1. - y_true * y_pred, 0.)), axis=-1))
Args:
y_true: The ground truth values. `y_true` values are expected to be -1 or 1.
If binary (0 or 1) labels are provided we will convert them to -1 or 1.
shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Squared hinge loss values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
y_true = _maybe_convert_labels(y_true)
return backend.mean(
math_ops.square(math_ops.maximum(1. - y_true * y_pred, 0.)), axis=-1)
@dispatch.add_dispatch_support
def hinge(y_true, y_pred):
"""Computes the hinge loss between `y_true` and `y_pred`.
`loss = mean(maximum(1 - y_true * y_pred, 0), axis=-1)`
Standalone usage:
>>> y_true = np.random.choice([-1, 1], size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.hinge(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> assert np.array_equal(
... loss.numpy(),
... np.mean(np.maximum(1. - y_true * y_pred, 0.), axis=-1))
Args:
y_true: The ground truth values. `y_true` values are expected to be -1 or 1.
If binary (0 or 1) labels are provided they will be converted to -1 or 1.
shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Hinge loss values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
y_true = _maybe_convert_labels(y_true)
return backend.mean(math_ops.maximum(1. - y_true * y_pred, 0.), axis=-1)
@dispatch.add_dispatch_support
def categorical_hinge(y_true, y_pred):
"""Computes the categorical hinge loss between `y_true` and `y_pred`.
`loss = maximum(neg - pos + 1, 0)`
where `neg=maximum((1-y_true)*y_pred) and pos=sum(y_true*y_pred)`
Standalone usage:
>>> y_true = np.random.randint(0, 3, size=(2,))
>>> y_true = tf.keras.utils.to_categorical(y_true, num_classes=3)
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.categorical_hinge(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> pos = np.sum(y_true * y_pred, axis=-1)
>>> neg = np.amax((1. - y_true) * y_pred, axis=-1)
>>> assert np.array_equal(loss.numpy(), np.maximum(0., neg - pos + 1.))
Args:
y_true: The ground truth values. `y_true` values are expected to be
either `{-1, +1}` or `{0, 1}` (i.e. a one-hot-encoded tensor).
y_pred: The predicted values.
Returns:
Categorical hinge loss values.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
pos = math_ops.reduce_sum(y_true * y_pred, axis=-1)
neg = math_ops.reduce_max((1. - y_true) * y_pred, axis=-1)
zero = math_ops.cast(0., y_pred.dtype)
return math_ops.maximum(neg - pos + 1., zero)
@dispatch.add_dispatch_support
def huber(y_true, y_pred, delta=1.0):
"""Computes Huber loss value.
For each value x in `error = y_true - y_pred`:
```
loss = 0.5 * x^2 if |x| <= d
loss = d * |x| - 0.5 * d^2 if |x| > d
```
where d is `delta`. See: https://en.wikipedia.org/wiki/Huber_loss
Args:
y_true: tensor of true targets.
y_pred: tensor of predicted targets.
delta: A float, the point where the Huber loss function changes from a
quadratic to linear.
Returns:
Tensor with one scalar loss entry per sample.
"""
y_pred = math_ops.cast(y_pred, dtype=backend.floatx())
y_true = math_ops.cast(y_true, dtype=backend.floatx())
delta = math_ops.cast(delta, dtype=backend.floatx())
error = math_ops.subtract(y_pred, y_true)
abs_error = math_ops.abs(error)
half = tensor_conversion.convert_to_tensor_v2_with_dispatch(
0.5, dtype=abs_error.dtype
)
return backend.mean(
array_ops.where_v2(abs_error <= delta, half * math_ops.square(error),
delta * abs_error - half * math_ops.square(delta)),
axis=-1)
@dispatch.add_dispatch_support
def log_cosh(y_true, y_pred):
"""Logarithm of the hyperbolic cosine of the prediction error.
`log(cosh(x))` is approximately equal to `(x ** 2) / 2` for small `x` and
to `abs(x) - log(2)` for large `x`. This means that 'logcosh' works mostly
like the mean squared error, but will not be so strongly affected by the
occasional wildly incorrect prediction.
Standalone usage:
>>> y_true = np.random.random(size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.logcosh(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> x = y_pred - y_true
>>> assert np.allclose(
... loss.numpy(),
... np.mean(x + np.log(np.exp(-2. * x) + 1.) - math_ops.log(2.), axis=-1),
... atol=1e-5)
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Logcosh error values. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
def _logcosh(x):
return x + math_ops.softplus(-2. * x) - math_ops.cast(
math_ops.log(2.), x.dtype)
return backend.mean(_logcosh(y_pred - y_true), axis=-1)
@dispatch.add_dispatch_support
def categorical_crossentropy(y_true,
y_pred,
from_logits=False,
label_smoothing=0,
axis=-1):
"""Computes the categorical crossentropy loss.
Standalone usage:
>>> y_true = [[0, 1, 0], [0, 0, 1]]
>>> y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
>>> loss = tf.keras.losses.categorical_crossentropy(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> loss.numpy()
array([0.0513, 2.303], dtype=float32)
Args:
y_true: Tensor of one-hot true targets.
y_pred: Tensor of predicted targets.
from_logits: Whether `y_pred` is expected to be a logits tensor. By default,
we assume that `y_pred` encodes a probability distribution.
label_smoothing: Float in [0, 1]. If > `0` then smooth the labels. For
example, if `0.1`, use `0.1 / num_classes` for non-target labels
and `0.9 + 0.1 / num_classes` for target labels.
axis: Defaults to -1. The dimension along which the entropy is
computed.
Returns:
Categorical crossentropy loss value.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
label_smoothing = tensor_conversion.convert_to_tensor_v2_with_dispatch(
label_smoothing, dtype=backend.floatx()
)
def _smooth_labels():
num_classes = math_ops.cast(array_ops.shape(y_true)[-1], y_pred.dtype)
return y_true * (1.0 - label_smoothing) + (label_smoothing / num_classes)
y_true = smart_cond.smart_cond(label_smoothing, _smooth_labels,
lambda: y_true)
return backend.categorical_crossentropy(
y_true, y_pred, from_logits=from_logits, axis=axis)
@dispatch.dispatch_for_types(categorical_crossentropy,
ragged_tensor.RaggedTensor)
def _ragged_tensor_categorical_crossentropy(y_true,
y_pred,
from_logits=False,
label_smoothing=0,
axis=-1):
"""Implements support for handling RaggedTensors.
Args:
y_true: Tensor of one-hot true targets.
y_pred: Tensor of predicted targets.
from_logits: Whether `y_pred` is expected to be a logits tensor. By default,
we assume that `y_pred` encodes a probability distribution.
label_smoothing: Float in [0, 1]. If > `0` then smooth the labels. For
example, if `0.1`, use `0.1 / num_classes` for non-target labels
and `0.9 + 0.1 / num_classes` for target labels.
axis: The axis along which to compute crossentropy (the features axis).
Defaults to -1.
Returns:
Categorical crossentropy loss value.
Expected shape: (batch, sequence_len, n_classes) with sequence_len
being variable per batch.
Return shape: (batch, sequence_len).
When used by CategoricalCrossentropy() with the default reduction
(SUM_OVER_BATCH_SIZE), the reduction averages the loss over the
number of elements independent of the batch. E.g. if the RaggedTensor
has 2 batches with [2, 1] values respectivly the resulting loss is
the sum of the individual loss values divided by 3.
"""
fn = functools.partial(
categorical_crossentropy,
from_logits=from_logits,
label_smoothing=label_smoothing,
axis=axis)
return _ragged_tensor_apply_loss(fn, y_true, y_pred)
@dispatch.add_dispatch_support
def sparse_categorical_crossentropy(y_true, y_pred, from_logits=False, axis=-1):
"""Computes the sparse categorical crossentropy loss.
Standalone usage:
>>> y_true = [1, 2]
>>> y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
>>> loss = tf.keras.losses.sparse_categorical_crossentropy(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> loss.numpy()
array([0.0513, 2.303], dtype=float32)
Args:
y_true: Ground truth values.
y_pred: The predicted values.
from_logits: Whether `y_pred` is expected to be a logits tensor. By default,
we assume that `y_pred` encodes a probability distribution.
axis: Defaults to -1. The dimension along which the entropy is
computed.
Returns:
Sparse categorical crossentropy loss value.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
return backend.sparse_categorical_crossentropy(
y_true, y_pred, from_logits=from_logits, axis=axis)
@dispatch.dispatch_for_types(sparse_categorical_crossentropy,
ragged_tensor.RaggedTensor)
def _ragged_tensor_sparse_categorical_crossentropy(y_true,
y_pred,
from_logits=False,
axis=-1):
""" Implements support for handling RaggedTensors.
Expected y_pred shape: (batch, sequence_len, n_classes) with sequence_len
being variable per batch.
Return shape: (batch, sequence_len).
When used by SparseCategoricalCrossentropy() with the default reduction
(SUM_OVER_BATCH_SIZE), the reduction averages the loss over the
number of elements independent of the batch. E.g. if the RaggedTensor
has 2 batches with [2, 1] values respectively, the resulting loss is
the sum of the individual loss values divided by 3.
"""
fn = functools.partial(
sparse_categorical_crossentropy, from_logits=from_logits, axis=axis)
return _ragged_tensor_apply_loss(fn, y_true, y_pred, y_pred_extra_dim=True)
@dispatch.add_dispatch_support
def binary_crossentropy(y_true,
y_pred,
from_logits=False,
label_smoothing=0,
axis=-1):
"""Computes the binary crossentropy loss.
Standalone usage:
>>> y_true = [[0, 1], [0, 0]]
>>> y_pred = [[0.6, 0.4], [0.4, 0.6]]
>>> loss = tf.keras.losses.binary_crossentropy(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> loss.numpy()
array([0.916 , 0.714], dtype=float32)
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
from_logits: Whether `y_pred` is expected to be a logits tensor. By default,
we assume that `y_pred` encodes a probability distribution.
label_smoothing: Float in [0, 1]. If > `0` then smooth the labels by
squeezing them towards 0.5 That is, using `1. - 0.5 * label_smoothing`
for the target class and `0.5 * label_smoothing` for the non-target class.
axis: The axis along which the mean is computed. Defaults to -1.
Returns:
Binary crossentropy loss value. shape = `[batch_size, d0, .. dN-1]`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
label_smoothing = tensor_conversion.convert_to_tensor_v2_with_dispatch(
label_smoothing, dtype=backend.floatx()
)
def _smooth_labels():
return y_true * (1.0 - label_smoothing) + 0.5 * label_smoothing
y_true = smart_cond.smart_cond(label_smoothing, _smooth_labels,
lambda: y_true)
return backend.mean(
backend.binary_crossentropy(y_true, y_pred, from_logits=from_logits),
axis=axis)
@dispatch.dispatch_for_types(binary_crossentropy, ragged_tensor.RaggedTensor)
def _ragged_tensor_binary_crossentropy(y_true,
y_pred,
from_logits=False,
label_smoothing=0,
axis=-1):
"""Implements support for handling RaggedTensors.
Args:
y_true: Tensor of one-hot true targets.
y_pred: Tensor of predicted targets.
from_logits: Whether `y_pred` is expected to be a logits tensor. By default,
we assume that `y_pred` encodes a probability distribution.
label_smoothing: Float in [0, 1]. If > `0` then smooth the labels. For
example, if `0.1`, use `0.1 / num_classes` for non-target labels
and `0.9 + 0.1 / num_classes` for target labels.
axis: Axis along which to compute crossentropy.
Returns:
Binary crossentropy loss value.
Expected shape: (batch, sequence_len) with sequence_len being variable
per batch.
Return shape: (batch,); returns the per batch mean of the loss values.
When used by BinaryCrossentropy() with the default reduction
(SUM_OVER_BATCH_SIZE), the reduction averages the per batch losses over
the number of batches.
"""
fn = functools.partial(
binary_crossentropy,
from_logits=from_logits,
label_smoothing=label_smoothing,
axis=axis)
return _ragged_tensor_apply_loss(fn, y_true, y_pred)
@dispatch.add_dispatch_support
def kl_divergence(y_true, y_pred):
"""Computes Kullback-Leibler divergence loss between `y_true` and `y_pred`.
`loss = y_true * log(y_true / y_pred)`
See: https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence
Standalone usage:
>>> y_true = np.random.randint(0, 2, size=(2, 3)).astype(np.float64)
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.kullback_leibler_divergence(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> y_true = tf.keras.backend.clip(y_true, 1e-7, 1)
>>> y_pred = tf.keras.backend.clip(y_pred, 1e-7, 1)
>>> assert np.array_equal(
... loss.numpy(), np.sum(y_true * np.log(y_true / y_pred), axis=-1))
Args:
y_true: Tensor of true targets.
y_pred: Tensor of predicted targets.
Returns:
A `Tensor` with loss.
Raises:
TypeError: If `y_true` cannot be cast to the `y_pred.dtype`.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
y_true = backend.clip(y_true, backend.epsilon(), 1)
y_pred = backend.clip(y_pred, backend.epsilon(), 1)
return math_ops.reduce_sum(y_true * math_ops.log(y_true / y_pred), axis=-1)
@dispatch.add_dispatch_support
def poisson(y_true, y_pred):
"""Computes the Poisson loss between y_true and y_pred.
The Poisson loss is the mean of the elements of the `Tensor`
`y_pred - y_true * log(y_pred)`.
Standalone usage:
>>> y_true = np.random.randint(0, 2, size=(2, 3))
>>> y_pred = np.random.random(size=(2, 3))
>>> loss = tf.keras.losses.poisson(y_true, y_pred)
>>> assert loss.shape == (2,)
>>> y_pred = y_pred + 1e-7
>>> assert np.allclose(
... loss.numpy(), np.mean(y_pred - y_true * np.log(y_pred), axis=-1),
... atol=1e-5)
Args:
y_true: Ground truth values. shape = `[batch_size, d0, .. dN]`.
y_pred: The predicted values. shape = `[batch_size, d0, .. dN]`.
Returns:
Poisson loss value. shape = `[batch_size, d0, .. dN-1]`.
Raises:
InvalidArgumentError: If `y_true` and `y_pred` have incompatible shapes.
"""
y_pred = tensor_conversion.convert_to_tensor_v2_with_dispatch(y_pred)
y_true = math_ops.cast(y_true, y_pred.dtype)
return backend.mean(
y_pred - y_true * math_ops.log(y_pred + backend.epsilon()), axis=-1)
@dispatch.add_dispatch_support
def cosine_similarity(y_true, y_pred, axis=-1):
"""Computes the cosine similarity between labels and predictions.
Note that it is a number between -1 and 1. When it is a negative number
between -1 and 0, 0 indicates orthogonality and values closer to -1
indicate greater similarity. The values closer to 1 indicate greater
dissimilarity. This makes it usable as a loss function in a setting
where you try to maximize the proximity between predictions and
targets. If either `y_true` or `y_pred` is a zero vector, cosine
similarity will be 0 regardless of the proximity between predictions
and targets.
`loss = -sum(l2_norm(y_true) * l2_norm(y_pred))`
Standalone usage:
>>> y_true = [[0., 1.], [1., 1.], [1., 1.]]
>>> y_pred = [[1., 0.], [1., 1.], [-1., -1.]]
>>> loss = tf.keras.losses.cosine_similarity(y_true, y_pred, axis=1)
>>> loss.numpy()
array([-0., -0.999, 0.999], dtype=float32)
Args:
y_true: Tensor of true targets.
y_pred: Tensor of predicted targets.
axis: Axis along which to determine similarity.
Returns:
Cosine similarity tensor.
"""
y_true = nn.l2_normalize(y_true, axis=axis)
y_pred = nn.l2_normalize(y_pred, axis=axis)
return -math_ops.reduce_sum(y_true * y_pred, axis=axis)
class CosineSimilarity(LossFunctionWrapper):
"""Computes the cosine similarity between labels and predictions.
Note that it is a number between -1 and 1. When it is a negative number
between -1 and 0, 0 indicates orthogonality and values closer to -1
indicate greater similarity. The values closer to 1 indicate greater
dissimilarity. This makes it usable as a loss function in a setting
where you try to maximize the proximity between predictions and targets.
If either `y_true` or `y_pred` is a zero vector, cosine similarity will be 0
regardless of the proximity between predictions and targets.
`loss = -sum(l2_norm(y_true) * l2_norm(y_pred))`
Standalone usage:
>>> y_true = [[0., 1.], [1., 1.]]
>>> y_pred = [[1., 0.], [1., 1.]]
>>> # Using 'auto'/'sum_over_batch_size' reduction type.
>>> cosine_loss = tf.keras.losses.CosineSimilarity(axis=1)
>>> # l2_norm(y_true) = [[0., 1.], [1./1.414, 1./1.414]]
>>> # l2_norm(y_pred) = [[1., 0.], [1./1.414, 1./1.414]]
>>> # l2_norm(y_true) . l2_norm(y_pred) = [[0., 0.], [0.5, 0.5]]
>>> # loss = mean(sum(l2_norm(y_true) . l2_norm(y_pred), axis=1))
>>> # = -((0. + 0.) + (0.5 + 0.5)) / 2
>>> cosine_loss(y_true, y_pred).numpy()
-0.5
>>> # Calling with 'sample_weight'.
>>> cosine_loss(y_true, y_pred, sample_weight=[0.8, 0.2]).numpy()
-0.0999
>>> # Using 'sum' reduction type.
>>> cosine_loss = tf.keras.losses.CosineSimilarity(axis=1,
... reduction=tf.keras.losses.Reduction.SUM)
>>> cosine_loss(y_true, y_pred).numpy()
-0.999
>>> # Using 'none' reduction type.
>>> cosine_loss = tf.keras.losses.CosineSimilarity(axis=1,
... reduction=tf.keras.losses.Reduction.NONE)
>>> cosine_loss(y_true, y_pred).numpy()
array([-0., -0.999], dtype=float32)
Usage with the `compile()` API:
```python
model.compile(optimizer='sgd', loss=tf.keras.losses.CosineSimilarity(axis=1))
```
Args:
axis: The axis along which the cosine similarity is computed
(the features axis). Defaults to -1.
reduction: Type of `tf.keras.losses.Reduction` to apply to loss.
Default value is `AUTO`. `AUTO` indicates that the reduction option will
be determined by the usage context. For almost all cases this defaults to
`SUM_OVER_BATCH_SIZE`. When used with `tf.distribute.Strategy`, outside of
built-in training loops such as `tf.keras` `compile` and `fit`, using
`AUTO` or `SUM_OVER_BATCH_SIZE` will raise an error. Please see this
custom training [tutorial]
(https://www.tensorflow.org/tutorials/distribute/custom_training) for more
details.
name: Optional name for the instance.
"""
def __init__(self,
axis=-1,
reduction=losses_utils.ReductionV2.AUTO,
name='cosine_similarity'):
super().__init__(
cosine_similarity, reduction=reduction, name=name, axis=axis)
# Aliases.
bce = BCE = binary_crossentropy
mse = MSE = mean_squared_error
mae = MAE = mean_absolute_error
mape = MAPE = mean_absolute_percentage_error
msle = MSLE = mean_squared_logarithmic_error
kld = KLD = kullback_leibler_divergence = kl_divergence
logcosh = log_cosh
huber_loss = huber
def is_categorical_crossentropy(loss):
result = ((isinstance(loss, CategoricalCrossentropy) or
(isinstance(loss, LossFunctionWrapper) and
loss.fn == categorical_crossentropy) or
(hasattr(loss, '__name__') and
loss.__name__ == 'categorical_crossentropy') or
(loss == 'categorical_crossentropy')))
return result
def serialize(loss):
"""Serializes loss function or `Loss` instance.
Args:
loss: A Keras `Loss` instance or a loss function.
Returns:
Loss configuration dictionary.
"""
return serialize_keras_object(loss)
def deserialize(name, custom_objects=None):
"""Deserializes a serialized loss class/function instance.
Args:
name: Loss configuration.
custom_objects: Optional dictionary mapping names (strings) to custom
objects (classes and functions) to be considered during deserialization.
Returns:
A Keras `Loss` instance or a loss function.
"""
return deserialize_keras_object(
name,
module_objects=globals(),
custom_objects=custom_objects,
printable_module_name='loss function')
def get(identifier):
"""Retrieves a Keras loss as a `function`/`Loss` class instance.
The `identifier` may be the string name of a loss function or `Loss` class.
>>> loss = tf.keras.losses.get("categorical_crossentropy")
>>> type(loss)
<class 'function'>
>>> loss = tf.keras.losses.get("CategoricalCrossentropy")
>>> type(loss)
<class '...keras.losses.CategoricalCrossentropy'>
You can also specify `config` of the loss to this function by passing dict
containing `class_name` and `config` as an identifier. Also note that the
`class_name` must map to a `Loss` class
>>> identifier = {"class_name": "CategoricalCrossentropy",
... "config": {"from_logits": True}}
>>> loss = tf.keras.losses.get(identifier)
>>> type(loss)
<class '...keras.losses.CategoricalCrossentropy'>
Args:
identifier: A loss identifier. One of None or string name of a loss
function/class or loss configuration dictionary or a loss function or a
loss class instance.
Returns:
A Keras loss as a `function`/ `Loss` class instance.
Raises:
ValueError: If `identifier` cannot be interpreted.
"""
if identifier is None:
return None
if isinstance(identifier, str):
identifier = str(identifier)
return deserialize(identifier)
if isinstance(identifier, dict):
return deserialize(identifier)
if callable(identifier):
return identifier
raise ValueError(
f'Could not interpret loss function identifier: {identifier}')
LABEL_DTYPES_FOR_LOSSES = {
losses_impl.sparse_softmax_cross_entropy: 'int32',
sparse_categorical_crossentropy: 'int32'
}