Checks if the targets are in the top-k predictions.
tf.keras.ops.in_top_k(
targets, predictions, k
)
Args |
targets
|
A tensor of true labels.
|
predictions
|
A tensor of predicted labels.
|
k
|
An integer representing the number of predictions to consider.
|
Returns |
A boolean tensor of the same shape as targets , where each element
indicates whether the corresponding target is in the top-k predictions.
|
Example:
targets = keras.ops.convert_to_tensor([2, 5, 3])
predictions = keras.ops.convert_to_tensor(
[[0.1, 0.4, 0.6, 0.9, 0.5],
[0.1, 0.7, 0.9, 0.8, 0.3],
[0.1, 0.6, 0.9, 0.9, 0.5]])
in_top_k(targets, predictions, k=3)
array([ True False True], shape=(3,), dtype=bool)