Deduplicates each row (0-th dimension) of the provided tensor.
SparseTensor. The first
dimension is assumed to be the batch or "row" dimension, and deduplication
is done on the 2nd dimension. If the Tensor is 1D it is returned as the
SparseTensor since the "row" is a scalar can't be further
Optional name for the operation.
SparseTensor containing the unique set of values from each
row of the input. Note: the original order of the input may not be