ML Community Day is November 9! Join us for updates from TensorFlow, JAX, and more Learn more

tfp.experimental.mcmc.GradientBasedTrajectoryLengthAdaptationResults

Internal state of GradientBasedTrajectoryLengthAdaptation.

tfp.experimental.mcmc.GradientBasedTrajectoryLengthAdaptationResults(
    inner_results, max_trajectory_length, step, adaptation_rate, jitter_amount,
    averaged_sq_grad, averaged_max_trajectory_length, criterion, seed
)

inner_results Results of the inner kernel.
max_trajectory_length Floating point scalar Tensor. Maximum HMC trajectory length.
step Int32 scalar Tensor. The number of steps this kernel has taken. Increases by 1 for every call to one_step.
adaptation_rate Floating point scalar Tensor. How rapidly to adapt the trajectory length.
jitter_amount Floating point scalar Tensor. How much to jitter the trajectory on the next step. The trajectory length is sampled from [(1 - jitter_amount) * max_trajectory_length, max_trajectory_length].
averaged_sq_grad Floating point scalar Tensor. Moving average of squared criterion gradients.
averaged_max_trajectory_length Floating point scalar Tensor. Moving average of the maximum of trajectory length. This is used after the burnin period.
criterion Floating point Tensor with shape [C0, ..., Cb] with b > 0. The value of the criterion returned by the criterion_fn corresponding to each Markov chain.
seed PRNG seed; see tfp.random.sanitize_seed for details. The random seed used by the kernel in the previous step.