tf.ragged.map_flat_values | TensorFlow v2.16.1 (original) (raw)
tf.ragged.map_flat_values
Stay organized with collections Save and categorize content based on your preferences.
Applies op to the flat_values of one or more RaggedTensors.
View aliases
Compat aliases for migration
SeeMigration guide for more details.
tf.compat.v1.ragged.map_flat_values
tf.ragged.map_flat_values(
op, *args, **kwargs
)
Used in the notebooks
| Used in the guide |
|---|
| Ragged tensors |
Replaces any RaggedTensor in args or kwargs with its flat_valuestensor (which collapses all ragged dimensions), and then calls op. Returns a RaggedTensor that is constructed from the input RaggedTensors'nested_row_splits and the value returned by the op.
If the input arguments contain multiple RaggedTensors, then they must have identical nested_row_splits.
This operation is generally used to apply elementwise operations to each value in a RaggedTensor.
Examples:
rt = tf.ragged.constant([[1, 2, 3], [], [4, 5], [6]])
tf.ragged.map_flat_values(tf.ones_like, rt)
<tf.RaggedTensor [[1, 1, 1], [], [1, 1], [1]]>
tf.ragged.map_flat_values(tf.multiply, rt, rt)
<tf.RaggedTensor [[1, 4, 9], [], [16, 25], [36]]>
tf.ragged.map_flat_values(tf.add, rt, 5)
<tf.RaggedTensor [[6, 7, 8], [], [9, 10], [11]]>
Example with a non-elementwise operation (note that map_flat_values andmap_fn return different results):
rt = tf.ragged.constant([[1.0, 3.0], [], [3.0, 6.0, 3.0]])
def normalized(x):
return x / tf.reduce_sum(x)
tf.ragged.map_flat_values(normalized, rt)
<tf.RaggedTensor [[0.0625, 0.1875], [], [0.1875, 0.375, 0.1875]]>
tf.map_fn(normalized, rt)
<tf.RaggedTensor [[0.25, 0.75], [], [0.25, 0.5, 0.25]]>
| Args | |
|---|---|
| op | The operation that should be applied to the RaggedTensor flat_values.op is typically an element-wise operation (such as math_ops.add), but any operation that preserves the size of the outermost dimension can be used. I.e., shape[0] of the value returned by op must matchshape[0] of the RaggedTensors' flat_values tensors. |
| *args | Arguments for op. |
| **kwargs | Keyword arguments for op. |
| Returns |
|---|
| A RaggedTensor whose ragged_rank matches the ragged_rank of all input RaggedTensors. |
| Raises | |
|---|---|
| ValueError | If args contains no RaggedTensors, or if the nested_splitsof the input RaggedTensors are not identical. |