https://github.com/GPflow/GPflow

sort by:
Revision Author Date Message Commit Date
d25eb15 fix vector handling and broadcasting in multivariate_normal density (#627) * fix vector handling and broadcasting in multivariate_normal density * change GPR to sum over likelihood values * add docstrings 24 January 2018, 19:32:04 UTC
2742332 Merge pull request #622 from GPflow/awav/documentation-update Update docstrings for core stuff. 23 January 2018, 17:25:18 UTC
71f7bef Merge branch 'master' into awav/documentation-update 23 January 2018, 16:59:12 UTC
8b7f36e Merge pull request #632 from GPflow/vincent/pickling Vincent/pickling 23 January 2018, 16:38:57 UTC
6d7c0f9 Merge branch 'master' into vincent/pickling 23 January 2018, 16:13:30 UTC
e23df44 Minor fix to allow unpickling models 23 January 2018, 16:12:11 UTC
9b13309 Update README.md to contain reference to gpflow-monitor (#623) 22 January 2018, 17:30:47 UTC
f2c8127 Merge branch 'master' into awav/documentation-update 19 January 2018, 18:21:05 UTC
ab0e2ab Merge pull request #621 from GPflow/awav/expectation-tests-clearance Expectation test clearance along with new test features. 17 January 2018, 22:33:55 UTC
cdf299a Remove unnecessary comments in Expectation tests. 16 January 2018, 22:31:08 UTC
db0f421 Merge branch 'master' into awav/expectation-tests-clearance 16 January 2018, 16:47:29 UTC
5f2a6d4 Gradient of `sqrt(RBF(Z, Z))` no longer NaN (#615) * Gradient of `sqrt(RBF(Z, Z))` no longer NaN Before this commit, `Kmms = sqrt(exp(-1/x)) = exp(-1/(2x))`. These two expressions are equivalent mathematically. However, they are not equivalent when it comes to computing finite-precision gradients (dKmms/dx) via backpropagation when `Kmms**2 == 0.0`. Thus, we change sqrt(exp(-1/x)) to exp(-1/(2x)). This solves #595 in the expectations framework. * Removed test for NaN gradients of eKxzKz 16 January 2018, 12:01:37 UTC
b9d40db Merge branch 'master' into awav/documentation-update 15 January 2018, 14:54:44 UTC
22272ea Clarify read index method. 15 January 2018, 14:51:11 UTC
33083a8 Merge branch 'master' into awav/expectation-tests-clearance 15 January 2018, 14:07:56 UTC
8e9a681 Merge branch 'master' of github.com:GPflow/GPflow 10 January 2018, 11:02:04 UTC
a75f24a improvements and clarifications to the handling of num_latent (#609) * improvements and clarifications to the handling of num_latent. Also fizes the coregion notebook * removing leftover debug lines * adding a regression test for the wrong number of latent functions in a switched likelihood * Pass num_latent to GPModel in SGPMC * Pass num_latent to GPModel in GPMC * Pass num_latent to GPModel in SVGP * Removed ignore notebooks argument in travis config * Let codecov ignore the doc notebooks 08 January 2018, 13:21:07 UTC
a0c0227 Merge pull request #612 from GPflow/qsqrt-order `q_sqrt` is now D x M x M. 06 January 2018, 17:29:18 UTC
2bfbaee Convert numbers to tensors explicitly. 06 January 2018, 16:13:54 UTC
5d63ecd Add auxilary functions at test_util. Untie expectation tests from single graph usage. 06 January 2018, 15:55:45 UTC
4dc651f `q_sqrt` is now D x M x M. 05 January 2018, 15:30:16 UTC
a1d1516 improvements and clarifications to the handling of num_latent. Also fizes the coregion notebook 04 January 2018, 16:28:15 UTC
1342c28 Merge pull request #607 from GPflow/awav/awav-backward-tensors Missing part in transform - backward_tensor 04 January 2018, 11:15:29 UTC
b84c44e Merge branch 'master' into awav/awav-backward-tensors 04 January 2018, 11:02:09 UTC
4827068 Expectations framework (#591) * Outline of new expectations code. * Quadrature code now uses TensorFlow shape inference. * General expectations work. * Expectations RBF kern, not tested * Add Identity mean function * General unittests for Expectations * Add multipledispatch package to travis * Update tests_expectations * Expectations of mean functions * Mean function uncertain conditional * Uncertain conditional with mean_function. Tested. * Support for Add and Prod kernels and quadrature fallback decorator * Refactor expectations unittests * Psi stats Linear kernel * Split expectations in different files * Expectation Linear kernel and Linear mean function * Remove None's from expectations api * Removed old ekernels framework * Add multipledispatch to setup file * Work on PR feedback, not finished * Addressed PR feedback * Support for pairwise xKxz * Enable expectations unittests * Renamed `TimeseriesGaussian` to `MarkovGaussian` and added tests. * Rename some variable, plus note for later test of <x Kxz>_q. * Update conditionals.py Add comment * Change order of inputs to (feat, kern) * Stef/expectations (#601) * adding gaussmarkov quad * don't override the markvogaussian in the quadrature * can't test * adding external test * quadrature code done and works for MarkovGauss * MarkovGaussian with quad implemented. All tests pass * Shape comments. * Removed superfluous autoflow functions for kernel expectations * Update kernels.py * Update quadrature.py 03 January 2018, 17:44:53 UTC
2182bf0 Merge pull request #604 from GPflow/awav/notebook-tests Notebook test clearance. 03 January 2018, 15:41:39 UTC
72a5d55 Small fix in transform tests. 02 January 2018, 19:59:37 UTC
cef4917 Fix typo in trasform tests. 02 January 2018, 19:45:50 UTC
e271730 Travis artful distribution is not available. 02 January 2018, 19:44:06 UTC
c1d9de3 Transform backward tensors. 02 January 2018, 19:40:39 UTC
8a80103 Add backward methods for LowerTriangular and Diagonal. 02 January 2018, 19:07:09 UTC
968e43b Add backward tensors. 01 January 2018, 21:06:48 UTC
1d458c3 Docstrings for core/parentable.py 31 December 2017, 14:59:35 UTC
09d4b14 Docstrings for core/node.py module. 31 December 2017, 13:19:06 UTC
35bee19 Add ignore for vscode folder. 29 December 2017, 20:25:29 UTC
786d0ab Add ignore for vscode folder. 29 December 2017, 20:23:32 UTC
69893a7 Docs for compilable.py 29 December 2017, 20:21:42 UTC
7ba6ece 'artful' distribution is not supported by Travis CI. 28 December 2017, 18:53:51 UTC
9a87bd7 Resolve conflicts left by previous merges. 27 December 2017, 12:33:56 UTC
e54e862 Simplify notebook tests. 27 December 2017, 12:10:31 UTC
02d5d09 Add documentation. First attempts clarify things in compilable interface. 17 December 2017, 17:01:52 UTC
140a53c Merge pull request #592 from GPflow/awav/pandas-pretty-printing Pandas Model/Parameters printing. 14 December 2017, 09:49:06 UTC
d0937b2 Merge branch 'master' into awav/pandas-pretty-printing 13 December 2017, 20:38:23 UTC
3c78e4f Merge pull request #596 from GPflow/remove_likelihood_checks removed _check_targets 11 December 2017, 01:28:40 UTC
01d0a7a Update notebooks from master branch. 10 December 2017, 23:55:21 UTC
b6ff40e removed _check_targets 10 December 2017, 22:16:11 UTC
158af0e Re-run notebooks with new printing. 08 December 2017, 22:39:28 UTC
186419c Cover pandas printing for Parameter, DataHolder and Parameterized instances. 08 December 2017, 22:08:53 UTC
6582f41 Merge pull request #586 from GPflow/awav/replace-tf-float-and-int Standardize float type 08 December 2017, 02:39:53 UTC
cfa59ce Merge branch 'master' into awav/replace-tf-float-and-int 08 December 2017, 02:29:21 UTC
e67b457 Tests for deprecation messages. 07 December 2017, 21:41:14 UTC
81c4b55 Use new pandas tables in notebooks. 07 December 2017, 20:28:56 UTC
e923230 Merge pull request #589 from GPflow/awav/documentation Documentation fixes. 07 December 2017, 16:55:09 UTC
22f450d Merge branch 'master' into awav/documentation 07 December 2017, 00:00:17 UTC
0ef8e3d Remove redundant destination string for comparison. 06 December 2017, 23:28:29 UTC
f413633 Parameter representation returns str type. 06 December 2017, 23:17:04 UTC
e153901 Pretty printing using pandas structure. 06 December 2017, 16:49:34 UTC
8a1376b Merge master into branch. Resolve conflicts. 05 December 2017, 14:29:24 UTC
942fe22 Rename np_float[int] to float[int]_type. 05 December 2017, 00:19:41 UTC
f70e2dc Merge branch 'master' into awav/replace-tf-float-and-int 04 December 2017, 19:33:04 UTC
c3d5f28 Improve nomenclature (issue #504) (#570) * improve kernel nomenclature #504 * add deprecation errors for renames * give uncertain_conditional explicit kwargs as well * conditional(..., whiten=...) -> conditional(..., white=...) * updates to regression notebook 04 December 2017, 11:54:56 UTC
c951c3a Fixes for notebooks. Renaming feat => feature. 04 December 2017, 09:41:08 UTC
e5b8cfe Small spelling 02 December 2017, 10:34:11 UTC
79e0815 Clarify variables naming at assign Parameter's method. 01 December 2017, 10:54:37 UTC
399e732 Normalize tf floats and ints. 01 December 2017, 00:45:12 UTC
cf629d5 Merge pull request #581 from GPflow/readme-py3 Mention Python 3 in README.md 29 November 2017, 00:08:29 UTC
0656d41 Merge branch 'master' into readme-py3 28 November 2017, 11:07:17 UTC
5b60cb4 Update README.md 28 November 2017, 11:07:09 UTC
c98cd3c updates to regression notebook (#584) 28 November 2017, 10:17:56 UTC
0588a3c Merge pull request #583 from GPflow/awav/improve-uncertain-conditional-test Improve uncertain conditional test 28 November 2017, 08:38:05 UTC
27ef3ce Fix spelling error. 28 November 2017, 08:33:45 UTC
04d5d95 Merge branch 'master' into awav/improve-uncertain-conditional-test 27 November 2017, 22:30:14 UTC
52ecfab [#577] Improvement to improvement of tests in uncertain conditional test. 27 November 2017, 22:27:32 UTC
4a58b6e Mention Python 3 in README.md 27 November 2017, 19:54:47 UTC
d86af28 Merge pull request #573 from GPflow/chain_transform Chain transform helper 27 November 2017, 14:36:44 UTC
a79d97f reverting test edits 27 November 2017, 14:06:46 UTC
8f2f491 Merge branch 'master' into chain_transform 27 November 2017, 14:00:50 UTC
3df130f Merge pull request #580 from GPflow/awav/minibatch-iterator Exclusive behaviour for Minibatch 26 November 2017, 20:15:01 UTC
559c6de Merge branch 'master' into awav/minibatch-iterator 26 November 2017, 15:54:58 UTC
4770c34 Merge pull request #579 from GPflow/awav/kernel-quadrature-speedup-test Speed up some ekernels test on 40%. 26 November 2017, 15:46:17 UTC
25520f4 Fix test. Minibatch must not be re-initialized at each iteration, but keep reading and shuffling data, so that previous values are different. 26 November 2017, 12:54:21 UTC
706cb48 Exclude minibatch iterators from auto-initialization. 26 November 2017, 12:32:07 UTC
ece68fb Remove unnecessary 'session_context' from top of the 'test_exp_quadrature_diag' test. 26 November 2017, 12:08:26 UTC
75435e3 Speed up some ekernels test on 40%. 26 November 2017, 09:40:37 UTC
28070e7 10x speed up likelihood test. (#576) 25 November 2017, 12:52:36 UTC
30120a1 Address issues #568, #561 (#575) * Add a bunch of test for Parameter and DataHolder. Minibatch seed can be changed after cleaning or in defer_build. * Add dataholder tests. * Add failure creation tests to data holders. * Shape and dtype for dataholders work for external tensors. * According fix in dataholder test. * Change travis file. * Experiments with travis. * Improvements in Parameterized tests. * Exclude external optimizer from coverage report. * Improvements in optimizer tests. * Indent functions in optimizer's tests. * Increase coverage in Parameterized. * Increase codecov for Parameterized. * Developing test coverage further. * Tests Parameterized feeds. * Call parameter size in tests. * 100% for ParamList. * Minibatch tests. * Increase coverage for Minibatch. * Setup.py clean up. Stick to 1.4 tensorflow version. Remove backwards for py2.7. * Increase coverage for Parameters. * Increase coverage for Parameters build check. * Core tests for TensorConverter. * Code coverage targets. * Change initialization policy at decors and trainers. * Check initialization status of variable using tf.is_variable_initialized tensor. * [#568] Improve initialization status checking performance. * [#568, #561] Test for Dataset iterators is not possible. * Increase coverage for Parameter. method was added, therefore required some tests. method inconsistently handled session inputs. * Added more tests for Parameter object covering cases when tensor is trainable and parameter is not. Also added test for Parameter string conversion. 25 November 2017, 12:33:07 UTC
6211222 Improve docstring. 24 November 2017, 17:32:37 UTC
6a5e6a0 fix tests 24 November 2017, 16:32:13 UTC
1a886fc Merge branch 'chain_transform' of github.com:GPflow/GPflow into chain_transform 24 November 2017, 14:20:16 UTC
f4559a3 Some simplifications for DiagMatrix class 24 November 2017, 14:18:47 UTC
13e16a6 minor cleanup 24 November 2017, 11:57:14 UTC
b313409 adding hosekeeping for chaining transforms and simple test 24 November 2017, 11:10:22 UTC
cef3876 merge in test)_transforms 24 November 2017, 10:55:52 UTC
2cb1a4a add test 23 November 2017, 21:24:36 UTC
f42cab0 add positiveRescale transform combination 23 November 2017, 18:18:31 UTC
01c2aa3 Chain transform (#541) * Chain transform * fix abstract interface ITransform * doc changes * fix Rescale jacobian * enable chaining of transforms using __call__ * tests for transformation chaining * add tests for code coverage * add docstrings * add docstring to ITransform 23 November 2017, 18:10:46 UTC
a429265 add docstring to ITransform 23 November 2017, 17:45:49 UTC
17f736a add docstrings 23 November 2017, 17:25:46 UTC
8153ffa add tests for code coverage 23 November 2017, 16:46:50 UTC
a3aa873 Merge branch 'master' into chain_transform 23 November 2017, 16:23:34 UTC
back to top