Skip to main content

Ctrl+K

Guides

Tutorials

Model Basics

Model
Var
Dist

MCMC Setup

LieselMCMC
MCMCSpec
EngineBuilder
Engine

MCMC Kernels

IWLSKernel
NUTSKernel
HMCKernel
RWKernel
MHKernel
MHProposal
- MHProposal.log_correction
- MHProposal.position
GibbsKernel

MCMC Results & Summary

SamplingResults
Summary
SamplesSummary
loo()

Plots

plot_trace()
plot_cor()
plot_pairs()
plot_scatter()
plot_density()
plot_param()

Optimization

optim_flat()
Stopper
history_to_df()
OptimResult

Model (Advanced)

log_prob_pointwise()
LogProb
FlatLogProb
Calc
Node
Value
PIT()
GaussianCopula
- GaussianCopula.cross_entropy()
- GaussianCopula.kl_divergence()
TransientCalc
- TransientCalc.value
- TransientCalc.monitor
TransientDist
- TransientDist.value
- TransientDist.monitor
TransientIdentity
- TransientIdentity.monitor
TransientNode
InputGroup
- InputGroup.value
- InputGroup.monitor
GraphBuilder

Model Interfaces

LieselInterface
DictInterface
DataclassInterface
NamedTupleInterface

P-Splines

basis_matrix()
equidistant_knots()
pspline_penalty()
MultivariateNormalDegenerate

MCMC (Advanced)

mh_step()
liesel.goose.da
liesel.goose.mm
- tune_inv_mm_diag()
- tune_inv_mm_full()
EpochConfig
EpochState
EpochType
Kernel
TransitionInfo
TuningInfo
- TuningInfo.error_code
- TuningInfo.time
DefaultTransitionInfo
DefaultTuningInfo
- DefaultTuningInfo.error_code
- DefaultTuningInfo.time
TransitionMixin
- TransitionMixin.transition()
TransitionOutcome
TuningOutcome
- TuningOutcome.info
- TuningOutcome.kernel_state
WarmupOutcome
- WarmupOutcome.error_code
- WarmupOutcome.kernel_state
ModelMixin

Experimental API

liesel.experimental
- liesel.experimental.arviz
  - to_arviz_inference_data()
- liesel.experimental.pymc
  - PyMCInterface

.rst

TuningOutcome.kernel_state

Contents

TuningOutcome.kernel_state

TuningOutcome.kernel_state#

TuningOutcome.kernel_state: TKernelState#: A kernel state object, see DAKernelState.

previous

TuningOutcome.info

next

WarmupOutcome

Contents

TuningOutcome.kernel_state

By Hannes Riebl & Paul Wiemann

© Copyright 2022, Hannes Riebl & Paul Wiemann.