Created by: alvations

When adding the settings.slots in the hook() function of dataprovider.py, is it right that both the dense_vector_sequence() and sparse_vector_sequence input must be a matrix (i.e. list of lists of floats)?

Is it right that for sparse_vector_sequence(), I can have None in my inner list, e.g.:

[ [0.1, 0.2, 0.2], [0.4, None, None] ]

But for dense_vector_sequence(), I cannot have None in my inner list and I have to fill them up with zeros, i.e.:

[ [0.1, 0.2, 0.2], [0.4, 0.0, 0.0] ]

Is that the difference in the sparse and dense vector sequence in Paddle?