v0.5.0

v0.5.0 (2019-08-23)

Diff since v0.0.1

Closed issues:

• The LBA model is different between Turing's and Stan's (#43)
• Model for DynamicHMC doesn't use truncated distribution (#38)
• How are keyword parameters passed (#37)
• Pin DynamicHMC and LogDensityProblems (#36)
• Does not run with low Nsamples and Nadapt (#35)
• Fix DataFrame Deprecations (#34)
• Standardize names for cross sampler rhat (#33)
• How do you generate the data for LBA? (#32)
• Drop support for DynamicNUTS (#31)
• Chain ranges differ (#30)
• Inconsistency between preconditoning matrix adaptation methods (#29)
• Add Tests (#27)
• How do I suppose to run this repo locally? (#25)
• Change output directory for figures (#24)
• Status of MCMCBenchmarks (#23)
• Memory differences for parallel and single chains (#19)
• plots for HPD (#18)
• Debugging functionality (#16)
• @JuliaRegistrator register\(\) (#15)
• Maximum iterations for LBA (#14)
• Speed up Stan compilation (#13)
• Sketch of Regression example (#12)
• Saving Dataframe for post processing and adding version info (#10)
• Error handling for DynamicNUTS (#3)
• List of models to add for benchmarking (#1)

Merged pull requests:

• Fix another mismatch in the LBA model and make it type stable (#44) (xukai92)
• Optimize Gaussian model for Turing (#42) (xukai92)
• Make LR model type stable and faster (#41) (xukai92)
• Improve scripts (#40) (xukai92)
• Use mv normal instead of iid normals in Poisson (#39) (xukai92)
• Fix Turing's burn-in samples (#26) (xukai92)