gpflow.optimizers.natgrad

gpflow.optimizers.natgrad.expectation_to_meanvarsqrt

gpflow.optimizers.natgrad.expectation_to_meanvarsqrt(eta1: tensorflow.Tensor, eta2: tensorflow.Tensor)[source]
Parameters
  • eta1 (tensorflow.Tensor) –

  • eta2 (tensorflow.Tensor) –

gpflow.optimizers.natgrad.expectation_to_natural

gpflow.optimizers.natgrad.expectation_to_natural(eta1: tensorflow.Tensor, eta2: tensorflow.Tensor)[source]
Parameters
  • eta1 (tensorflow.Tensor) –

  • eta2 (tensorflow.Tensor) –

gpflow.optimizers.natgrad.meanvarsqrt_to_expectation

gpflow.optimizers.natgrad.meanvarsqrt_to_expectation(m: tensorflow.Tensor, v_sqrt: tensorflow.Tensor)[source]
Parameters
  • m (tensorflow.Tensor) –

  • v_sqrt (tensorflow.Tensor) –

gpflow.optimizers.natgrad.meanvarsqrt_to_natural

gpflow.optimizers.natgrad.meanvarsqrt_to_natural(mu: tensorflow.Tensor, s_sqrt: tensorflow.Tensor)[source]
Parameters
  • mu (tensorflow.Tensor) –

  • s_sqrt (tensorflow.Tensor) –

gpflow.optimizers.natgrad.natural_to_expectation

gpflow.optimizers.natgrad.natural_to_expectation(nat1: tensorflow.Tensor, nat2: tensorflow.Tensor)[source]
Parameters
  • nat1 (tensorflow.Tensor) –

  • nat2 (tensorflow.Tensor) –

gpflow.optimizers.natgrad.natural_to_meanvarsqrt

gpflow.optimizers.natgrad.natural_to_meanvarsqrt(nat1: tensorflow.Tensor, nat2: tensorflow.Tensor)[source]
Parameters
  • nat1 (tensorflow.Tensor) –

  • nat2 (tensorflow.Tensor) –

gpflow.optimizers.natgrad.swap_dimensions

gpflow.optimizers.natgrad.swap_dimensions(method)[source]

Converts between GPflow indexing and tensorflow indexing method is a function that broadcasts over the first dimension (i.e. like all tensorflow matrix ops):

method inputs [D, N, 1], [D, N, N] method outputs [D, N, 1], [D, N, N]

Returns

Function that broadcasts over the final dimension (i.e. compatible with GPflow): inputs: [N, D], [D, N, N] outputs: [N, D], [D, N, N]