Reuse tuning for next sampling call

junpenglao · February 8, 2019, 11:00am

Try also specifying the max_treedepth and step_size

pymc-devs/pymc3/blob/f375f27fdcf6568dd662d8db4474497b9bda1f58/pymc3/step_methods/hmc/nuts.py#L159-L162


if self.tune and self.iter_count < 200:
    max_treedepth = self.early_max_treedepth
else:
    max_treedepth = self.max_treedepth

github.com

pymc-devs/pymc3/blob/f375f27fdcf6568dd662d8db4474497b9bda1f58/pymc3/step_methods/hmc/base_hmc.py#L158




adapt_step = self.tune and self.adapt_step_size
step_size = self.step_adapt.current(adapt_step)
self.step_size = step_size


if self._step_rand is not None:
    step_size = self._step_rand(step_size)


hmc_step = self._hamiltonian_step(start, p0, step_size)


self.step_adapt.update(hmc_step.accept_stat, adapt_step)
self.potential.update(hmc_step.end.q, hmc_step.end.q_grad, self.tune)
if hmc_step.divergence_info:
    info = hmc_step.divergence_info
    if self.tune:
        kind = WarningType.TUNING_DIVERGENCE
        point = None
    else:
        kind = WarningType.DIVERGENCE
        self._num_divs_sample += 1
        # We don't want to fill up all memory with divergence info

Topic		Replies	Views
Pause and resume training a model that can change during training Questions	2	1438	May 2, 2019
Restarting sampling from stored multitrace: mixed sampling and tuning Questions	2	586	October 5, 2020
Saving the state of the sampler in pymc3? Questions	11	1579	August 20, 2020
Reusing tuned NUTS steps to save time Questions	0	438	February 19, 2021
Just using NUTS Questions	7	1168	May 5, 2018

Reuse tuning for next sampling call

Related topics