ott.neural.methods.flows.genot.GENOT.transport

GENOT.transport(source, condition=None, t0=0.0, t1=1.0, rng=None, **kwargs)

Transport data with the learned plan.

This function pushes forward the source distribution to its conditional distribution by solving the neural ODE.

Parameters:

• source (Array) – Data to transport.

• condition (Optional[Array]) – Condition of the input data.

• t0 (float) – Starting time of integration of neural ODE.

• t1 (float) – End time of integration of neural ODE.

• rng (Optional[Array]) – Random generate used to sample from the latent distribution.

• kwargs (Any) – Keyword arguments for odesolve().

Return type:

Array

Returns:

The push-forward defined by the learned transport plan.