CondorOrdinalCrossEntropy
CondorOrdinalCrossEntropy(importance_weights=None, from_type='ordinal_logits', name='ordinal_crossent', kwargs)
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))
```
Methods
call(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]`.
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()
Returns the config dictionary for a Loss instance.
ordinal_loss(logits, levels, importance)
Cross-entropy loss function designed for ordinal outcomes.
Parameters
logits: tf.Tensor, shape=(num_samples,num_classes-1) Logit output of the final Dense(num_classes-1) layer.
levels: tf.Tensor, shape=(num_samples, num_classes-1)
Encoded lables provided by CondorOrdinalEncoder.
importance_weights: tf or np array of floats, shape(numclasses-1,)
Importance weights for each binary classification task.
Returns
loss: tf.Tensor, shape=(num_samples,) Loss vector, note that tensorflow will reduce it to a single number automatically.