Applies softmax to a batched N-D
Compat aliases for migration
See Migration guide for more details.
tf.raw_ops.SparseSoftmax( sp_indices, sp_values, sp_shape, name=None )
The inputs represent an N-D SparseTensor with logical shape
[..., B, C]
N >= 2), and with indices sorted in the canonical lexicographic order.
This op is equivalent to applying the normal
tf.nn.softmax() to each innermost
logical submatrix with shape
[B, C], but with the catch that the implicitly
zero elements do not participate. Specifically, the algorithm is equivalent
to the following:
tf.nn.softmax() to a densified view of each innermost submatrix
[B, C], along the size-C dimension;
(2) Masks out the original implicitly-zero locations;
(3) Renormalizes the remaining elements.
SparseTensor result has exactly the same non-zero indices and
||A name for the operation (optional).|