v0.6.1

SliceSampling v0.6.1

Diff since v0.6.0

Merged pull requests:

• CompatHelper: bump compat for Turing in [weakdeps] to 0.34, (keep existing compat) (#11) (@github-actions[bot])

v0.6.0

SliceSampling v0.6.0

Diff since v0.5.0

• Add support for Turing.Experimental.Gibbs.

v0.5.0

SliceSampling v0.5.0

Diff since v0.4.2

• A default initializer has been added for Turing models. Thus, sample can be called without specifying initial_params.
• Univariate slice samplers can now be used directly on 1-dimensional targets through the AbstractMCMC interface.

v0.4.2

SliceSampling v0.4.2

Diff since v0.4.1

v0.4.1

SliceSampling v0.4.1

Diff since v0.4.0

v0.4.0

SliceSampling v0.4.0

Diff since v0.3.0

Breaking Changes

• Univariate slice samplers now need to wrapped into RandPermGibbs to be used on multivariate targets.

New Features

• Added the hit-and-run sampling strategy for using univariate slice samplers on multivariate targets.

Merged pull requests:

• Add hit-and-run sampler, separate Gibbs sampling interface (#8) (@Red-Portal)
• enable docs for PRs (#9) (@Red-Portal)

v0.3.0

SliceSampling v0.3.0

• Added Gibbsian polar slice sampling
• Added optional limit on the maximum number of proposals to avoid infinite loops

Merged pull requests:

• Support Turing's externalsampler interface, return Transition instead of raw samples. (#1) (@Red-Portal)
• Add window size check for univariate slice samplers (#2) (@Red-Portal)
• fix doc urls (#3) (@Red-Portal)
• Add Gibbsian polar slice sampler (#4) (@Red-Portal)
• fix README badges (#6) (@Red-Portal)
• Improve docs for GPSS (#7) (@Red-Portal)

Closed issues:

• TagBot trigger issue (#5)