tfl.kronecker_factored_lattice_lib.evaluate_with_hypercube_interpolation

Evaluates a Kronecker-Factored Lattice using hypercube interpolation.

Kronecker-Factored Lattice function is the product of the piece-wise linear interpolation weights for each dimension of the input.

inputs Tensor representing points to apply lattice interpolation to. If units = 1, tensor should be of shape: (batch_size, ..., dims) or list of dims tensors of same shape (batch_size, ..., 1). If units > 1, tensor should be of shape: (batch_size, ..., units, dims) or list of dims tensors of same shape (batch_size, ..., units, 1). A typical shape is (batch_size, dims).
scale Kronecker-Factored Lattice scale of shape (units, num_terms).
bias Kronecker-Factored Lattice bias of shape (units).
kernel Kronecker-Factored Lattice kernel of shape `(1, lattice_sizes, units

  • dims, num_terms). </td> </tr><tr> <td>units</td> <td> Output dimension of the Kronecker-Factored Lattice. </td> </tr><tr> <td>num_terms</td> <td> Number of independently trained submodels per unit, the outputs of which are averaged to get the final output. </td> </tr><tr> <td>lattice_sizes</td> <td> Number of vertices per dimension. </td> </tr><tr> <td>clip_inputs`
If inputs should be clipped to the input range of the Kronecker-Factored Lattice.

Tensor of shape: (batch_size, ..., units).