v188.8.131.52 (16. Aug 2018)¶
- Update Stan source to v2.18.0 (release notes)
- Fit method
to_dataframeorganizes draws into a Pandas DataFrame. Only works if Pandas is installed. Thanks to Liam Brannigan and Ari Hartikainen.
- Improved effective sample size calculation. Thanks to Aki Vehtari.
pystan.diagnostics. Thanks to @jjramsey.
- Several minor bug fixes.
v184.108.40.206 (16. Jan 2018)¶
- Update Stan source to v2.17.1 (release notes) (bugfix release)
- Remove deprecation warnings (Thanks to Alexander Rudiuk)
- Drop testing of Python 3.4.
v220.127.116.11 (6. Oct 2017)¶
- Update Stan source to v2.17.0 (release notes),
pystan.stansummaryfunction. Patch by @ahartikainen.
pystan.stanas deprecated. It will be removed in version 3.0. Please compile and use a Stan program as two separate steps.
- Reminder: Consider using Python 3.5 or higher. In a future release, Python 3.4 wheels will no longer be built.
- Reminder: Stan v2.16.0 is the final release which will not require a C++11 compatible compiler. Future releases will require a C++11 compatible compiler. (This does not affect most users.)
v18.104.22.168 (22. June 2017)¶
- Update Stan source to v2.16.0 (release notes),
- Ari Hartikainen (Aalto University) @ahartikainen joins the Stan development team. 🎉
pystan.lookup(contributed by Marco Inacio, @randommm)
- NOTE: Stan v2.16.0 is the final release which will not require a C++11 compatible compiler. Future releases will require a C+11 compatible compiler. The vast majority of users have a compatible compiler.
v22.214.171.124 (2. May 2017)¶
- Python 2.7 compatibility fix (#332). Thanks to @monga for the report.
v126.96.36.199 (21. Apr 2017)¶
- Update Stan source to v2.15.0 (release notes)
- Allow UTF-8 comments in Stan Program code. Thanks to @ahartikainen
- Expose constrain_pars method, thanks to Lars Mescheder.
V188.8.131.52 (1. Jan 2017)¶
- Update Stan source to v2.14.0 (release notes), includes important fix to the default sampling algorithm (NUTS). All users are encouraged to upgrade.
- Several documentation and minor bug fixes (thanks @ahartikainen, @jrings, @nesanders)
- New OpenPGP signing key for use with PyPI. Key fingerprint is C3542448245BEC68F43070E4CCB669D9761F0CAC.
V184.108.40.206 (15. Sept 2016)¶
- Update Stan source to v2.12.0 (release notes)
- #239 Fix bug in array indexing (thanks @stephen-hoover)
- #254 FIx off-by-one error in estimated sample size calculation
V220.127.116.11 (18. July 2016)¶
- Update Stan source to v2.10.0 (release notes)
- Sampling in
Fixed_parammode now works. Thanks to @luac for the fix and @axch for the original report.
- Detailed installation instructions from @chendaniely added to the documentation.
v18.104.22.168 (7. Jan 2016)¶
- Update Stan source to v2.9.0 (release notes)
- Bugs fixed in _chains.pyx and model.py (thanks to @stephen-hoover, Paul Kernfeld)
v22.214.171.124 (6. Nov 2015)¶
- Cython 0.22 or higher requirement included on PyPI
v126.96.36.199 (5. Nov 2015)¶
- Python 3.5 support added
- Cython 0.22 or higher now required
- Compiler optimization (
-O2) turned on for model compilation. This should increase sampling speed.
- Significant bug fixes (pickling, pars keyword)
v188.8.131.52 (22. August 2015)¶
- Minor Cython 0.23.1 compatibility fixes
- Bug preventing mean_pars from being recorded
v184.108.40.206 (30. Jan 2015)¶
- Fix bug in rdump (for >1 dimensional arrays)
v220.127.116.11 (14. Nov 2014)¶
- Support for pickling fit objects (experimental)
- Fix bug that occurs when printing fit summary
v18.104.22.168 (21. Oct 2014)¶
- Update Stan source to v2.5.0
- Fix several significant bugs in the
v22.214.171.124 (9. Sept 2014)¶
- Performance improvements for the printed summary of a fit.
v126.96.36.199 (6. Sept 2014)¶
- Performance improvements for the
extractmethod (5-10 times faster)
- Performance improvements for the printed summary of a fit. Printing
a summary of a model with more than a hundred parameters is not recommended.
extractand calculating summary statistics for the parameters of interest.
v188.8.131.52 (31. July 2014)¶
- Sets LBFGS as default optimizer.
- Adds preliminary support for Python binary wheels on OS X and Windows.
- Fixes bug in edge case in new summary code.
v184.108.40.206 (26. July 2014)¶
- Stan 2.4 (LBFGS optimizer added, Nesterov removed)
- Improve display of fit summaries
v220.127.116.11 (26. June 2014)¶
- Stan 2.3 (includes user-defined functions, among other improvements).
- Optimizing returns a vector (array) by default instead of a dictionary.
v18.104.22.168 (30. April 2014)¶
- Add support for reading Stan’s R dump files.
- Add support for specifying parameters of interest in
- Add Windows installation instructions. Thanks to @patricksnape.
- Lighten source distribution.
v22.214.171.124 (16. February 2014)¶
- Updates Stan to v2.2.0.
v126.96.36.199 (27. January 2014)¶
- Implement model name obfuscation. Thanks to @karnold
- Improve documentation of StanFit objects
v188.8.131.52 (26. December 2013)¶
- Updates Stan code to v2.1.0.
v184.108.40.206 (18. December 2013)¶
- Sampling is parallel by default.
grad_log_probmethod of fit objects is available.
v220.127.116.11 (1. December 2013)¶
- Improves setuptools support.
- Allows sampling chains in parallel using multiprocessing. See the
- Allows users to specify initial values for chains.
v18.104.22.168 (18. November 2013)¶
- Clean up random_seed handling (Stephan Hoyer).
- Add fit methods get_seed, get_inits, and get_stancode.
v22.214.171.124 (24. October 2013)¶
- Updated to Stan 2.0.1.
sample_filenow works as expected.
v126.96.36.199 (23. October 2013)¶
arrayparameters are now handled correctly.
- Ancillary methods added to fit instances.
- Fixed bug that caused parameters in
controldict to be ignored.
v188.8.131.52 (21. October 2013)¶
- Stan source updated to to 2.0.0.
- PyStan version now mirrors Stan version.
- Rudimentary plot and traceplot methods have been added to fit instances.
- Warmup and sampling progress now visible.
v.0.2.2 (28. September 2013)¶
log_probmethod available from StanFit instances.
- Estimated sample size and Rhat included in summary.
v.0.2.1 (17. September 2013)¶
StanModelinstances can now be pickled.
- Adds basic support for saving output to
v.0.2.0 (25. August 2013)¶
optimizingmethod working for scalar, vector, and matrix parameters
- stanfit objects now have
__str__methods à la RStan
- stan source updated to commit cc82d51d492d26f754fd56efe22a99191c80217b (July 26, 2013)
- IPython-relevant bug fixes
v.0.1.1 (19. July 2013)¶
- Support for Python 2.7 and Python 3.3
stancworking with common arguments