40ccb10 | Chris Fonnesbeck | 09 January 2017, 14:08:10 UTC | Merge pull request #1647 from pymc-devs/v3.0_release REL Update version to 3.0. | 09 January 2017, 14:08:10 UTC |
5b6bb33 | Thomas Wiecki | 09 January 2017, 11:41:53 UTC | BLD Remove Python 3.6 compatibility. | 09 January 2017, 11:41:53 UTC |
36457cb | Thomas Wiecki | 09 January 2017, 11:00:22 UTC | DOC Update date. Split up authos and outside contributors. | 09 January 2017, 11:00:22 UTC |
69b9461 | Chris Fonnesbeck | 09 January 2017, 03:49:16 UTC | Merge pull request #1637 from pymc-devs/sqlite_record_indent Revert indentation in sqlite.record and clarify syntax | 09 January 2017, 03:49:16 UTC |
4953515 | Osvaldo Martin | 06 January 2017, 20:53:21 UTC | auto-scale the max value of y when njobs > 1 (#1648) | 08 January 2017, 20:47:50 UTC |
7c8cd15 | Christopher Fonnesbeck | 02 January 2017, 04:32:08 UTC | Reverted c0351c19 and clarified reference to varnames element | 08 January 2017, 20:47:50 UTC |
6d02627 | Osvaldo Martin | 06 January 2017, 20:53:21 UTC | auto-scale the max value of y when njobs > 1 (#1648) | 06 January 2017, 20:53:21 UTC |
5740e84 | Colin | 06 January 2017, 20:16:31 UTC | Add Colin's email | 06 January 2017, 20:16:31 UTC |
832ad65 | Thomas Wiecki | 06 January 2017, 12:22:31 UTC | DOC Update release notes. | 06 January 2017, 12:22:31 UTC |
4f6bdf6 | Thomas Wiecki | 06 January 2017, 12:16:55 UTC | REL Update version to 3.0. | 06 January 2017, 12:16:55 UTC |
2394fac | Osvaldo Martin | 06 January 2017, 10:38:39 UTC | add varnames argument for consintency (#1641) * add varnames argument for consintency * do not pass flat_names argument (or varnames) | 06 January 2017, 10:38:39 UTC |
19fcf4b | Thomas Wiecki | 05 January 2017, 20:29:40 UTC | Merge pull request #1645 from aloctavodia/init_sampling use tranformed variables as starting points | 05 January 2017, 20:29:40 UTC |
88125bc | aloctavodia | 05 January 2017, 15:07:48 UTC | use tranformed variables as starting points | 05 January 2017, 15:07:48 UTC |
704a897 | Thomas Wiecki | 03 January 2017, 12:04:46 UTC | Merge pull request #1639 from pymc-devs/adding_book DOCS: Adding Osvaldo Martin book | 03 January 2017, 12:04:46 UTC |
270f42a | springcoil | 03 January 2017, 09:48:29 UTC | DOCS: Adding Osvaldo Martin book | 03 January 2017, 09:48:29 UTC |
0bc5f8f | Thomas Wiecki | 02 January 2017, 13:10:23 UTC | Merge pull request #1623 from huynhnguyen/master fix issue with advi when running on GPU | 02 January 2017, 13:10:23 UTC |
cad8ec2 | Peadar Coyle | 02 January 2017, 00:28:08 UTC | Merging (#1635) | 02 January 2017, 00:28:08 UTC |
cc57a0c | Chris Fonnesbeck | 01 January 2017, 23:17:42 UTC | Removed doc from packages list (#1634) | 01 January 2017, 23:17:42 UTC |
4079f6e | Chris Fonnesbeck | 01 January 2017, 23:03:29 UTC | Fixed indentation in record (#1632) | 01 January 2017, 23:03:29 UTC |
581b704 | Chris Fonnesbeck | 01 January 2017, 20:09:27 UTC | Merge pull request #1630 from pymc-devs/rc5 REL: Incrementing to release candidate 5 | 01 January 2017, 20:09:27 UTC |
87ae225 | springcoil | 01 January 2017, 16:31:15 UTC | REL: Incrementing to release candidate 5 | 01 January 2017, 16:31:15 UTC |
5376771 | huynhnguyen | 01 January 2017, 06:18:14 UTC | avdi: change cast type to theano standard type | 01 January 2017, 06:18:14 UTC |
f20aca4 | Osvaldo Martin | 31 December 2016, 22:35:01 UTC | add jitter to fast_kde (#1629) * add jitter to fast_kde, prevents errors when input values are all the same * remove unnecessary print | 31 December 2016, 22:35:01 UTC |
ca7f68d | Peadar Coyle | 31 December 2016, 19:23:39 UTC | Sampling Output checks and Normality tests to check for bugs (#1618) * Fixing up tests Updating tests Updating tests again Added in one test for each sampler Updating tests Sampling errors Trying this commit Removing parameterized nose tests as they didn't work with the model context Moving asserts and removing arguments which are unnecessary - removing requirements.txt reference to nose_parameterized Changing shape Removing unnecessary test, and checking they run locally * TST Improve test. * TST Add thinning. | 31 December 2016, 19:23:39 UTC |
f450940 | Chris Fonnesbeck | 31 December 2016, 14:30:22 UTC | Updates to README (#1626) | 31 December 2016, 14:30:22 UTC |
f222ffd | huynhnguyen | 29 December 2016, 06:05:11 UTC | fix issue with advi when running on GPU | 29 December 2016, 06:05:11 UTC |
ab437e4 | Colin | 19 December 2016, 23:34:14 UTC | Bounded now works with array bounds (#1604) * Bounded now works with array bounds * Bound is now a function returning a type | 19 December 2016, 23:34:14 UTC |
d9e57e8 | John Salvatier | 19 December 2016, 21:28:29 UTC | Merge pull request #1596 from pymc-devs/fix_bounds2 Make bound broadcast again. | 19 December 2016, 21:28:29 UTC |
6c7f127 | Thomas Wiecki | 19 December 2016, 21:25:22 UTC | MAINT Remove double import of bound. | 19 December 2016, 21:25:22 UTC |
1b01b53 | Thomas Wiecki | 19 December 2016, 15:54:03 UTC | MAINT Instead of separate functions, add broadcast kwarg and make it be used by multivariate distributions. | 19 December 2016, 18:51:28 UTC |
ed89f10 | Thomas Wiecki | 13 December 2016, 17:01:53 UTC | BUG Wrong search/replace. | 19 December 2016, 16:02:27 UTC |
59f68c3 | Thomas Wiecki | 13 December 2016, 16:49:22 UTC | Make all univariate distributions use bound_elemwise. Addd tests. | 19 December 2016, 16:02:27 UTC |
1ad526a | Thomas Wiecki | 13 December 2016, 08:16:57 UTC | MAINT Add back bound_elemwise. Port DiscreteUniform and Uniform to use it. | 19 December 2016, 16:02:27 UTC |
1a1a36e | Colin | 17 December 2016, 18:06:39 UTC | Merge pull request #1609 from ColCarroll/make_scripts_executable Adds executable mode to `start_container.sh` | 17 December 2016, 18:06:39 UTC |
466b38b | colin | 16 December 2016, 21:08:04 UTC | Adds executable mode to `start_container.sh` | 16 December 2016, 21:08:04 UTC |
628cb52 | Colin | 16 December 2016, 20:51:45 UTC | Merge pull request #1608 from tyarkoni/fix-batches Dynamically set number of batches to appropriate value | 16 December 2016, 20:51:45 UTC |
c378850 | tyarkoni | 16 December 2016, 19:01:14 UTC | dynamically sets batches to min(n_samples, 100); closes #1606 | 16 December 2016, 19:01:14 UTC |
7584b3b | Maxim Kochurov | 13 December 2016, 19:42:54 UTC | Merge pull request #1599 from Spaak/master replacing mutable default value with None | 13 December 2016, 19:42:54 UTC |
cfa709a | Eelke Spaak | 13 December 2016, 16:31:53 UTC | replacing mutable default value with None | 13 December 2016, 16:31:53 UTC |
5be027f | Colin | 11 December 2016, 04:25:55 UTC | Merge pull request #1592 from ColCarroll/fix_travis_build Actually build on miniconda with python2.7 | 11 December 2016, 04:25:55 UTC |
7b23fb5 | colin | 11 December 2016, 03:53:36 UTC | Fix for py2/py3 | 11 December 2016, 03:53:36 UTC |
0150b75 | colin | 11 December 2016, 01:49:55 UTC | Fix python2 build errors | 11 December 2016, 01:49:55 UTC |
e7dc271 | colin | 11 December 2016, 00:40:17 UTC | Actually build on miniconda with python2.7 | 11 December 2016, 00:44:21 UTC |
60c7209 | Thomas Wiecki | 10 December 2016, 12:26:33 UTC | DOC Add list of software using and papers citing PyMC3. (#1588) * DOC Add list of software using and papers citing PyMC3. * DOC Add BEAT package. | 10 December 2016, 12:26:33 UTC |
ab51b6a | Thomas Wiecki | 09 December 2016, 07:34:02 UTC | Merge pull request #1586 from aloctavodia/master traceplot: auto-scale the max value of y | 09 December 2016, 07:34:02 UTC |
c8fdf88 | aloctavodia | 09 December 2016, 02:00:39 UTC | traceplot: auto-scale the max value of y | 09 December 2016, 02:00:39 UTC |
1a83358 | Thomas Wiecki | 08 December 2016, 17:05:16 UTC | TST Remove stray debug print. | 08 December 2016, 17:05:16 UTC |
9b380da | Thomas Wiecki | 08 December 2016, 16:59:51 UTC | Merge pull request #1582 from pymc-devs/fix_scaling Fix NUTS scaling bounds #1581 | 08 December 2016, 16:59:51 UTC |
74db63b | Thomas Wiecki | 08 December 2016, 16:16:28 UTC | TST Update Hamiltonian and NUTS samples. | 08 December 2016, 16:16:28 UTC |
84d3b06 | Thomas Wiecki | 08 December 2016, 11:57:33 UTC | BUG 362c8ee introduced a bug by switching the bounds of the scaling matrix. The bounds 1e-3 are also too stringent for some data types, like returns. This thus adjusts bounds to 1e-8 instead of 1e-3. | 08 December 2016, 11:57:33 UTC |
2d632de | Thomas Wiecki | 07 December 2016, 08:25:21 UTC | DOC Improve axes doc string in response to #1577. | 07 December 2016, 08:25:21 UTC |
784a773 | Thomas Wiecki | 06 December 2016, 13:57:20 UTC | DOC Add Probabilistic ML to summary. | 06 December 2016, 13:57:20 UTC |
2edbe6c | Colin | 03 December 2016, 18:29:35 UTC | Add mkl-service to build (#1572) | 03 December 2016, 18:29:35 UTC |
493d281 | Thomas Wiecki | 02 December 2016, 06:35:57 UTC | DOC Add baseball model. | 02 December 2016, 06:35:57 UTC |
67c1cd0 | Colin | 02 December 2016, 06:32:21 UTC | Add test to warn for algorithmic changes (#1570) | 02 December 2016, 06:32:21 UTC |
bdd1dc3 | Chris Fonnesbeck | 01 December 2016, 22:45:24 UTC | Incremented release version to 3.0.4c4 | 01 December 2016, 22:45:24 UTC |
bbb4d57 | Chris Fonnesbeck | 01 December 2016, 08:06:27 UTC | Tweaked the BEST notebook (#1568) | 01 December 2016, 08:06:27 UTC |
010de60 | Chris Fonnesbeck | 01 December 2016, 02:06:34 UTC | Merge pull request #1546 from usptact/master Adding hierarhical partial pooling example | 01 December 2016, 02:06:34 UTC |
aab8625 | usptact | 01 December 2016, 00:09:01 UTC | Adding v2 of the hierarchical partial pooling notebook credits: Chris Fonnesback | 01 December 2016, 00:09:01 UTC |
a54a53e | Osvaldo Martin | 30 November 2016, 20:18:14 UTC | Kde boundaries (#1567) * use a fft-based kde instead of scipy.gaussian_kde * fix typo | 30 November 2016, 20:18:14 UTC |
68e53eb | Yves Delley | 30 November 2016, 10:15:45 UTC | BUG placed context stack inside thread-local data space (#1555) * placed context stack inside thread-local data space * fixed thread-local context manager, and added a regression test * fixed docstring of #1552 regression test | 30 November 2016, 10:15:45 UTC |
8a6d87e | Thomas Wiecki | 29 November 2016, 16:30:01 UTC | Revert "ENH Add burn and thin kwargs to sample." (#1564) | 29 November 2016, 16:30:01 UTC |
1c9adc6 | Thomas Wiecki | 29 November 2016, 10:30:52 UTC | MAINT Re-enable trust_input. | 29 November 2016, 10:30:52 UTC |
a7410c2 | notoraptor | 29 November 2016, 10:29:33 UTC | Fix the TypeError issue #994 on Windows. (#1550) * Suggestion to fix the TypeError issue #994 on Windows. Errors occurs because some ndarrays with dtype=int32 are passed to the function delta_logp which expects ndarrays with dtype=int64 as inputs. Inputs are casted with `ndarray.astype(int)`, but on Windows, the Python 'int' type is treated as a 32-bit integer by NumPy. I suggest to remove the line with `f.trust_input = True` to ensure that Theano always checks the inputs and casts it if necessary and when possible. This fix resolves the issue on my computer: ``` Python version: 2.7.12 |Continuum Analytics, Inc.| (default, Jun 29 2016, 11:07:13) [MSC v.1500 64 bit (AMD64)] OS version: Windows-8.1-6.3.9600 ``` * Discrete super-class raises TypeError if dtype != 'int64'. Arguments for delta_logp in Metropolis.astep() are now casted to 'int64' when Metropolis handles a discrete model. | 29 November 2016, 10:29:33 UTC |
f61d8cd | Thomas Wiecki | 29 November 2016, 10:09:32 UTC | ENH Add burn and thin kwargs to sample. (#1562) * ENH Add burn and thin kwargs to sample. * MAINT Move new thin and burn behind start kwarg. Make test use kwargs. | 29 November 2016, 10:09:32 UTC |
2caa005 | Yves Delley | 28 November 2016, 21:46:40 UTC | Fix #895: Bug pickling model instances (#1560) * added regression test for #895 * fixes #895 | 28 November 2016, 21:46:40 UTC |
c0cc253 | Thomas Wiecki | 28 November 2016, 04:36:13 UTC | BUG Convert mean and mode to tensors before applying theano indexing. | 28 November 2016, 04:36:13 UTC |
8a59d87 | Chris Fonnesbeck | 28 November 2016, 04:20:11 UTC | DOC tidy notebooks (#1535) * Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks WIP: Notebooks updated rugby analytics * Removed discrete_find_MAP.ipynb * Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks WIP: Notebooks updated rugby analytics * Cleanup of several notebooks * Additional cleanup of notebooks * Cleanup of additional notebooks * More cleaned up notebooks (including divergent LKJ) * Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks Removed discrete_find_MAP.ipynb Cleanup of several notebooks Additional cleanup of notebooks Cleanup of additional notebooks WIP: Notebooks updated rugby analytics * Cleanup of several notebooks * More cleaned up notebooks (including divergent LKJ) * DOC Rename pmf example. * DOC Remove glm-logistic as it timeouts on reading data. * DOC Revert LKJ model use Metropolis sampler as NUTS runs into non-singular errors. * DOC Remove NUTS scaling NB as it stops running and is mostly obsolete now that we auto-init in this manner. * DOC Change some samplers in model selection NB to make it faster. * DOC Change getting started guide to use auto-init and auto-assigned samplers. * DOC Remove link to paper as it's out-of-date. | 28 November 2016, 04:20:11 UTC |
9543afc | Mehmet Ozan Kabak | 28 November 2016, 04:19:01 UTC | Fixed bug in WAIC and LOO computation, also switched to SciPy's (#1557) logsumexp function for better numerical stability. | 28 November 2016, 04:19:01 UTC |
50d0e02 | Yves Delley | 26 November 2016, 16:45:10 UTC | fix computation of test-values when distribution parameters are given by a theano TensorConstant (#1554) | 26 November 2016, 16:45:10 UTC |
8db9069 | Vladislavs Dovgalecs | 26 November 2016, 03:30:03 UTC | Adding notebook on Hierarchical Partial Pooling | 26 November 2016, 03:30:03 UTC |
0bdff96 | Thomas Wiecki | 25 November 2016, 15:26:35 UTC | Do not compute gradient in find_MAP() if it's not required. (#1551) * ENH Do not compute gradient in find_MAP() if it's not required. Closes #639. * MAINT Remove unused Poisson import. * BUG Add missing allfinite() call. Refactor model to models.py * TST Missed import. | 25 November 2016, 15:26:35 UTC |
818e4cd | Thomas Wiecki | 24 November 2016, 13:56:14 UTC | Merge pull request #1548 from Spaak/master some small changes to make advi respect theano.config.floatX | 24 November 2016, 13:56:14 UTC |
39c3281 | Eelke Spaak | 24 November 2016, 11:57:29 UTC | some small changes to make advi respect theano.config.floatX and improve GPU compatibility | 24 November 2016, 11:57:29 UTC |
df93e9d | Thomas Wiecki | 23 November 2016, 15:58:23 UTC | MAINT Forward *kwargs to init_nuts to NUTS. | 23 November 2016, 16:01:39 UTC |
7cc7668 | usptact | 22 November 2016, 21:22:27 UTC | Adding hierarhical partial pooling example Credits: http://mc-stan.org/documentation/case-studies/pool-binary-trials.html | 22 November 2016, 21:22:27 UTC |
808d7dd | griosd | 22 November 2016, 20:32:44 UTC | Update text.py (#1545) | 22 November 2016, 20:32:44 UTC |
57b6f13 | Chris Fonnesbeck | 22 November 2016, 18:55:31 UTC | Merge pull request #1543 from pymc-devs/reduce_advi_n_init MAINT Reduce number of n_init advi steps. Closes #1540. | 22 November 2016, 18:55:31 UTC |
47dd7d7 | Thomas Wiecki | 21 November 2016, 08:34:18 UTC | MAINT Reduce number of n_init advi steps. Closes #1540. | 21 November 2016, 08:34:18 UTC |
27f663d | Thomas Wiecki | 21 November 2016, 08:25:44 UTC | Merge pull request #1542 from ColCarroll/scientific_notation ADVI uses scientific notation for greater than 5 digits. | 21 November 2016, 08:25:44 UTC |
e672ee9 | colin | 21 November 2016, 03:48:24 UTC | ADVI uses scientific notation for greater than 5 digits. | 21 November 2016, 03:48:24 UTC |
ff28eff | Chris Fonnesbeck | 20 November 2016, 22:28:07 UTC | Removed tutorial.ipynb -- seems redundant in the presence of getting_started.ipynb | 20 November 2016, 22:28:07 UTC |
00eb604 | Austin Rochford | 20 November 2016, 20:00:33 UTC | Merge pull request #1539 from AustinRochford/example-dp-marginal-mixture Use marginalized mixture classes for DP example | 20 November 2016, 20:00:33 UTC |
2e4e842 | AustinRochford | 20 November 2016, 19:23:57 UTC | Use marginalized mixture classes for DP example | 20 November 2016, 19:23:57 UTC |
296ed60 | Austin Rochford | 20 November 2016, 18:29:55 UTC | Merge pull request #1538 from springcoil/docker DOCS: Clarified the developing in docker part of CONTRIBUTING.md | 20 November 2016, 18:29:55 UTC |
806ad51 | springcoil | 20 November 2016, 18:10:39 UTC | DOCS: Clarified the developing in docker part of Contributing | 20 November 2016, 18:10:39 UTC |
08556d7 | Thomas Wiecki | 20 November 2016, 14:19:22 UTC | Merge pull request #1537 from ColCarroll/rm_hierarchical Just remove hierarchical test | 20 November 2016, 14:19:22 UTC |
83a638f | colin | 19 November 2016, 17:40:42 UTC | Just remove hierarchical test | 19 November 2016, 17:40:42 UTC |
b233972 | Thomas Wiecki | 19 November 2016, 14:34:20 UTC | DOC Add math to docs. Fix other doc string problems. | 19 November 2016, 14:34:32 UTC |
fffd959 | Chris Fonnesbeck | 18 November 2016, 16:29:21 UTC | Merge pull request #1495 from pymc-devs/constant Renamed ConstantDist to Constant | 18 November 2016, 16:29:21 UTC |
976393b | Thomas Wiecki | 18 November 2016, 15:43:02 UTC | Merge pull request #1533 from AustinRochford/advi-init-remove-sample_vp-progress Remove sample_vp progressbar for ADVI initialization | 18 November 2016, 15:43:02 UTC |
a70689e | Chris Fonnesbeck | 18 November 2016, 15:27:55 UTC | Merge branch 'constant' of github.com:pymc-devs/pymc3 into constant | 18 November 2016, 15:27:55 UTC |
bd80b46 | Chris Fonnesbeck | 15 November 2016, 18:56:51 UTC | Remove Constant step method test | 18 November 2016, 15:27:40 UTC |
93bd8e3 | Chris Fonnesbeck | 15 November 2016, 18:55:18 UTC | Removed Constant step method | 18 November 2016, 15:27:40 UTC |
45720d2 | Chris Fonnesbeck | 03 November 2016, 03:19:14 UTC | Renamed ConstantDist to Constant | 18 November 2016, 15:27:40 UTC |
71be307 | AustinRochford | 18 November 2016, 14:57:59 UTC | Remove sample_vp progressbar for ADVI initialization | 18 November 2016, 14:57:59 UTC |
d57000d | Thomas Wiecki | 18 November 2016, 14:11:30 UTC | DOC Add None option for NUTS intialization. | 18 November 2016, 14:11:30 UTC |
dbf0654 | Thomas Wiecki | 18 November 2016, 13:59:18 UTC | ENH Add ADVI initializing for continuous models sampled using NUTS. | 18 November 2016, 13:59:18 UTC |
ad7dc8a | Chris Fonnesbeck | 18 November 2016, 10:21:34 UTC | Replaced fixed kappa with passed value (#1531) | 18 November 2016, 10:21:34 UTC |
59f500c | Thomas Wiecki | 18 November 2016, 08:33:20 UTC | TST Fix Rhat bad test by changing seed and starting point. | 18 November 2016, 08:33:20 UTC |