How to use layers.Discretization()
across dimention when a ragged tensor innermost dimension is 1 ? Is there any layers.Reshape()
trick?
Like [[[10],[20],[30]]]
to become [[[1],[2],[3]]]
And [[[10],[20],[30],[40]]]
to become [[[1],[2],[3],[4]]]
Since it is ragged, reshape()
need to flexiable right?
Hi @ZacharyLaw,
As far as I’m aware, layers.Discretization
works only with fixed-shape tensors.So, I suggest converting ragged to dense tensors for discretization, and then change them back as ragged. Please take a look into the attached gist for ragged discretization, and let us know if you have any questions.
Thank You.