* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [mzwiessele]
* Merge pull request #421 from SheffieldML/beckdaniel-wgps_improvements. [Max Zwiessele]
Beckdaniel wgps improvements
* Merge branch 'devel' into beckdaniel-wgps_improvements. [mzwiessele]
* [merge] devel. [mzwiessele]
* Revert "Revert "[kern] Add kernel was swallowing parts #fix#412"" [mzwiessele]
This reverts commit 0abb9b835ffeb020410bdf9a1e0532139ffa5cfc.
* Revert "[kern] Add kernel was swallowing parts #fix#412" [mzwiessele]
This reverts commit b8867f1552c05244dcd5ba38a7a57b6f1056312c.
* [kern] Add kernel was swallowing parts #fix#412. [mzwiessele]
* Merge branch 'devel' into beckdaniel-wgps_improvements. [mzwiessele]
* Merge branch 'wgps_improvements' of https://github.com/beckdaniel/GPy into beckdaniel-wgps_improvements. [mzwiessele]
* Improving coverage and removing py2 print. [beckdaniel]
* Merged last devel. [beckdaniel]
* Removed the check on f(y), it was only useful in logtanh. [beckdaniel]
* Added tests for closed inverse in identity and log. [beckdaniel]
* Changed imports to relative ones. [beckdaniel]
* Replicated the cubic sine example into warped_gp tests for code coverage. [beckdaniel]
* Removed logtanh, put into a new branch. [beckdaniel]
* Added an exception when you input 0 or negative values to logtanh function. [beckdaniel]
* Added a log-tanh function. [beckdaniel]
* Cleaning. [beckdaniel]
* Moved cubic sine from tests to examples. [beckdaniel]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into wgps_improvements. [beckdaniel]
Merging new devel
* Refactored the numeric inverse into the mother class, to test Identity and Log. [beckdaniel]
* Cleaning. [beckdaniel]
* Added a rate to inverse calculation. [beckdaniel]
* Logistic seems working but more tests are needed. [beckdaniel]
* First try on logistic function. [beckdaniel]
* Doctest on TanhFunction. [beckdaniel]
* Renamed TanhWarpingFunction to TanhFunction. [beckdaniel]
* Some cleaning on warped gp. [beckdaniel]
* Deleted old tanh_warp and renamed warp_tanh_d to warp_tanh. [beckdaniel]
* Cleaning on warping functions. [beckdaniel]
* Cleaning on warping functions. [beckdaniel]
* Added a log warping function. [Daniel Beck]
* Stuff. [beckdaniel]
* Skipping the wgps Snelson's test (comment the skip line to see the plots) [beckdaniel]
* (wpgs) fixing newton-raphson for f_inv and fixing plotting stuff. [beckdaniel]
* [priorizable] small edit to be usable with paramz 0.6.1 and greater. [mzwiessele]
* [priorizable] small edit to be usable with paramz 0.6.1 and greater. [mzwiessele]
* [appveyor] warning on existing whls. [mzwiessele]
* [plotting] small edit. [mzwiessele]
* [readme] added deploy status to readme. [mzwiessele]
* [Add] add kernel swallowed parts fix #412. [mzwiessele]
* Merge pull request #422 from SheffieldML/offset_and_clustering. [Mike Smith]
Offset and clustering: Utility to do clustering (greedy, pairs best clusters until likelihood stops increasing). Also includes a model that has an offset parameter to allow aligning of time series.
* Removing 'threaded' version. [Michael T Smith]
* Push just to rerun testing. [Michael T Smith]
* Don't use message added to cluster code. [Michael T Smith]
* Added threaded option - but this doesn't work due to the global interpreter lock. [Ubuntu]
* Made initial lengthscale!=1 to ensure we're properly testing gradients. [Michael T Smith]
* Modified set code in test to work with python 2 and python 3. [Michael T Smith]
* Corrected mistake in gradients: Was finding d(Xi-Xj)/dOffset instead of dr/dOffset. Fixed by scaling by kernel lengthscale. [Michael T Smith]
* Added GPy import. [Michael T Smith]
* More useful message from testing re offset estimate. [Michael T Smith]
* Corrected v2 missing print brackets. Added test code for new model and util. [Michael T Smith]
* Offset model and clustering utility. [Michael T Smith]
* Adding refs to new clustering and offset model and utility. [Michael T Smith]
* Merge pull request #420 from SheffieldML/deploy. [Max Zwiessele]
Deploy
* Merge pull request #419 from SheffieldML/devel. [Max Zwiessele]
fixed issues
* [merge] [mzwiessele]
* Merge pull request #418 from SheffieldML/plotting_tests. [Max Zwiessele]
Plotting tests and mean funcs
* [secret] coveralls only on travis. [mzwiessele]
* [prod] fix #388. [mzwiessele]
* [stationary] hint at confusing definition in GPy. [mzwiessele]
* [coverage] both coveralls and codecov, showing codecov. [mzwiessele]
* Coveralls token. [mzwiessele]
* Coveralls token in appveyor.yml. [mzwiessele]
* Revert "Revert "[coverage reports] change to coveralls as test"" [mzwiessele]
This reverts commit ee23da6dd9405120bec62402abf7aaa228a87a19.
* Revert "[coverage reports] change to coveralls as test" [mzwiessele]
This reverts commit 040ac72b82b6aa39720abe9817619103892b27a1.
* [coverage reports] change to coveralls as test. [mzwiessele]
* [plotting] last full upadate of baselinge. [mzwiessele]
* [plotting] more test updates and check for errors. [mzwiessele]
* [tests] all tests. [mzwiessele]
* [merge] devel and plotting tests from meanfunc. [mzwiessele]
* [plotting] updated for test skips. [mzwiessele]
* [mean_func] added parameters in additive mean func and tests for mean functions. [mzwiessele]
* Merge branch 'devel' into plotting_tests. [mzwiessele]
* [datasets] rnaseq changed up. [Max Zwiessele]
* [baseline] images adjusted and checked for testing including gplvm. [Max Zwiessele]
* [testsave] saved the testmodel for quicker and more robust plotting. [Max Zwiessele]
* Merge branch 'devel' into plotting_tests. [Max Zwiessele]
* [plotting] adjusting tests for quicker plotting. [Max Zwiessele]
* Bump version: 1.2.0 → 1.2.1. [mzwiessele]
* Update README.md. [Max Zwiessele]
* Merge pull request #415 from SheffieldML/devel. [Max Zwiessele]
deploy
* Bump version: 1.1.3 → 1.2.0. [mzwiessele]
* [travis] not tagged. [mzwiessele]
* Merge pull request #413 from SheffieldML/devel. [Max Zwiessele]
appveyor deploy
* Bump version: 1.1.2 → 1.1.3. [mzwiessele]
* [coverage] default to devel. [mzwiessele]
* Bump version: 1.1.1 → 1.1.2. [mzwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [mzwiessele]
* Merge branch 'deploy' into devel. [Max Zwiessele]
* Merge pull request #402 from SheffieldML/devel. [Max Zwiessele]
1.0.9 on deploy
* [codecov] default branch. [mzwiessele]
* [appveyor] twine? [mzwiessele]
* [appveyor] elsif ps. [mzwiessele]
* [appveyor] elsif ps. [mzwiessele]
* [appveyor] stop deploy on github. [mzwiessele]
* [appveyor] stop deploy on github. [mzwiessele]
* Bump version: 1.1.0 → 1.1.1. [mzwiessele]
* [appveyor] test deploy and full deploy. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [deploy] branch check. [mzwiessele]
* [version] for github releases? [mzwiessele]
* [secure] yes but not on prs. [mzwiessele]
* [secure] does it actually work? [mzwiessele]
* [secure] does it actually work? [mzwiessele]
* [secure] does it actually work? [mzwiessele]
* [secure] does it actually work? [mzwiessele]
* [appveyor] secure password 3. [mzwiessele]
* [appveyor] secure password 3. [mzwiessele]
* [appveyor] secure password 2. [mzwiessele]
* [appveyor] secure password 2. [mzwiessele]
* [appveyor] secure password. [mzwiessele]
* [appveyor] twine. [mzwiessele]
* [appveyor] echo empty line. [mzwiessele]
* [appveyor] pypirc. [mzwiessele]
* [appveyor] pypirc. [mzwiessele]
* [appveyor] pypirc. [mzwiessele]
* [appveyor] pypirc. [mzwiessele]
* [appveyor] twine upload. [mzwiessele]
* [setup] more setup changes. [mzwiessele]
* [setup] more setup changes. [mzwiessele]
* [setup] more setup changes. [mzwiessele]
* [appveyor] upload is not working? [mzwiessele]
* [appveyor] version in gh releases. [mzwiessele]
* [appveyor] version in gh releases. [mzwiessele]
* [appveyor] use twine. [mzwiessele]
* [appveyor] use twine. [mzwiessele]
* [appveyor] trying out github relesaessese. [mzwiessele]
* [appveyor] on_success. [mzwiessele]
* [tests] skip on climin import error. [mzwiessele]
* [appveyor] build script. [mzwiessele]
* [py33] removed from builds. [mzwiessele]
* [appveyor] tests. [mzwiessele]
* [appveyor] user. [mzwiessele]
* Bump version: 1.0.9 → 1.1.0. [mzwiessele]
* [secure] password added to appveyor. [mzwiessele]
* Merge pull request #408 from mikecroucher/devel. [Mike Croucher]
Automated Windows builds using Appveyor
* Appveyor: Py27 and Py35 builds. [Mike Croucher]
* Another appveyor bug fix. [Mike Croucher]
* Appveyor bug fix. [Mike Croucher]
* Appveyor: Use Miniconda for numpy and scipy. [Mike Croucher]
* Appveyor: Only build the deploy branch. [Mike Croucher]
* First attempt to use appveyor for windows builds. [Mike Croucher]
* Trying to be more specific. [Ricardo Andrade]
* Trying to be more specific. [Ricardo Andrade]
* Merge branch 'deploy' into devel. [Max Zwiessele]
* [py3] iterator .next fixes. [Max Zwiessele]
* [imports] fix #392. [Max Zwiessele]
* [#403] fix of inconsistent config naming. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge pull request #398 from SheffieldML/integral. [lionfish0]
Integral kernels added
* Modified testing to allow not implemented exceptions in update_gradients_full. [Michael T Smith]
* References integral kernel classes. [Michael T Smith]
* Merge remote-tracking branch 'origin/devel' into integral. [Michael T Smith]
* Merge pull request #397 from avehtari/dev-python3. [Alan Saul]
Python2->Python3
* Merge pull request #4 from alansaul/dev-python3. [Aki Vehtari]
Added test for subarray in util to dev-python3
* Added test for subarray in util. [Alan Saul]
* More Python 3 compatibility fixes. [Aki Vehtari]
* Python2->Python3. [Aki Vehtari]
* Removed references to integral kernels from __init__ [Michael T Smith]
* TEST: Rename parameters is test function. [Alexander Grigorievskiy]
* TEST: Tests use 'lbfgsb' optimization function. Also some syntactic changes. [Alexander Grigorievskiy]
* FIX: Some fixes which prevented tests passing on python3.5. [Alexander Grigorievskiy]
* FIX: SDE inference. Couple of bug fixes and minor syntactic madifications. [Alexander Grigorievskiy]
* FIX: Fixint the bug with matrix exponential computation. [Alexander Grigorievskiy]
* FIX: Get rid of unused imports in state_space_model file. [Alexander Grigorievskiy]
* ENH: Copying sde kernels to the '/src' directory. [Alexander Grigorievskiy]
* TEST: Modifying tests so that their ruunig time is shorter. [Alexander Grigorievskiy]
* FIX: Fixing the unit test which gave an error in Travis. [Alexander Grigorievskiy]
* BUG: change import from absolute to relative. [Alexander Grigorievskiy]
* UPD: Added testing, and bug fixing. [Alexander Grigorievskiy]
* UPD: Major update, changed interface of the module, Cython support added. Although cython gives almost no speed-up. [Alexander Grigorievskiy]
* UPD: Added SVD Kalman Filter, EM algorithm for gradient calculation (only for discrete KF) [Alexander Grigorievskiy]
* TEST: Remove test file which is incompatible with other tests in GPy. [Alexander Grigorievskiy]
* ENH: Adding tests for SDE kernels. [Alexander Grigorievskiy]
* ENH: Added SDE for all basic kernels except Rationale Quadratic. Some necessary modifications for the previous code are performed. [Alexander Grigorievskiy]
* ENH: Adding SDE representation of addition, sumation and standard periodic kernel. [Alexander Grigorievskiy]
All changes have been tested tests are added in later commits.
* ENH: Added templates for state-space kernels. [Alexander Grigorievskiy]
Those are childs of regular kernels with extra "sde" function.
* FIX: Fixe bug with "expm" function in "state_space_new". Also some minor changes. [Alexander Grigorievskiy]
Test function has been modified also.
* EXT: State-Space modelling functionality is untied with the GPy models. [Alexander Grigorievskiy]
Currently, these new functionality is added on the side, not intervening
the old state-space functionality. Example file has been changed and minimal
example where descripancies appear is cunstructed.
* Example of sde_Matern covarince function is added, along with other small changes. [Alexander Grigorievskiy]
State-space example is slightly modified.
Imports are corrected accordingly.
* Added summation of kernels under the state space formalism. [Arno Solin]
* Add the SDE for one kernel. [Arno Solin]
* Updated Kalman filter implementation to new GPy. [Arno Solin]
* Merge pull request #350 from SheffieldML/fixed_inputs. [Max Zwiessele]
Fixed inputs and BGPLVM prediction tests
* Small convenience function for extracting fixed_inputs, fixed inputs can be set to their mean, median, or zero. [Alan Saul]
* Test for BGPLVM predictions, for linear case which is possible to do analytically. [Alan Saul]
* Added back fixed_inputs. [Alan Saul]
* [coverage] dont complain if tests dont hit defensive assertion code. [Max Zwiessele]
* [plotting] was not covered in tests, now is. [Max Zwiessele]
* [kernel] added structural tests for ind outputs kernel, but problem with gradients persist. [Max Zwiessele]
* [plotting] still testing the testing. [Max Zwiessele]
* [plotting] skipping on fail. [Max Zwiessele]
* [plotting] back to png... [Max Zwiessele]
* [plotting] svg? [Max Zwiessele]
* [plotting] test for pdf? [Max Zwiessele]
* [plotting] tests. [Max Zwiessele]
* [plotting] tests. [Max Zwiessele]
* [Merge] merge devel. [Max Zwiessele]
* Merge pull request #339 from zhenwendai/devel_pullrequest. [Zhenwen Dai]
numerical stable implementation of rational qudratic
* Implement sensitivie of periodic kernel for plot_ARD. [Zhenwen Dai]
* Merge from Sheffield/GPy devel branch. [Zhenwen Dai]
* Merge pull request #314 from zhenwendai/devel_pullrequest. [Max Zwiessele]
Fix the issue of negative predicted variance of normal GP
* Implement the gradient_X for standard_periodic kernel. [Zhenwen Dai]
* Numerical stable implementation of rational qudratic. [Zhenwen Dai]
* Move _raw_predict into posterior object. [Zhenwen Dai]
* Bug fix for mcmc sampler and add test case. [Zhenwen Dai]
* Merge with upstream. [Zhenwen Dai]
* Merge pull request #324 from AlexGrig/std_periodic_kernel. [Max Zwiessele]
[kern] Standard periodic kernel. Changes parameter name from 'waveleng…
* [kern] Standard periodic kernel. Changes paramter name from 'wavelenght' to 'period'. This seems to be more clear. Also some minor modifications in the same file. [Alexander Grigorievskiy]
* Merge pull request #326 from SheffieldML/kern. [Max Zwiessele]
[kernel] fix #218 and #325
* [plotting] added plotting for missing data. [mzwiessele]
* [plotting] and ignoring it again. [mzwiessele]
* [plotting] adding plotting tests, due to many problems with plotting, when not checked. [mzwiessele]
* [util] tests for util/debug.py. [mzwiessele]
* [util] tests for util/debug.py. [mzwiessele]
* [kernel] fix #218 and #325. [mzwiessele]
* Merge pull request #323 from SheffieldML/stochastics. [Max Zwiessele]
[stochastics] update for new stochastic iptimizers in gpy
* [kern] added doc string. [mzwiessele]
* [autograd] added install instr for autograd. [mzwiessele]
* [sparse gp] commented out print statements, which are never used. [mzwiessele]
* [minibatch] added coverage for branching, spottet bug in X_variance. [mzwiessele]
* [climin] added tests and install directions for travis. [mzwiessele]
* [stochastics] added optimization for a few runs. [mzwiessele]
* [stochastics] update for new stochastic iptimizers in gpy. [mzwiessele]
* [white hetero] additional check for prediction. [mzwiessele]
* [white hetero] additional check for prediction. [mzwiessele]
* Merge pull request #322 from SheffieldML/minibatch. [Max Zwiessele]
[sparsegplvm] added sparsegplvm and tests for minibatch sparsegplvm
* Merge branch 'devel' into minibatch. [mzwiessele]
* Merge pull request #321 from SheffieldML/limit=3. [Max Zwiessele]
[chaching] changing all chacher limits to 3
* [chaching] changing all chacher limits to 3. [mzwiessele]
* Merge pull request #320 from SheffieldML/white_hetero. [Max Zwiessele]
Heteroscedastic White Kernel
* [white] added heteroscedastic white kernel for specific number of samples. [mzwiessele]
* [sparsegplvm] added sparsegplvm and tests for minibatch sparsegplvm. [mzwiessele]
* Merge pull request #318 from SheffieldML/gpy_one_fixes. [Max Zwiessele]
Gpy one fixes
* [plotting] info heatmap plotly. [Max Zwiessele]
* [deprecated] deprecated spelling mistake in wishart embedding. [Max Zwiessele]
* [kern] inner transformation to types, start for the multitype pandas arrays. [Max Zwiessele]
* Merge pull request #317 from SheffieldML/fix-276. [Max Zwiessele]
Last minute adjustements for plotly
* [plotly] fixes for mrd. [Max Zwiessele]
* [mrd] matplotlib specific fig_kwargs matplotlib specific. [Max Zwiessele]
* Merge pull request #316 from SheffieldML/fix-276. [Max Zwiessele]
Fix for #276
* [mrd] plot_scales and plot_latent added. [Max Zwiessele]
* Add ssgplvm model test. [Zhenwen Dai]
* Merge GPy upstream. [Zhenwen Dai]
* Fallback original slvm kl divergence. [Zhenwen Dai]
* Fix gpu initialziation. [Zhenwen Dai]
* Fix gpu initialziation. [Zhenwen Dai]
* Get rid of mpi4py import. [Zhenwen Dai]
* Merge remote-tracking branch 'upstream/devel' into devel. [Zhenwen Dai]
* Fix the issue of negative prediction variance of normal GP. [Zhenwen Dai]
* Slvm gamma mean-field. [Zhenwen Dai]
* Merge remote-tracking branch 'upstream/devel' into devel. [Zhenwen Dai]
* Implement slvm. [Zhenwen Dai]
* Add the import of transformation __fixed__ [Zhenwen Dai]
* [plotting util] faster generation of grid. [Max Zwiessele]
* Merge pull request #309 from SheffieldML/travis_scripts. [Max Zwiessele]
[travis_scripts] loading scripts from github repo
* [travis_scripts] loading scripts from github repo. [Max Zwiessele]
* [svgp] python 3.x fix for next. [Max Zwiessele]
* [latent plots] legend was always plotted. [Max Zwiessele]
* [latent plots] legend was always plotted. [Max Zwiessele]
* [latent plots] legend was always plotted. [Max Zwiessele]
* [latent plots] legend was always plotted. [Max Zwiessele]
* [latent plots] legend was always plotted. [Max Zwiessele]
* [travis] added retries for installing conda, as it seemed to fail on 404 errors. [Max Zwiessele]
* [Poly] added bias and scale. [Max Zwiessele]
* [Poly] added bias and scale. [Max Zwiessele]
* Bump version: 0.9.6 → 0.9.7. [Max Zwiessele]
* [gp_plots] transposed plotting of 2d contours. [Max Zwiessele]
* [matplotlib_dep] added the baseplots utility for backcompatibility. [Max Zwiessele]
* [matplotlib_dep] added the baseplots utility for backcompatibility. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* [plotting&kern] bugfixes in plotting and kernel size. [Max Zwiessele]
* Apidoc in conf. [Alan Saul]
* Added paramz to requirement file for docs. [Alan Saul]
* [plotly] scatter plotting was defaulting to color='white' [Max Zwiessele]
* [vardtc] these two lines are overridden by the next two lines... [Max Zwiessele]
* [plotting] catching 3d error for plotting latent space in other then 2 dimensions. [Max Zwiessele]
* Bump version: 0.9.4 → 0.9.5. [Max Zwiessele]
* [plotting] skipping plotting tests, as they are inconsistent across platforms. [Max Zwiessele]
* [plotting] baseline updates for 3d plotting. [Max Zwiessele]
* [plotting] was failing on some 3 dimensional plots (latent) [Max Zwiessele]
* Changed add_paraters to link_paramters. [cdguarnizo]
* Merge pull request #271 from zhenwendai/devel. [Zhenwen Dai]
Fix the dL_dK and dL_dpsi2 symmetric issue. The bug of the Independent_output kernel and Hierarchical kernel from the original implementation still remains.
* Mark the kernel test for independent kernel and hierarchical kernel as expectedFailure. [Zhenwen Dai]
* Merge with current GPy devel. [Zhenwen Dai]
* Bump version: 0.9.3 → 0.9.4. [Max Zwiessele]
* [examples] added bgplvm stochastic example and parameter for dimensionality. [Max Zwiessele]
* Bump version: 0.9.2 → 0.9.3. [mzwiessele]
* [plotting] latent plotting had dimension mix up in it. [mzwiessele]
* [plotting] latent plotting had dimension mix up in it. [mzwiessele]
* [plotting] latent plotting had dimension mix up in it. [mzwiessele]
* Bump version: 0.9.1 → 0.9.2. [mzwiessele]
* [plotting] latent plotting had dimension mix up in it. [mzwiessele]
* Bump version: 0.9.0 → 0.9.1. [mzwiessele]
* [readme] Reinstall update. [mzwiessele]
* [testing] testing the error messages for plotting. [mzwiessele]
* [testing] testing the error messages for plotting. [mzwiessele]
* [defaults.cfg] updated so we can ship it. [mzwiessele]
* [plotting] import updated so that the config is handled better. [mzwiessele]
* Bug fixed. [Ricardo Andrade]
ICM-RBF is used as default Kernel, but the user should be able to define a multiple output kernel outside and pass it to the model.
* Bug fixed. [Ricardo Andrade]
* Bump version: 0.8.31 → 0.9.0. [mzwiessele]
* Merge pull request #280 from SheffieldML/manifest-include. [Max Zwiessele]
Update MANIFEST.in
* Update MANIFEST.in. [Max Zwiessele]
* Update MANIFEST.in. [Max Zwiessele]
* Merge pull request #279 from SheffieldML/fixing_likelihoods. [Max Zwiessele]
Fixing likelihoods and EP
* Documentation. [Alan Saul]
* Merge branch 'devel' into fixing_likelihoods. [Alan Saul]
* Bump version: 0.8.30 → 0.8.31. [Max Zwiessele]
* [travis] register failes.. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Update .travis.yml. [Max Zwiessele]
* Bump version: 0.8.29 → 0.8.30. [mzwiessele]
* [travis] testing deployment code. [mzwiessele]
* [readme] added paramz update into readme. [Max Zwiessele]
* Removed old code. [Alan Saul]
* Allow EP to have a auto reset option. [Alan Saul]
* Merge branch 'devel' into fixing_likelihoods. [Alan Saul]
* Bump version: 0.8.28 → 0.8.29. [mzwiessele]
* [travis] testing deployment code. [mzwiessele]
* [travis] testing deployment code. [mzwiessele]
* Bump version: 0.8.27 → 0.8.28. [mzwiessele]
* [pypi. [mzwiessele]
* Bump version: 0.8.26 → 0.8.27. [mzwiessele]
* Bump version: 0.8.25 → 0.8.26. [mzwiessele]
* Bump version: 0.8.24 → 0.8.25. [mzwiessele]
* [coverage] updated coveragerc. [mzwiessele]
* Bump version: 0.8.23 → 0.8.24. [mzwiessele]
* [coverage] updated coveragerc. [mzwiessele]
* Bump version: 0.8.22 → 0.8.23. [mzwiessele]
* Bump version: 0.8.21 → 0.8.22. [mzwiessele]
* Merge. [Max Zwiessele]
* [optimization] deleted and backwardscompatible. [Max Zwiessele]
* [pickle] load errors bc of kernel changes, backwards compatibility issues fixed. [Max Zwiessele]
* Merge. [Max Zwiessele]
* Merge pull request #269 from SheffieldML/paramz. [Neil Lawrence]
Paramz
* Merge pull request #274 from SheffieldML/bumpversion. [Max Zwiessele]
Have pypi show the link to github
* Update __version__.py. [Max Zwiessele]
* Update setup.cfg. [Max Zwiessele]
* Have pypi show the link to github. [Max Zwiessele]
rendering of the rst file is way too strict to fuzz about with.
* Merge pull request #273 from SheffieldML/bumpversion. [Neil Lawrence]
Update setup.cfg
* Update __version__.py. [Max Zwiessele]
* Update setup.cfg. [Max Zwiessele]
* Bump version: 0.8.18 → 0.8.19. [mzwiessele]
* [paramz] changes in regexp. [mzwiessele]
* Bump version: 0.8.17 → 0.8.18. [mzwiessele]
* [description] was not being converted to rst. [mzwiessele]
* Bump version: 0.8.16 → 0.8.17. [mzwiessele]
* [description] was not being converted to rst. [mzwiessele]
* Bump version: 0.8.15 → 0.8.16. [mzwiessele]
* [description] was still in. [mzwiessele]
* Bump version: 0.8.14 → 0.8.15. [mzwiessele]
* Bump version: 0.8.13 → 0.8.14. [mzwiessele]
* [rst] pypandoc. [mzwiessele]
* Bump version: 0.8.12 → 0.8.13. [mzwiessele]
* [rst] pypandoc. [mzwiessele]
* Bump version: 0.8.11 → 0.8.12. [mzwiessele]
* [desc] no long desc. [mzwiessele]
* Bump version: 0.8.10 → 0.8.11. [mzwiessele]
* [plotting_tests] failed, because of same name? [mzwiessele]
* [plotting_tests] failed, because of same name? [mzwiessele]
* [plotting_tests] failed, because of same name? [mzwiessele]
* Bump version: 0.8.9 → 0.8.10. [mzwiessele]
* [doc] manifest includes tutorials. [mzwiessele]
* [paramz] updated. [mzwiessele]
* [MANIFEST] README added. [mzwiessele]
* [paramz] test file update. [mzwiessele]
* [paramz] test verbose error. [mzwiessele]
* [paramz] test warning catches. [mzwiessele]
* [merge] devel changes to regression objects. [mzwiessele]
* [paramz] right imports. [mzwiessele]
* [pickling] pickling causes seg fault, wtf? [mzwiessele]
* [core] import parameterization. [mzwiessele]
* Bump version: 0.8.8 → 0.8.9. [mzwiessele]
* [travis] using testpypi. [mzwiessele]
* [manifest] include right files. [mzwiessele]
* [unpickle] with latin1 encoding. [mzwiessele]
* [devel] pickling files... [mzwiessele]
* Merge branch 'devel' into paramz. [mzwiessele]
There was a conflict and I only had access to the web interface.
* Merge pull request #243 from SheffieldML/ep. [Max Zwiessele]
Ep inference updates
* [#186] fixed distribution across files and added base class for reusability. [Max Zwiessele]
* [inference] changed gaussian variance to precision (which it really is) [Max Zwiessele]
* [ep] now calling exact inference instead of copying code. [Max Zwiessele]
* [setup] include headers in source dist. [Max Zwiessele]
* [classification] sparse gp classification and dtc update. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Bug fix for set_XY. [Zhenwen Dai]
* Add set_XY test. [Zhenwen Dai]
## v0.8.8 (2015-09-10)
### Other
* Bump version: 0.8.7 → 0.8.8. [Max Zwiessele]
* Bump version: 0.8.6 → 0.8.7. [Max Zwiessele]
* [py3] print statement. [Max Zwiessele]
* [pred_var] added predictive variable as property now. [Max Zwiessele]
* Bump version: 0.8.5 → 0.8.6. [Max Zwiessele]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Max Zwiessele]
* Apply bug fix for set_XY. [Zhenwen Dai]
* Converting .md to .rst automatically now. [Max Zwiessele]
* Bump version: 0.8.4 → 0.8.5. [Max Zwiessele]
* Merge pull request #240 from SheffieldML/devel. [Max Zwiessele]
Devel
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'master' into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Updated sampling and plots to be correct shape, and changed plotting of sampling to be posterior samples p(f*|f), like it used to be, and samples_y to plot samples of p(y*|y) [Alan Saul]
* Add ARD to MLP kernel. [Zhenwen Dai]
* Merge pull request #239 from mikecroucher/master. [mikecroucher]
Cython fix
* Cython fix. [Mike Croucher]
* [setup] check if darwin. [Max Zwiessele]
* Update README.md. [Max Zwiessele]
* [verbosity] option in tests. [Max Zwiessele]
* [plotting] no printing of warning unless you actually use plotting. [Max Zwiessele]
* Bug fix: the name of parameterable object is not removed when unlinking. [Zhenwen Dai]
* Allow Y to be uncertain. [Zhenwen Dai]
* The bug fix for the cblas.h problem in Mac os x. [Zhenwen Dai]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mzwiessele]
* Merge the changes. [Fariba]
* Fix DGPLVM prior. [Zhenwen Dai]
* Improve the documentation of infer_newX. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Change crescent data to optimize with .optimize() [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Commit to pull. [Neil Lawrence]
* Fixed model test. [Alan Saul]
* Changed quantile computation via sampling and added fallback for predictive mean and variance if conditional mean and variance are not implemented yet. [Alan Saul]
* Passing metadata. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Generalized the variatinoal Gaussian approximatino revisited code for any likelihood. [James Hensman]
* Cache some values. [Alan Saul]
* Changes. [Fariba]
* [bgplvm] technical new stuff. [mzwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mzwiessele]
* Some correction for ibp ssgplvm. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Python 3 fixes. [Mike Croucher]
* Passing on y_metadata. [Alan Saul]
* Allowing set initial noise variance for GPRegression. [javiergonzalezh]
* Sqrt pi term was missing in variational expectations. [James Hensman]
* Lots of work on cython choleskies. [James Hensman]
* Fiddling with cholesky backprop. [James Hensman]
* Parallelizing backprop of cholesky. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Jacobians should not be computed only for transforms. [James Hensman]
* Jacobians should not be computed only for transforms. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Changed <> to != for Py3 compatibility. [Mike Croucher]
* Updated Py3 work. [Mike Croucher]
* Removed import urllib2 since it wasn't being used. [Mike Croucher]
* Urllib2 fixes for Py3 compatibility. [Mike Croucher]
* Fixed cPickle import for Python 3. [Mike Croucher]
* Exception raising fix for Python 3. [Mike Croucher]
* Put weave import in a try block so it fails gracefullt in Py3. [Mike Croucher]
* Merge remote-tracking branch 'upstream/devel' into devel. [Mike Croucher]
* Fixed ConfigParser for Python 3 compat. [Mike Croucher]
* Commented out cholupdate since it uses weave and appears not to be used. [Mike Croucher]
* Added Python 3 progress to README.md. [Mike Croucher]
* Resolved merge conflict. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Exception fixes for Python 3 compat. [Mike Croucher]
* Typo. [Mike Croucher]
* Fixed Python 2 compatibility. [Mike Croucher]
* Convert print to function for Python 3 compatibility. [Mike Croucher]
* Convert print to function for Python 3 compatibility. [Mike Croucher]
* Convert print to function for Python 3 compatibility. This breaks compatibility for versions of Python < 2.6. [Mike Croucher]
* Convert print to function for Python 3 compatibility. [Mike Croucher]
* Convert print to function for Python 3 compatibility. This breaks compatibility for versions of Python < 2.6. [Mike Croucher]
* Convert print to function for Python 3 compatibility. [Mike Croucher]
* Merge remote-tracking branch 'upstream/devel' into devel. [Mike Croucher]
* Added details of Python 3 work. [Mike Croucher]
* Convert print to function for Python 3 compatibility. This breaks compatibility for versions of Python < 2.6. [Mike Croucher]
* More relative import fixes for Python 3 compatibility. [Mike Croucher]
* Relative import fixes for Python 3 compatibility. [Mike Croucher]
* Updated README.md to refer to GPy/testing for running the tests. [Mike Croucher]
* Speed ups for normal cdf. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Added Y_metadata to log_predictive_density. [Alan Saul]
* [minor edits] [mzwiessele]
* [basis func kernels] added support for simple basis function kernels, can be easily extended by implementing phi function in BasisFuncKern. [mzwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mzwiessele]
* Small bugfix in white kernel. [James Hensman]
* Tests to probe the mean-function functionality. [James Hensman]
* Merge branch 'mean_functions' into devel. [James Hensman]
* Merges. [James Hensman]
* Whitespace. [James Hensman]
* Mean functions now working for svgp. with tests. [James Hensman]
* Fixed up product kernel tests. [James Hensman]
* Mappings, including tests. [James Hensman]
* Working mean function examples. [James Hensman]
* Merge branch 'devel' of github.com:/sheffieldml/GPy into mean_functions. [James Hensman]
* Merge branch 'devel' of github.com:/sheffieldml/GPy into mean_functions. [James Hensman]
* Merged. ish. [James Hensman]
* Merge pull request #174 from beckdaniel/kernel_prod_bugfix. [James Hensman]
Kernel product bugfix
* Test + code change in gradients_X. [Daniel Beck]
* A cleaner test. [Daniel Beck]
* Changed operator.mul to np.multiply for consistency. [Daniel Beck]
* Added decorator that changes numpy invalid op warning to exception. [Daniel Beck]
* First attempt. [Daniel Beck]
* Merge pull request #177 from mellorjc/master. [James Hensman]
matplotlib interactive mode only in IPython
* Catch only a specific error. [mellorjc]
catch only NameError, rather than everything.
* Matplotlib interactive mode only in IPython. [mellorjc]
have interactive mode only in IPython so that running scripts that plot from python behave like normal.
* A temporal fix for the problem of sometimes the model not being updated. [Zhenwen Dai]
* Add mcmc into inference import. [Zhenwen Dai]
* Bug in linalg jitchol!!! [Alan Saul]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolo Fusi]
* Merge branch 'devel' [Max Zwiessele]
* Implement update_gradients_diag for MLP kernel. [Zhenwen Dai]
* Re-implemented warpedGP for new release of GPy. [Nicolo Fusi]
* Fixed minor bug in sparse gp minibatch. [Nicolo Fusi]
* Minimual edits to exact_inference. [James Hensman]
* Added mean function into the prediction. [James Hensman]
* Added parseing of mean func to gp.py. [James Hensman]
* Mean functions in place. [James Hensman]
* Merge branch 'devel' of github.com:/sheffieldml/GPy into devel. [James Hensman]
* Optimize sslinear kernel. [Zhenwen Dai]
* Fallback the implementation of spike and slab prior. [Zhenwen Dai]
* Change the name of kernel DiffGenomeKern to DEtime. [Zhenwen Dai]
* More samples for predictive quantile. [Alan Saul]
* Merge branch 'saul_merge' into devel. [Alan Saul]
* Added safe_exp and tests. [Alan Saul]
* Small tidying up. [Alan Saul]
* Merge branch 'devel' into saul_merge. [Alan Saul]
* Merging with private repo, mostly fixed. [Alan Saul]
* Added block matrix dot product. [Alan Saul]
* Adding likelihoods and block matrices. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Relaxed inference test requirement. [Alan Saul]
* Adding a comment to clarify predictive_gradeints (Thanks AT) [James Hensman]
* Extra kernel stressing in benchmarks, bugfix in svgp. [James Hensman]
* Adding the beginnings of some benchmarks. [James Hensman]
* Shape changes for gradeitns of likelihood parameters in variational_expectations. [James Hensman]
* Added some clarifying comments with NDL. [James Hensman]
* Some tests for the svgp, and some changes to the likelihoods. [James Hensman]
* Derivatives of likelihood things now working for svgp. [James Hensman]
* Stupid bug. [James Hensman]
* Merge branch 'devel' of github.com:/sheffieldml/GPy into devel. [James Hensman]
* Lots of changes to mappings. [James Hensman]
* [var plots] wrong return values. [mzwiessele]
* [variational] plot needed kwargs. [mzwiessele]
* [verbose opt] ipython notebook new version widget changes. [mzwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mzwiessele]
* A little work on mappings. [James Hensman]
* [verbose opt] ipython notebook new version widget changes. [mzwiessele]
* [optimization] html widget api changes in ipython notebook? [mzwiessele]
* [sparse gp] doc changes for missing data. [mzwiessele]
* [ploting init] minor. [mzwiessele]
* [optimization] model prints how many parameters there are to optimize. [mzwiessele]
* Remove printing. [Zhenwen Dai]
* Fix the param renaming problem. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Version change (early beta, do not change until everythin works. [Max Zwiessele]
* Fixed the SCG optimizer, thanks to Yarin Gal. [James Hensman]
* Version now 48. [Max Zwiessele]
* Version update. [Max Zwiessele]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Max Zwiessele]
* Pickling now allways binary as well as protocol -1. [Max Zwiessele]
* Pickling now allways binary as well as protocol -1. [Max Zwiessele]
* Using lbfgs algorithm from scipy.minimize, starting to convert all optimizers to minimize format. [Max Zwiessele]
* Windows -.- [Max Zwiessele]
* Versions update. [Max Zwiessele]
* Image is a PIL requirement and should only be imported when actually using it. [Max Zwiessele]
* Plot handling greatly improved for latent space visualizations. [Max Zwiessele]
* Version file added. [Max Zwiessele]
* Fixed Ctrl-C behaviour on Windows. [Nicolo Fusi]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolo Fusi]
* Fixed come path issues in sympykern. [James Hensman]
* Rename and redoc. [Alan Saul]
* Moving imports, attempting to update RTD. [Alan Saul]
* Changed initalise_latent to take lower of init=PCA and corrected import. [Alan Saul]
* Change order of imports for RTD. [Alan Saul]
* Merge branch 'devel' [Alan Saul]
* Merge branch 'master' into devel. [Alan Saul]
* Removed variational.py. [Alan Saul]
* Fixed plot_latent failure. [Alan Saul]
* Ignore example tests. [Alan Saul]
* Removed yes pipe for travis. [Alan Saul]
* Seems to handle without answering now. [Alan Saul]
* Reverse travis to see what it asks for. [Alan Saul]
* Fixed some tests. [Alan Saul]
* Merged and fixed conflict in ODE_UY.py. [Alan Saul]
* Merge completed. [Max Zwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mu]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Andreas]
* Small changes in svigp. [Andreas]
* ODE UY dkdtheta. [mu]
* UY dkdtheta. [mu]
* UY dkdtheta. [mu]
* BGPLVM with missing data. [Max Zwiessele]
* Pca adjustements to lvm models. [Max Zwiessele]
* Ppca added, ppca missing data not working yet. [Max Zwiessele]
* Diagonal operations. [Max Zwiessele]
* Subarray indexing. [Max Zwiessele]
* Documenting. [Max Zwiessele]
* Bug in ODE_UY fix. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Fixed the numerical quadrature, won't work with large f unless normalized. [Alan Saul]
* Fixed lots of breaking tests, reduced step size for checkgrad to 1e-4 in tests (perhaps this should be global), added some missing attributes to data_resources.json. [Alan Saul]
* Set warnings for truncated hessian, it has been noted that that by truncating we can have incorrect posteriors, though at convergence this should not be a problem, could be fixed by not using Cholsky as the decomposition as it cannot handle non-positive definite mats. [Alan Saul]
* Merge pull request #90 from jamesmcm/master. [Alan Saul]
* Adding data_resources.json to setup data files. [James McMurray]
* Testing modification for ReadTheDocs to stop docstring errors. [James McMurray]
* Ensure_defaiult constraints in svigp. [James Hensman]
* Fixed Ctrl-C behaviour on Windows. [Nicolo Fusi]
* Removed print statements from config parser, commented out ODE kerns. [Nicolo Fusi]
* Merge branch 'devel' [Nicolo Fusi]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mu]
* Ode UY. [mu]
* Dk dparameter. [mu]
* ODE_UY. [mu]
* Merge branch 'devel' [Nicolo Fusi]
* Added some more error checking for downloading datasets. [Neil Lawrence]
* Fixed some bugs in mocap.py where errors weren't being raised when file format was incorrect and made datasets.py check for 404 errors which previously were occuring silently ... shhhhh. [Neil Lawrence]
* Fixed bugs in cmu_mocap loader where cmu_url was missing and loading in mocap data twice in same session led to incorrect url through copy error. [Neil Lawrence]
* Fixed examples tests, started changing datasets code which has a few bugs. [Alan Saul]
* Minor changes to naming of signitures. [Alan Saul]
* Changed more examples to accept optimize and plot. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Added comments for models module and adjusted setup. [Max Zwiessele]
* Merge branch 'naniny' into devel. [Max Zwiessele]
* Rename _models to models_modules to include in doc. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Some tidying in the regression examples. [James Hensman]
* Added optimize and plot for classification, non_gaussian and stochastic examples. [Alan Saul]
* Fixed gp_base and svigp for sampling (doesn't use it but needs the arguments) [Alan Saul]
* Added constant to Z_tilde, now log likelihoods are equal! [Alan Saul]
* Changed some parameters of the laplace, tidied up examples. [Alan Saul]
* Dimensionality reduction examples updated with optimize, plot and verbose. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Changing the seed seems to fix Alan's bug. [James Hensman]
* Fixed symmetry in checkgrad issue. [James Hensman]
* Reverted the brent optimisation in laplace. [James Hensman]
(For the 1D linesearch using Brent)
* Improved detectino of sympy. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Fixed exception handling bug in GPy/util/linalg.py:29. [Teo de Campos]
* Argghdfklg. [James Hensman]
* Better warings for cathcing of blaslib detection. [James Hensman]
* Changeing models to _models in setup.py. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Fixed step size for likelihood tests and allowed randomizing of laplace. [Alan Saul]
* Fixed student_t approximation demo and changed convergence critera to difference of f. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Added cfg file to manfiest and package_data. [Alan Saul]
* Dimensionality reduction example (oil) updated. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Removed ipdb statement from kern, cleaned up some nasty whitespace. [James Hensman]
* More readme. [James Hensman]
* More readme stuff. [James Hensman]
* More readme edits. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Minor edits to the README. [James Hensman]
* Gradient checker comments and import updates. [Max Zwiessele]
* Gradientchecker added as a model. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Push minor fix to eq_sympy kernel test. [Neil Lawrence]
* Added some tips to the readme. [James Hensman]
* Fixed import errors in tests. [Max Zwiessele]
* Sympykern kern_tests now passing, code is inefficient but should be numerically stable. [Neil Lawrence]
* Modified to improve part of stability, gradient checks still passing. [Neil Lawrence]
* Added gradient of sympy kernel, seems to pass tests, but know it's not numerically stable. Checking in before making numerically stable. [Neil Lawrence]
* Fixed test in kern.py to request correct output dim for multioutput covariances. [Neil Lawrence]
* ODE_UY. [mu]
* Fixing up the blas detectino in linalg. [James Hensman]
* Removing ipdb statements. [James Hensman]
* Lots of medding with the likelihoods to get the tests working. the tests still don;t work. [James Hensman]
* Formatting docstring. [James Hensman]
* Changed nasty whitespace. [James Hensman]
* Eq_ode1 working but test failing? [Neil Lawrence]
* Added eq_ode1 to constructors.py. [Neil Lawrence]
* Fixed bug in sympy kernel and added sympolic.py back into utils __init__.py. [Neil Lawrence]
* Merge with James's changes. [Neil Lawrence]
* Merge. [James Hensman]
* Removed some sympy stuff. [James Hensman]
* Merge with James's changes. [Neil Lawrence]
* Committing change for master check out. [Neil Lawrence]
* Skipping crossterm tests instead of expected failure. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Added a path for the data resources. not all users will be working in the GPy directory. [James Hensman]
* Moved data resource information to a json file. [Neil Lawrence]
* Bug fix for single output sympy kernel. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mu]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Fixed problem in warping. [Nicolo Fusi]
* Constructor and init for ODE_UY. [mu]
* Working eq_ode1 in sympy now. [Neil Lawrence]
* Dim reduction imports. [Max Zwiessele]
* Testing imports update and expected failure for crossterms. [Max Zwiessele]
* Rename models to _models and import models in models.py. [Max Zwiessele]
* Psistattests update. [Max Zwiessele]
* Updated crossterms, rbf x any not working yet (derivatives) [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Weird merge. [James Hensman]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Part implementation of ode_eq functionality. Not yet numerically stable or efficient (some horrible use of cut and paste to get things working ...) [Neil Lawrence]
* A trial namespace renaming. [James Hensman]
* Better handling of missing config files. [James Hensman]
* Debugging the config paths. [James Hensman]
* Allowing the passing of 1D X to a GP. with warning of course. [James Hensman]
* More fiddling with the windows path for config. [James Hensman]
Where is the windows guru? out playing beach volley?
* Changed how we search for config files on windows. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Fixed up symmetric kern. [James Hensman]
* Half way through crossterm objective. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Added block matrix utility. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Added **likelihood_params to predictive_values. [Ricardo]
* Changes in plot function: sampling vs numerical approximation. [Ricardo]
* Adding docstring for symmetric kern. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Bug fixed in numerical approx. to the predictive variance. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Numerical predictions fixed, sampling predictions are not working. [Ricardo]
* Predictive_mean and predictive_variance now use gp_var as a parameter, rather than gp_std. [Ricardo]
* Fixed product kern get and set state. [James Hensman]
* Added getstate/setstate for product kernel. [James Hensman]
* In the middle of crossterms. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* 2D plots fixed. [Ricardo]
* Passing **noise_args into predictive_values. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Missing term in the likelihood. [Ricardo]
* Reverted broken kern. [Max Zwiessele]
* Added variational distribution for latent space. [Max Zwiessele]
* BGPLVM test for crossterms. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Even more data plotting. [James Hensman]
* Fixed plotting isue with plot_f. [James Hensman]
* Fixed the dpotrs use.. [Alan Saul]
* Added dpotrs instead of cho_solve. [Alan Saul]
* Removed ipython dependency from kern. [Alan Saul]
* Sped up sampling a lot for student t, bernoulli and poisson, added sampling for gaussian and exponential (untested) [Alan Saul]
* Added sampling to student_t noise distribution, very slow and is possible to speed up. predictive mean analytical and variance need checking. [Alan Saul]
* Tidying up and fixed objective being vector. [Alan Saul]
* Added sampling for predictive quantiles and also mean and variance where necessary. [Alan Saul]
* Rederived gamma distribution. [Alan Saul]
* Added derivatives for poisson and a couple of examples, need to fix for EP. [Alan Saul]
* Merged with devel. [Alan Saul]
* Reimplemented gradients for exponential, seems to work for laplace now, needs a visual test though. [Alan Saul]
* Was a bug in the examples_tests.py, fixed and added brendan faces to ignore list. [Alan Saul]
* Updated boston tests (more folds, allow a bias as the datasets are not normalized once split) and more folds. Tweaked some laplace line search parameters, added basis tests for ep. [Alan Saul]
* Fixed bug in gradient checker where it worked differently given a integer parameter to a float. [Alan Saul]
* Removed derivatives of variance wrt gp and derivatives of means with respect to gp from noise models. [Alan Saul]
* Tore out code no longer used from noise_distributions due to rewriting using quadrature. [Alan Saul]
* Added numerical mean and variance with quadrature, about to clean up. [Alan Saul]
* Merge branch 'master' into merge_branch. [Alan Saul]
* Changed the gradients (perhaps for the worse) [Alan Saul]
* A few typos. [Alan Saul]
* Gaussian likelihood errors, still not working. [Alan Saul]
* Added gaussian checker and gaussian likelihood, not checkgrading yet. [Alan Saul]
* Started adding gaussian sanity checker. [Alan Saul]
* Got rid of some overdoing the approximation. [Alan Saul]
* Trying to debug kernel parameters learning (fails even when noise fixed) may be some instablility, seems like it can get it if it starts close. [Alan Saul]
* Playing trying to find what makes it want to go so low. [Alan Saul]
* Fixed bug where B wasn't refering to current f location. [Alan Saul]
* Everything seems to be gradchecking again. [Alan Saul]
* Added minimizer for finding f, doesn't help. [Alan Saul]
* Now checkgrads a lot more of the time, but still fails in optimisation, seems also odd that when parameter is fixed kernel parameters go to infinity. [Alan Saul]
* Added another optimisation which doesn't use gradients. Seems like F is almost always found, but Y can be off, suggesting that Wi__Ki_W is wrong, maybe W? [Alan Saul]
* Trying to fix optimisation problem, fixed a few bugs but still fails at very low noise. [Alan Saul]
* Starting to fiddle with mode finding code. [Alan Saul]
* Fixed a sign wrong, now gradchecks weirdly only above certain points. [Alan Saul]
* Now gradchecks everytime but student_t fit is bad, noise is underestimated by a long way. [Alan Saul]
* Checkgrads with explicit and implicit components half the time. [Alan Saul]
* About to input new derivations for Z's... [Alan Saul]
* Took out all the asserts and using pure broadcasting method of diagonal now. [Alan Saul]
* Made it use the fact that W is diagonal and put assertions in to ensure that the results are the same. [Alan Saul]
* Broken it by getting rid of squeeze, but now working on making it faster using proper vector multiplciation for diagonals. [Alan Saul]
* Made more numerically stable in a hope that it will work and I will find a bug... [Alan Saul]
* Lots of name changing and went through all likelihood gradients again. [Alan Saul]
* Ripped out all things Laplace parameter estimation, starting again with new tactic. [Alan Saul]
* About to rip out old chain rule method of learning gradients. [Alan Saul]
* Trying to fix dL_dytil gradient. [Alan Saul]
* Changed name. [Alan Saul]
* FIXED DYTIL_DFHAT. [Alan Saul]
* Workong on doing explicit gradients. [Alan Saul]
* Gradients almost there for dytil_dfhat, diagonal terms are right. [Alan Saul]
* Merged with devel. [Alan Saul]
* Still getting closer to grads for likelihood. [Alan Saul]
* Almost have likelihood gradients working but kernels still way off. [Alan Saul]
* Working on putting callback to update laplace in callback. [Alan Saul]
* Seem to have gradients much closer now. [Alan Saul]
* Scale and switch KW+I. [Alan Saul]
* Merged with upstream. [Alan Saul]
* Added a debug examples. [Alan Saul]
* Merging. [Alan Saul]
* Merged with master. [Alan Saul]
* Plotting problematic kernel. [Alan Saul]
* Adding gradients, shapes starting to make sense. [Alan Saul]
* Attempted to introduce gradient methods, won't work yet I doubt. [Alan Saul]
* Merge branch 'old_repo' into merge_trial. [Alan Saul]
* Make directory structure match that of GPy. [Alan Saul]
* Tidy up comments. [Alan Saul]
* Merged in the GPy upstream. [Alan Saul]
* Merged in branch which had old_repo merged in. [Alan Saul]
* Adding weibull likelihood, requires 'extra_data' to be passed to likelihood, i.e. the censoring information. [Alan Saul]
* Fixed the z scalings. [Alan Saul]
* Fixed laplace approximation and made more numerically stable with cholesky decompositions, and commented. [Alan Saul]
* Stabalised most of the algorithm (apart from the end inversion which is impossible) [Alan Saul]
* Added timing and realised mdot can be faster as its almost always a diagonal matrix its multiplying with. [Alan Saul]
* Got the mode finding without computing Ki. [Alan Saul]
* Fixed broadcasting bug, rasm now appears to work. [Alan Saul]
* Still working on rasmussen, link function needs vectorizing I think. [Alan Saul]
* Tidying up. [Alan Saul]
* Added predicted values for student t, works well. [Alan Saul]
* Working laplace, just needs predictive values. [Alan Saul]
* Seemed to be working, now its not. [Alan Saul]
* Changing definitions again... [Alan Saul]
* Worked out in terms of W, needs gradients implementing. [Alan Saul]
* Just breaking some things... [Alan Saul]
* Following naming convention better, lots of inverses which should be able to get rid of one or two, unsure if it works. [Alan Saul]
* Got an idea of how to implement! written in docs. [Alan Saul]
* Trying to 'debug' [Alan Saul]
* Got most of laplace approximation working. [Alan Saul]
* Added some comments. [Alan Saul]
* Initial commit, setting up the laplace approximation for a student t. [Alan Saul]
* Initial commit. [James Hensman]
* Use bfgs for laplace instead. [Alan Saul]
* Moved transf_data to make data -1 or 1 from 0 or 1 for bernoulli with probit into the analytical moment match (but it 10% slower), needs removing from epmixednoise. [Alan Saul]
* Changed naming from old derivatives of likelihoods to new ones in noise distributions. [Alan Saul]
* Fixed breakage of dvar, tidied up to make more efficient. [Alan Saul]
* Doc stringing. [Alan Saul]
* Added quadrature numerical moment matching (but not predictive yet) [Alan Saul]
* Fixed a few laplace bits. [Alan Saul]
* Refactored gradients wrt parameters slightly, need to future proof against _get_param_names() disappearing. [Alan Saul]
* Added more options to generic tests (constraining link function values as bernoulli requies R^{0,1}) and implemented new gradients for bernoulli. [Alan Saul]
* Rename Binomial to Bernoulli (maybe generalise it with the constant later, but tilted distribution may change) [Alan Saul]
* Added pdf_link's for gaussian and student t, added third derivatives for transformations and tests for them. [Alan Saul]
* Renamed laplace_tests to likelihoods_tests. [Alan Saul]
* Tidying up laplace_tests.py. [Alan Saul]
* Fixed some bugs, added third derivative for log transformation, and did some doccing. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Still tidying up, laplace now working again, gaussian and student_t likelihoods now done. [Alan Saul]
* Started on chaining, must remember to chain _laplace_gradients aswell! [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Finished tearing gaussian noise down, time for student t. [Alan Saul]
* Beginning to merge lik_functions and derivatives with richardos. [Alan Saul]
* Docs. [Alan Saul]
* Removed fit as it is unused. [Alan Saul]
* More doc strings. [Alan Saul]
* Doccing and testing for D dimensional input (not multiple dimensional Y yet) [Alan Saul]
* Tidying up a lot, works for 1D, need to check for more dimensions. [Alan Saul]
* Tidied up laplace. [Alan Saul]
* Integrated Laplace and merged Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
Conflicts:
GPy/core/gp.py
GPy/likelihoods/__init__.py
GPy/likelihoods/likelihood_functions.py
GPy/likelihoods/link_functions.py
* Fixed white variance. [Alan Saul]
* Boston housing works (apart from variance of student t is not valid below 2) [Alan Saul]
* Tests setup but not fitting properly yet. [Alan Saul]
* Changed the examples (started boston data) and increased tolerance of finding fhat. [Alan Saul]
* Added some stability and tidied up. [Alan Saul]
* Tidying up. [Alan Saul]
* Student t likelihood function checkgrads (summed gradients wrt to sigma2), maybe some numerical instability in laplace. [Alan Saul]
* Now checkgrads for gaussian, and ALMOST for student t. [Alan Saul]
* All gradients now gradcheck. [Alan Saul]
* Merge remote-tracking branch 'gpy_real/devel' into merge_branch. [Alan Saul]
* Refactored tests. [Alan Saul]
* Tidied up grad checking. [Alan Saul]
* Added tests and fixed some naming. [Alan Saul]
* Modified gradient_checker to allow for variable 'f' [Alan Saul]
* Renamed some things, made some small (incorrect) gradient changes, generalised the gp regression for any likelihood, and added a place holder link function waiting for Richardos changes. [Alan Saul]
* Removed unneeded dependency. [Alan Saul]
* Merged GP models. [Alan Saul]
* Dragged likelihood_function changes in. [Alan Saul]
* Checked out relavent files. [Alan Saul]
* Merged in real gpy. [Alan Saul]
* Empty branch. [Alan Saul]
* SPELLAFSDIUN. [Max Zwiessele]
* Fixed up plotting in sparse_gp also. [James Hensman]
* Fixed up the plotting. [James Hensman]
* Fixed up plot in GP_base. [James Hensman]
* Started changing the plotting in examples to remove plot_single_output. [James Hensman]
* General tidying in models. [James Hensman]
* Improved docstrings in svigp. [James Hensman]
* Some tidying in gp.py. [James Hensman]
* Docstrings and removal of duplicated plotting code in gp_base. [James Hensman]
* Turned omp off by default as discussed. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Added configuration file. [Nicolò Fusi]
this was done to solve the OpenMP problem on Windows/mac, but I think it
is useful in general. All unit tests pass except the sympy kern ones.
* Added xw_pen data. [Neil Lawrence]
* Added xw_pen data. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Basic sim code functional. [Neil Lawrence]
* Minor change in tutorial. [mu]
* Added link to user mailing list. [James Hensman]
## v0.6.0 (2014-11-21)
### Other
* Small random perturbations in kernel tests helps with the symmetry gradcheck bug. [James Hensman]
* More cooooooopyrighting. [James Hensman]
* More coooooopyrighting. [James Hensman]
* Removing old notes.py, issues are now all on github. [James Hensman]
* More cooooopyrighting. [James Hensman]
* More cooooopyrighting. [James Hensman]
* More coooopyrighting. [James Hensman]
* More cooopyrighting. [James Hensman]
* More coopyrighting. [James Hensman]
* More copyrighting. [James Hensman]
* More ]#copyrighting. [James Hensman]
* Branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* [kernel plots] updates on bar plots. [Max Zwiessele]
* Linear kernel speed up. [Zhenwen Dai]
* Performance improvement for sslinear kernel. [Zhenwen Dai]
* Copyrighting. [James Hensman]
* Working One vs All sparse gp classification wrapper. [Andreas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Added lengthscales for a standard GPLVM with ARD. [Andreas]
* Small changes to 1vsAll. [Andreas]
* New file, sparse one vs all classification. [Ricardo]
* Missing file. [Ricardo]
* One vs all classification. [Ricardo]
* [inferenceX] with missing data. [mzwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Add test cases for inference new X for bayesian GPLVM. [Zhenwen Dai]
* Bug fix for inferenceX. [Zhenwen Dai]
* Add new inference X. [Zhenwen Dai]
* [tests] pickle tests updated to new structure. [Max Zwiessele]
* [MRD] running again, using missing_data classes, more details needed for missing data though. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Refine the docstring for hmc. [Zhenwen Dai]
* Removed pod dependency for pickle tests. [Alan Saul]
* Tidied up laplace warnings. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Add __init__.py to mcmc. [Zhenwen Dai]
* Add documentation for hmc. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Changed init for mcmc. [James Hensman]
* Removing dead bayesopt file. [James Hensman]
* Catch nosetests import error if not installed, now ignore GPy.tests() when nosetests GPy is called, but allows GPy.tests() to be called, and throws error if this is tried without nose being installed. [Alan Saul]
* [normalizer] only mean, because variance could be not Gaussian... [mzwiessele]
* [normalize] [mzwiessele]
* [normalizer] first commit for normalizer in GPy. [mzwiessele]
Conflicts:
GPy/core/sparse_gp.py
GPy/models/bayesian_gplvm.py
* Further bug fix for sparsegp_mpi. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Noise parameters built depending on Y_metadata. [Ricardo]
* Generalize the interface of mpi. [Zhenwen Dai]
* Recover the ss_gplvm.py. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* [ard] enhanced ard handling and plotting. [mzwiessele]
Conflicts:
GPy/kern/_src/linear.py
GPy/models/ss_gplvm.py
* [linear] einsums. [mzwiessele]
* [pickling] wb as write parameter. [mzwiessele]
* Improve numerical stability of vardtc_parallel. [Zhenwen Dai]
* A bug fix for psi statistics related model pickle. [Zhenwen Dai]
* Update additive kernel for SSGPLVM. [Zhenwen Dai]
* Fix the pickle problem for models with psi statistics. [Zhenwen Dai]
* Some further performance improvement for linear kernel psi statistics. [Zhenwen Dai]
* Linear kernel psi statistics performance optimization. [Zhenwen Dai]
* Minor changes on SSGPLVM. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* [parameterization] Parameter adding more robust and better error handling. [mzwiessele]
* Added kronecker and variational gaussian approximation gp's, vargpapprox needs generalising to any factorizing likelihood. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Gradients of predictions for Trevor. [James Hensman]
* More bugfix. [James Hensman]
* Bugfix merge! [James Hensman]
* Strange bug in np.einsum fixed when using the _out_ argument (thanks T. Cohn) [James Hensman]
* Minor changes in ICM. [Ricardo]
* Fix the bug of caching w.r.t. ignore arguments. [Zhenwen Dai]
* Change for ssgplvm example. [Zhenwen Dai]
* Merge the current devel into psi2. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* [setup] no import of os, nanana alan :) [mzwiessele]
* [parameter core] offset for can be done without parameter slices. [mzwiessele]
* [gradcheck] some performance enhancement. [mzwiessele]
* [bgplvm] gradient settings. [mzwiessele]
* Merge branch 'hmc' into devel. [Zhenwen Dai]
A HMC sampler for GP parameters
* Debug HMC shortcut. [Zhenwen Dai]
* Hmc shortcut. [Zhenwen Dai]
* Correct the initial distribution of p. [Zhenwen Dai]
* Remove prints. [Zhenwen Dai]
* Debug hmc code. [Zhenwen Dai]
* Add hmc.py. [Zhenwen Dai]
* Initial implementation of hmc. [Niu]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* [minor] minor changes. [mzwiessele]
* Coregionalized 2D plotting fixed. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Bug fixed. [Ricardo]
* Fizxed a numerical bug in stationary.py. [James Hensman]
* Returned setup.py read to old version. [Alan Saul]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Changes to datasets.py. [Neil Lawrence]
* Added forced extraction of eggs (as we have a fair few non-py files and use the directory structure) added some files to MANIFEST and setup.py's package_data so its included upon distributing. [Alan Saul]
* [inference] less constant jitter, and jitter adjustements. [mzwiessele]
Conflicts:
GPy/util/linalg.py
* Fixed a bug in optimize restarts: it now used optimizer_array. [James Hensman]
* [optimizer array] bugfix, when updating the model the optimizer array would not update. [mzwiessele]
* [optmimize] bugfix. [mzwiessele]
* [whitespaces] & [opt] minor fix of optimizer, when Optimizer is provided (set model to self) [mzwiessele]
* [model] optimizer can now be an optimizer instance, instead of a string. [mzwiessele]
* [parallel vardtc] minor adjustments to work with current implementation of psi stats. [mzwiessele]
* [randomize] adjusted parameters to go into random generator right. [mzwiessele]
* [scg] minor adjustements based on original publication. [mzwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* [linalg] fixed scipy 0.14 bugfix. sciy.linalg.lapack.dpotri was fixed to work right with lower=1, thus, the hack is gone now from GPy.util.linalg.dpotri, when using scipy 0.14 and higher. [Max Zwiessele]
* Re-doing the lee dataset. [James Hensman]
* Revert "Minor edits to reading Lee data in datasets.py" [James Hensman]
This reverts commit 730e229238062fa22b726e8c30c891d0819b3c6e.
* Fixed unnecessary warnings when using periodic kernels. [durrande]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Datasets.py updates should have been committed before. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Reverting Maxs linalg changes. [James Hensman]
* Linalg had lowers missing for windows libraries to work correctly. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Merge branc( 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Editied whitespace. [James Hensman]
* [splitkern] support idx_p==0. [Zhenwen Dai]
* Minor changes. [Zhenwen Dai]
* DiffGenomeKern bug fix. [Zhenwen Dai]
* [splitkern] some more changes. [zhenwen]
* [splitkern] some additional implmentation. [Zhenwen Dai]
* [splitkern] bug fix. [Zhenwen Dai]
* [splitkern] buf fix. [Zhenwen Dai]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Zhenwen Dai]
* Merging. [James Hensman]
* Reverting the fixing behaviour. [James Hensman]
two reasons: 1) the new behaviour is confusing for new users. Either
something is fixed, or it's not. 2) the fixing didn't work! things that
should have been fixed were passed to the optimizer for optimization.
If we really want to save keystrokes, consider this:
m.foo.fix()
m.foo.unfix()
m.foo.constrain_positive()
is the same as
m.foo.fix()
m.foo.constrain_positive()
but the latter throws a warning.
* Merge the bug of fixing function. [Zhenwen Dai]
* Developing split kernel. [Zhenwen Dai]
* Developing split kernel. [Zhenwen Dai]
* Minor edits to reading Lee data in datasets.py. [Neil Lawrence]
* Modified Spellman et al data load. [Neil Lawrence]
* Changes to configuration file set up: now uses defaults.cfg, installation.cfg and searches locally for .gpy_user.cfg in the users home directory. [Neil Lawrence]
* Added CIFAR-10 data to data sets. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Fixed an assertion, it was checking the dimensionality of the input data matrix, rather than that of the labels. [teodecampos]
* [bgplvm] do test latents updated for devel branch. [mzwiessele]
* [copy] is now fully functional, only hierarchy observers will be copied and pickled. [mzwiessele]
* [active_dims] all kernels now have int arrays as active_dims. [Max Zwiessele]
* [combination kernel] some fixing with error messages. [Max Zwiessele]
* [caching] id fix. [Max Zwiessele]
* [caching] done right. [Max Zwiessele]
* Merge branch 'devel' of github.com:/SheffieldML/GPy into devel. [James Hensman]
* Added citation to readme. [James Hensman]
* Sparse GPs can now accept kerns for predicting. [James Hensman]
* Added polynomial kernel. [James Hensman]
* [SSGPLVM] linear kernel cpu ready. [Zhenwen Dai]
* [ssgplvm] linear kernel. [Zhenwen Dai]
* Merge branch 'psi2' of github.com:SheffieldML/GPy into psi2. [Zhenwen Dai]
* Proper whitespace. [James Hensman]
* [mpi] add mpi into ssgplvm. [Zhenwen Dai]
* [GPU] varDTC_gpu ready. [Zhenwen Dai]
* [GPU] [Zhenwen Dai]
* Switch psi2 statistics design. [Zhenwen Dai]
* Some hacking on image_show in viaualize. [James Hensman]
* Removed another import of non-added file. [James Hensman]
* Removed import of non-added file (Mu) [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* St. [mu]
* Merge branch 'params' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge branch 'params' of https://github.com/SheffieldML/GPy into params. [Neil Lawrence]
* Pre-devel-move check in. [Neil Lawrence]
* Added the ability for GPs to predict with a different kernel. [James Hensman]
* Some work on the hierarchical kern. [James Hensman]
* Grabbed readme from old devel branch. [James Hensman]
* Bugfix: slicing checks needed to be suspended for combination kernels, checks are done in inner kernels now. [mzwiessele]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [mzwiessele]
* Merged. [James Hensman]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Ricardo]
Conflicts:
GPy/kern/_src/kern.py
* Missing file, import line commented. [Ricardo]
* Double quote deleted. [Ricardo]
* Some horrible hacking on hierarchical. [James Hensman]
* Bugfix: slicing. [mzwiessele]
* Bugfix: kern input_sens errir. [mzwiessele]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [mzwiessele]
* Merge kernel source. [Neil Lawrence]
* Removed imports of files the are not added to the repo. [James Hensman]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [James Hensman]
* Removing Neils mid-atlantic commit because he failed to add the relevant files to the repo. [James Hensman]
* Added some missing files. [Neil Lawrence]
* Slicing: slicing now thorughs the right error. [mzwiessele]
* HACK: plot_ARD is additive, should think of making it selectable through parameter handle. [mzwiessele]
* Bugfix: lineplot in visualize. [mzwiessele]
* Added documentation for parameterized objects, needs more detail and fleshing out with proper english. [mzwiessele]
* Part working on symbolics. Replacing data_resources.json with the correct full file (-hapmap). Don't know why we've gone for separate create file ... [Neil Lawrence]
* Need to fix missing data in likelihoods. [Neil Lawrence]
* Merge of kern/__init__.py. [Neil Lawrence]
* New test heteroscedastic noise model. [Ricardo]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Ricardo]
* Bugfix: mixed up global and local index in unfixing. [mzwiessele]
* New file, special request. [Ricardo]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Ricardo]
Conflicts:
GPy/models/gp_classification.py
GPy/models/sparse_gp_classification.py
* New file. [Ricardo]
* Minor change. [Ricardo]
* EPDTC added. [Ricardo]
* Minor change. [Ricardo]
* Changes according to new definitions. [Ricardo]
* Changes according to new definitions. [Ricardo]
* Pseudo_EM is not available for the moment. [Ricardo]
* Just had to do a check in from midlantic (showing off). [Neil Lawrence]
* Merge branch 'params' of https://github.com/SheffieldML/GPy into params. [Neil Lawrence]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [James Hensman]
* [datasets] merged hapmap dataset into params. [Max Zwiessele]
* Bugfix: slicing was still in stationary somehow. [Max Zwiessele]
* Removed random.seed from the base of kernel_tests.py (the tests still pass) [James Hensman]
* [tests] added test for fixing through regular expressions. [Max Zwiessele]
* Bugfix: fixing through regular expression matching. [Max Zwiessele]
* Bugfix: couldn't constrain single parameters, because of indexing of np. [Max Zwiessele]
* [Kern] added assertion for inputs X being matrix (ndim == 2) [Max Zwiessele]
* [Kern] added assertion for inputs X being matrix (ndim == 2) [Max Zwiessele]
* Bug fix: slicing was not checking dimensions. [Max Zwiessele]
* Bug fix: slicing can now be turned of by passing None as active_dims. [Max Zwiessele]
* Slicing .... maybe needs to be deleted. [Max Zwiessele]
* Enabled some more getting/setting parameters, such as regular expressions and params. [Max Zwiessele]
* Pickle test error fixed. [Max Zwiessele]
* First draft of base symbolic object, compiling with symbolic mapping. [Neil Lawrence]
* Copy and paste observable_array from repository to try and resolve bizzare merge request. [Neil Lawrence]
* Psi stat and kernel tests new parameterization. [Max Zwiessele]
* Linear kern variational updates. [Max Zwiessele]
* Final touches to gradchecker. [Max Zwiessele]
* Gradchecker now with fixed inputs. [Max Zwiessele]
* Fixes added for gradchecking. [Max Zwiessele]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Max Zwiessele]
Conflicts:
GPy/core/parameterization/param.py
* Fixed a couple of small params bugs. [Alan Saul]
* Merged parameterized fixing. [Alan Saul]
* Laplace now appears to be grad checking again. [Alan Saul]
* Moved fix parameter to constrainable. [Alan Saul]
* Checkgrad now global and callable from any parameter. [Max Zwiessele]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Max Zwiessele]
Conflicts:
GPy/core/parameterization/param.py
GPy/core/parameterization/parameter_core.py
GPy/core/parameterization/parameterized.py
* Stupid error, needed to actually USE the gradients in student t... Looks like s2 of rasm's may have an extra -? dW_df == -d2logpdf_df not just d2logpdf_df? [Alan Saul]
* Have most of the likelihood testing working, laplace likelihood parameters need fixing, some of the signs are wrong I believe. [Alan Saul]
* Have most of the likelihood testing working, laplace likelihood parameters need fixing, some of the signs are wrong I believe. [Alan Saul]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Alan Saul]
* Fixed likelihood tests for new parameters structure. [Alan Saul]
* Fixed parameter bugs. [Alan Saul]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Alan Saul]
* Fixed parameterized oddity where it was updating all constrained parameters as soon as any were constrained rather than after all are constrained@ @ [Alan Saul]
* _highest_parent_ now follows the tree, dK_dX > gradient_X, added update_grads_variational to linear, bgplvm for new framework. [Max Zwiessele]
* Adjusted periodic exponential to the new parameterization. [Max Zwiessele]
* Small changes to parameterization init. [Max Zwiessele]
* Added some docstrings and the posterior class structure. [James Hensman]
* Messing wih he inference directory now. [James Hensman]
* Some more messing with the likelihood directory. [James Hensman]
* First commit in new structure. [James Hensman]
* Changes to symbolic. [Neil Lawrence]
* Latest updates for ties, still bery buggy, considering restructuring... [Max Zwiessele]
* Simple tieing now working, still buggy though, progress with care. [Max Zwiessele]
* Biiig changes in tieing, and printing -> hirarchy now always shown. [Max Zwiessele]
* Minor edits, like spacing, spelling. [James Hensman]
* Added variational. [Max Zwiessele]
* Bgplvm integrated. [Max Zwiessele]
* Added BGPLVM in parameterized. [Max Zwiessele]
* Added gplbm and sparse gp to new parameterized structure. [Max Zwiessele]
* Merge branch 'params' of github.com:SheffieldML/GPy into params. [Max Zwiessele]
* Adjusted parameters to report their changes. [Max Zwiessele]
* Updated scg messages. [Max Zwiessele]
* Tutorial update and X observer. [Max Zwiessele]
* X caching is not yet done, parameter caching working fine. X cache must be adjusted to update at the right times. [Max Zwiessele]
* Changing all parameterized objects to be compatible with the new parameterization. [Max Zwiessele]
* Adjusted gaussian likelihood to new parameterization. [Max Zwiessele]
* Gp_base newly paramterized. [Max Zwiessele]
* Updated white, bias and rbf to new parameter handling. [Max Zwiessele]
* GPRegression working, gradients still todo. [Max Zwiessele]
* First adjustments to the model and gps -> updates and gradient transforms. [Max Zwiessele]
* Parameters have a update hirarchy, in which updates to parameters are hirarchically pursuit. [Max Zwiessele]
* Likelihood is now parameterized. [Max Zwiessele]
* Kern is now parameterized. [Max Zwiessele]
* Kern is now parameterized. [Max Zwiessele]
* Parameters now work efficiently, tieing is iwth observer pattern. [Max Zwiessele]
* Still todo: untie, gradients, priors, print ties. [Max Zwiessele]
* Parameter_testing has to be written new. [Max Zwiessele]
* Starting to sort out likelihoods WARNING not working. [Max Zwiessele]
* Make sure _init_ is not overriden. [Max Zwiessele]
* WARNING: half way through commit, this is a non working middle thing! everything should be in place now, figure tieing and printing with broadcasting. [Max Zwiessele]
* Parameter indexing now linear in number of printed values. [Max Zwiessele]
* Index ops now with own dict. [Max Zwiessele]
* Documentation added. [Max Zwiessele]
* More testing. [Max Zwiessele]
* Warning messages optional for re-constraining. [Max Zwiessele]
* Indexing fixed, some equality testing. [Max Zwiessele]
* Updated sympy code, multioutput grad checks pass apart from wrt X. Similar problems with prediction as to sinc covariance, needs investigation. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Seems to work on windows now. [Nicolò Fusi]
not everything works yet, but I've identified the main issues. Still
* Added olivetti faces data set. It required adding netpbmfile.py a bsd licensed pgm file reader from Christoph Gohlke, which doesn't seem to have a spearate installer. Also modified image_show to assume by default that array ordering is python instead of fortran. Modified brendan_faces demo to explicilty force fortran ordering. Notified Teo of change. [Neil Lawrence]
* Domain and trtansformations namespace prettyfying. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into params. [Max Zwiessele]
* Some fixes and changes to the sympykern. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Nparam changes to num_params. [James Hensman]
* Changes Nparts for num_parts in kern. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Andreas]
* Minor changes. [Andreas]
* Normalize Y given as an argument to constructor. [Andreas]
* Fixed stick datasets bug ... but sympykern is currently in a rewrite so will be broken. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Plots tidied up. [Ricardo]
* Modifications to allow noise_model related parameters. [Ricardo]
* Coregionalization examples fixed. [Ricardo]
* Sampling function added. [Ricardo]
* Added first draft of functionality for multiple output sympy kernels. [Neil Lawrence]
* Tests updated. [Max Zwiessele]
* Dim reduction examples Q= > input_dim= [Max Zwiessele]
* Numpy non hashable AHHHHHH. [Max Zwiessele]
* Dont print brackets in transformations. [Max Zwiessele]
* Parameterized first beta test. [Max Zwiessele]
* Merge branch 'devel' into params. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [mu]
* Testing ODE. [mu]
* Added capability for sinc covariance via sympy kernel. [Neil Lawrence]
* Minor changes to della_gatta example (multiple optima). [Neil Lawrence]
* Replaced check for sympy in constructors.py. [Neil Lawrence]
* Change to criterion on positive definite check (epsilon*10 instead of epsilon). [Neil Lawrence]
* Remove coregionalization test as it's causing a core dump! Need to chase this up. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Some tidying in the EP likelihood. [James Hensman]
Changes self.N to self.num_data for consistency with everywhere else
added the factor of 2pi to Z.
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Allowed passing of factr to bfgs algorithm. [James Hensman]
* Updates to sympykern including bug fixes and ability to name covariance. Include test for rbf_sympy in kernel tests. Remove coregionalization test as it's causing a core dump! Need to chase this up. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Fixing W_columns and num_outputs nomenclature. [Ricardo]
* Added missing files. [Neil Lawrence]
* Added missing file. [Neil Lawrence]
* Missing file. [Neil Lawrence]
* Merging changed files. [Neil Lawrence]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge pull request #82 from jamesmcm/devel. [James McMurray]
Devel
* Fixed more errors in docs 2. [James McMurray]
* Fixed more errors in docs. [James McMurray]
* Adding extra tests for bgplvm. [James Hensman]
* Merging changed files. [Neil Lawrence]
* Add eq_ode1 covariance. [Neil Lawrence]
* More samples for higher sampling accuracy. [Max Zwiessele]
* Sparse_gp_multioutput test added. [Ricardo]
* Removing unnecessary stuff... [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
Conflicts:
GPy/examples/classification.py
* Fixed a bug in Neil's otherwise tidy hetero kernel. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge pull request #80 from jamesmcm/devel. [James McMurray]
Devel
* Merge branch 'devel' of git://github.com/SheffieldML/GPy into devel. [James McMurray]
* Fixed docstring warnings - could still be mistakes. [James McMurray]
* Started fixing docs. [James McMurray]
* Fixed three tests by being _slightly_ less stringeent about poositive-definiteness. [James Hensman]
* Small change in crescent demo. [Ricardo]
* Please stop breaking this module. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Do_test_latents appears to be working now. [James Hensman]
* Added hetero back to the init. [James Hensman]
* Epsilon and power_ep now are parameters of update_likelihood. [Ricardo]
* Crescent data example is better organized. [Ricardo]
* Cross term testing switched on. [Max Zwiessele]
* Comment linear x linear for testing purposes. [Max Zwiessele]
* Parameter_testing. [Max Zwiessele]
* Index operations now work on flattened indices. [Max Zwiessele]
* Unoptimized parameter, still slower than current implementation. [Max Zwiessele]
* Ndarray subclass cleaned up. [Max Zwiessele]
* Subclassing ndarray almost functional. [Max Zwiessele]
* Parameters ndarray, stuck at using strides for transformations. [Max Zwiessele]
* Transformations are singletons now, weak refs for memory managment. [Max Zwiessele]
* Almost there with array inheriting. [Max Zwiessele]
* With subclassing ndarray, current_slice problems not solved... [Max Zwiessele]
* Without inheriting from numpy.ndarray. ndarray functionality missing. [Max Zwiessele]
* NegativeLogexp Pep8ted. [Max Zwiessele]
* Index operations finalized. [Max Zwiessele]
* Added index_operations and deleted them from paramter. [Max Zwießele]
* Added parameter files - Alan. [Max Zwiessele]
* Transformations singleton. [Max Zwiessele]
* Merge branch 'devel' into params. [Max Zwiessele]
Conflicts:
GPy/core/transformations.py
GPy/kern/parts/kernpart.py
* Merge pull request #77 from jamesmcm/devel. [James McMurray]
Devel
* Merge branch 'devel' of git://github.com/SheffieldML/GPy into devel. [James McMurray]
* Bug in prod-coreg kernels fixed, not in the most elegant way though. [Ricardo]
* Disp=False in scipy optimizations. [Ricardo]
* Function grep_model added, works like print model, but accepts regexp. [Ricardo]
* Reverting error. [Ricardo]
* Rebuilt documentation. [James McMurray]
* Changes to fix autodoc - need to test with ReadTheDocs still. [James McMurray]
* Fixed readme. [James McMurray]
* Updated readme with instructions for compiling documentation, running unit tests. [James McMurray]
* Normalize_Y is passed to core function. [Ricardo]
* Redundant code commented. [Ricardo]
* Samples in plot_f fixed for sparse_models. [Ricardo]
* Comments added to assertions. [Ricardo]
* Coreg_spars fixed. [Ricardo]
* Heaviside instead of heavyside. [Ricardo]
* 1 docstring line completed. [Ricardo]
* Random 2 bug neutralized... not fixed. [Ricardo]
* Heaviside transformation fixed. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Likelihoods are now Parameterized objects. [Ricardo]
* Change in gaussian noise to avoid confusion with Gaussian likelihood. [Ricardo]
* AddedHeraviside functionality to EP. [James Hensman]
* Fixed bug in rat_quad for RW. [James Hensman]
* Fixed Zsolts bug in prod.py. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
Conflicts:
GPy/examples/regression.py
GPy/kern/constructors.py
GPy/testing/kernel_tests.py
GPy/util/multioutput.py
* Other local changes. [Neil Lawrence]
* Actually changing coregionalise to coregionalize. [James Hensman]
* Renamed coregionalise to coregionalize to make it consistent with other spellings (optimize etc.) [Neil Lawrence]
* Merged conflict of model.py. [Neil Lawrence]
* Changed kern.py so that X2 is correctly passed as None to the kern parts for dK_dX. Modified several kern parts so that dK_dX is correctly computed (factors of 2 missing). Removed spurious factors of 2 from gplvm, bcgplvm, sparse_gp and fitc code. [Neil Lawrence]
* Print_all function removed, print m works as before. [Ricardo]
* Coregionalisation changed to coregionalization. [Ricardo]
* LinK2_functions2 merged. [Ricardo]
* Merge branch 'linK_functions2' into devel. [Ricardo]
Conflicts:
GPy/core/gp.py
GPy/core/gp_base.py
GPy/core/sparse_gp.py
GPy/examples/regression.py
GPy/kern/constructors.py
GPy/kern/parts/coregionalise.py
GPy/models/__init__.py
GPy/models/sparse_gp_classification.py
GPy/util/__init__.py
* Useless files deleted. [Ricardo]
* Moved to (sparse_)gp_multioutput_regression. [Ricardo]
* Doesn't matter I'll erase this file. [Ricardo]
* Duplicated line erased. [Ricardo]
* Sparse multiple outputs model with Gaussian noise. [Ricardo]
* Multiple outputs model with Gaussian noise. [Ricardo]
* Likelihood gradients for heteroscedastic noise. [Ricardo]
* Changes in plot functions, to allow 1D multiple outputs visualization. [Ricardo]
* Docstrings in new functions. [Ricardo]
* Works well now. [Ricardo]
* Redundant variable eliminated. [Ricardo]
* R paramter renamed as W_columns, and Nout renamed as num_outputs. [Ricardo]
* New files added. [Ricardo]
* Build_cor_kernel is now called build_lcm. [Ricardo]
* Docstring completed. [Ricardo]
* Changes according to new likelihoods definition. [Ricardo]
* Errors fixed. [Ricardo]
* Docstrings completed. [Ricardo]
* Docstrings completed. [Ricardo]
* Bug fixed. [Ricardo]
* Step function modified, now the output is either 1 or -1. [Ricardo]
* Step transformation added. [Ricardo]
* New Gaussian likelihood for multiple outputs. [Ricardo]
* Error in plot corrected. [Ricardo]
* Multioutput models added. [Ricardo]
* Changes to allow multiple output plotting. [Ricardo]
* Plots for multiple outputs. [Ricardo]
* Multioutput is working. [Ricardo]
* Gamma noise added. [Ricardo]
* Random changes. [Ricardo]
* The next step is to optimize the noise models' parameters. [Ricardo]
* Files re-organized. [Ricardo]
* Confidence interval fixed. [Ricardo]
* Massive changes. [Ricardo]
* More files. [Ricardo]
* Further corrections. [Ricardo]
* Massive changes. [Ricardo]
* Predictive mean done. [Ricardo]
* More changes. [Ricardo]
* Predictive values, new method. [Ricardo]
* Some cool stuff for EP. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Fixed args to bfgs. [James Hensman]
* Fixed Alans checkgrad bug. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Working on the Poisson likelihood. [Ricardo]
* Parameterization changes take a while. [Max Zwiessele]
* Correcterd minor errors (imports etc) [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Added gibbs.py, although test is still failing. [Neil Lawrence]
* Added slicing to kern.py. [Max Zwiessele]
* Cleaned up imports. [Max Zwiessele]
* Removed white kernel, bc of likelihood variance. [Max Zwiessele]
* Plotting adjusted. [Max Zwiessele]
* Added print m and print m.all differentiation. [Max Zwiessele]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Merge branch 'devel' of git@github.com:SheffieldML/GPy.git into devel. [Max Zwiessele]
* Merge dim reduction. [Max Zwiessele]
* Fixed gplvm unittest which was sampling an output matrix of size input_dim (and corresponding error in bcgplvm unittest which was based on it). [Neil Lawrence]
* Added mlp mapping (with possibility of having multiple layers). [Neil Lawrence]
* Added simple back constrained GPLVM model. [Neil Lawrence]
* Added unit tests for mapping functions. [Neil Lawrence]
* Implemented Mapping framework and associated linear and kernel mappings. This is needed for mean functions, back constrained GPLVM and the non-stationary Gibbs kernel. [Neil Lawrence]
* Implemented MLP gradients with respect to X. [Neil Lawrence]
* Polynomial kernel gradients wrt X working. [Neil Lawrence]
* Changed default values of W and kappa for coregionalisation kernel. Changed names of keyword arguments from Nout and R to output_dim and rank. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Fixed target slicing bug in prod kernel. [James Hensman]
* Added gpx dataset. [Neil Lawrence]
* Added missing poly.py file. [Neil Lawrence]
* Added models for testing kernel gradients in unit tests. [Neil Lawrence]
* Various doc string edits and consistency changes in comments. [Neil Lawrence]
* Minor fixes. [Neil Lawrence]
* Added notes file for issues raised while looking through code, some are things I need to raise on github, others need some informal discussion, but for the moment thought to put them informally here, given flakiness of internet connection. [Neil Lawrence]
* Minor fixes to classification to allow kernel choice, change of oil example to use full test set and full training set. [Neil Lawrence]
* Minor fixes to regression example with robot_wireless. [Neil Lawrence]
* Added robot_wireless data set and examples. [Neil Lawrence]
* Added ability to load in cmu motion capture data bases in the new data base loading format. [Neil Lawrence]
* Finished first draft of system for downloading data sets. [Neil Lawrence]
* Added first draft of polynomial kernel. [Neil Lawrence]
* Added automatic computation of dKdiag_dtheta in kernpart.py using dK_dtheta. [Neil Lawrence]
* Added mlp covariance (x gradient not working) [Neil Lawrence]
* Modifications to transformations ... not sure which tests to run to make sure I haven't messed things up. New code avoids exponentiating values greater than -log(eps) or less than log(eps). Also changed negative code to call the positive code (I think they should inherit the positive code ... but maybe not. [Neil Lawrence]
* Mocking matplotlib pyplot as readthedocs is failing to import it (again...) [Alan Saul]
* Prodpart bugfix. [Max Zwiessele]
* Linear kernel old implementation. [Max Zwiessele]
* Better bound extimate for plot_latent background. [Max Zwiessele]
* Visual? [Max Zwiessele]
* Changed psi0 of white to be zero. [Max Zwiessele]
* Constant jitter to Kmm, deleted some white kernels in models and examples. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Added connections.txt. [Neil Lawrence]
* Implemented smart pickling for svigp. [Andreas]
* Added plot of magnification factor. [Alessandra Tosi]
* Added jacobian and magnification factor. [Alessandra Tosi]
* Oops, that was silly bug. Don't code tired, kids. [James Hensman]
problem summing psi2 in sparse_GP
* Some profilng and debugging in sparse_GP. [James Hensman]
* Input_sensitivity right way. [Max Zwiessele]
* Import handling... [Max Zwiessele]
* Import handling... [Max Zwiessele]
* Added imshow controller and label controller. [Max Zwiessele]
* Minor/pep8. [Max Zwiessele]
* Minor adjustment to fast_array_equal. [Max Zwiessele]
* Mrd plot_scales improved. [Max Zwiessele]
* Gradient checker more robust against name changes. [Max Zwiessele]
* Experimental plotting of gradient in latent space (and addition of steepest gradient dimensions) [Max Zwiessele]
* Pca initializer more robust to lower dimensions. [Max Zwiessele]
* Ard and latent plotting improved. [Max Zwiessele]
* Merge of rbf_inv failed, corrected with AD. [Max Zwiessele]
* Gradient checker implemented. [Max Zwiessele]
* Gradient checker implemented. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Andreas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Changes to psi2 in linear. [James Hensman]
* Minor fix. [Andreas]
* Gradient checker implemented. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Fixed asarray in example toy_ARD. [Andreas]
* Rbf kern chaching bug fixed. [Max Zwiessele]
* Merge rbf_inv changes. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Small changes to rbf and rbf_inv. [James Hensman]
* Added gradient checker model. [Max Zwiessele]
* Caching bugfix for psi2 computations. [Max Zwiessele]
* Reverted stupid merge error. [Max Zwiessele]
* Dimensionality reduction demo. [Max Zwiessele]
* Ard plotting enhanced. [Max Zwiessele]
* Rewritten most parts wrt inv_lengthscale (still missing dK_dtheta and dPsi2_dtheta can be written better) [Andreas]
* Input sensitivity for rbf_inv. [Andreas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Andreas]
* Model robustness greatly improved and avisualize adjustments for plot_latent. [Max Zwiessele]
* Robust failure handling in model objective and gradient. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Fixed blaslib bug, I hope. [James Hensman]
* Catch linalg errors inside model and more sopihisticated non-pd checks. [Max Zwiessele]
* Adjusted doc to new pep8 format. [Max Zwiessele]
* Fixed a transpose bug in the sparse GP prediction. [James Hensman]
* Merged with master. [Alan Saul]
* Removed slices from GP regression wrapper. [James Hensman]
Xslices no longer exists in the GPy framework.
* Commented out a buggy ax.set_aspect line. [Teo de Campos]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Removed the unnecessary test skip. [Nicolas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Un-commented out visualize.py. [Teo de Campos]
* Resurrected visualize.py. [Teo de Campos]
* Fixed mrd init call. [Alan Saul]
* Removed overwriting_b in lapack. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Fixed lapack importing for old scipy version. [James Hensman]
* Ignoring example tests again. [Alan Saul]
* Changed deprecation supression to 12. [Alan Saul]
* Merged. [Alan Saul]
* Merge branch 'psi1_transpose_fix' into devel. [James Hensman]
* Psi1 is now the right way around. [James Hensman]
* Fixing lapacks. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Removing unused link_functions. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Modified lengthscale gradients - demo works now. [James Hensman]
* Removed visualize dependencies and added dKdiag_dX for linear. [Nicolas]
* Factored out lapack into utils so we can check version and give deprecation warnings. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Making sure GPy imports. [Neil Lawrence]
* Part changes to datasets.py and mocap.py to download data resources for examples. Not working currently! [Neil Lawrence]
* Placed back in examples for motion capture! Added spheres to visualization of figure. [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Added visualization for motion capture data using python visual module. [Neil Lawrence]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* _Xmean is now Xoffset and _Xstd is now _Xscale. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Removed examples with non public datasets. [Nicolas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Removed examples with non public datasets. [Nicolas]
* FITC example: bound for lengthscale. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Fixed tuto example. [Nicolas]
* Non_Gaussian exampless deleted. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Fitc and generalized_fitc models deleted. [Ricardo]
* Reduced number of iterations for a couple of things. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Poisson likelihood implementations needs to be thought carefully. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Dimensionality reduction mrd example less interations. [Max Zwiessele]
* Link_function class renamed as LinkFunction. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Minor changes. [Ricardo]
* Pseudo_EM modified. [Ricardo]
* Examples corrected. [Ricardo]
* MRD fix. [Max Zwiessele]
* Nice plot handling in tutorials. [Nicolas]
* Example multiple optima now returns a model. [Nicolas]
* Refactored example tests. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Kern constructors now have input_dim instead of D. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merged conflict in tutorial's tests (again) [Nicolas]
* Merged conflict in tutorial's tests. [Nicolas]
* Bugs fixed in tutorial's tests. [Nicolas]
* Dim reduction adaption. [Max Zwiessele]
* Num_data refactoring. [Max Zwiessele]
* Fixed merge conflict in examples_tests. [Alan Saul]
* Git pushMerge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
Conflicts:
GPy/core/fitc.py
* FITC example added. [Ricardo]
* FITC is back. [Ricardo]
* Psi stat tests adapted (psi1 still not tested, bc of transpose confusion) [Max Zwiessele]
* Merged num_data conflicts. [Max Zwiessele]
* Merge kern conflicts in examples. [Max Zwiessele]
* Kern params adapted: Nparams > num_params and fixes of input_dim. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Nparams > num_params and Nparam_tranformed > num_params_transformed. [Max Zwiessele]
* Made examples possible to run all examples and throw out a dictionary of problems at the end (and whilst it's running, tried to ignore deprecation warnings. [Alan Saul]
* Lots of bugfixes after refactoring. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Alan Saul]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* New file. [Ricardo]
* Fixed merge conflicts, M now num_inducing. [Alan Saul]
* Refactoring: self.D > self.input_dim in kernels. [Max Zwiessele]
* Output_dim instead of input_dim. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* An assortment of fixes. [James Hensman]
* Change input for output. [Ricardo]
* Seems to run now. [Ricardo]
* Change input for output. [Ricardo]
* New file. [Ricardo]
* Changed all M's for num_inducing. [Alan Saul]
* Fixed naming to standardized PEP8. [Ricardo]
* Merged and fixed conflicts, names still need changing accordingly. [Ricardo]
* Refactoring files added. [Max Zwiessele]
* Merged regression example, corrected refactoring. [Max Zwiessele]
* Preferred optimiser is now scg bydefault. [James Hensman]
* Fixed the multiple optima demo. [James Hensman]
* Getting examples running. [Max Zwiessele]
* REFACTORING: model names, lowercase, classes uppercase. [Max Zwiessele]
* New FITC model and other stuff. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* How did that happen? [James Hensman]
* Merged an emty line... [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Dim reduc plotting. [Max Zwiessele]
* Merged plotting and input_dim. [Max Zwiessele]
* Dimensionalityreduction plotting adjusted to new syntax. [Max Zwiessele]
* Constrain fixed now unconstrains first. [Nicolas]
* Changed the behaviour of checkgrad. [James Hensman]
checkgrad usd to check the passed string (for name matching) against the
list of _get_param_names(). Then it would index along
_get_param_names_transformed()!
this led to inconsistensies when fixed or tied variables were used,
which screwed up the ordering of the variable names.
We now match against _get_param_names_transformed().
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Replaced Q by input_dim. [Alan Saul]
* Tutorials updated to comply with changes throught the code. [Nicolas]
* Modified the regression demos r. [James Hensman]
they all seem to work now.
* Made input_ubncertainty plotting work, modified example a little. [James Hensman]
* Minor changes. [Ricardo]
* Correction to some tests. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Plotting in the model behaves better. [James Hensman]
* Plotting behaviour adapted for BGPLVM. [Max Zwiessele]
* Plotting behaviour adapted for BGPLVM. [Max Zwiessele]
* Plotting behaviour adapted for kern and mrd. [Max Zwiessele]
* Making examples working. [Nicolas]
* New model. [Ricardo]
* New model. [Ricardo]
* New models included. [Ricardo]
* Examples corrected. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Fixed printing, some example. [James Hensman]
* Added priors behaviour as intended and issue #38 closed and fixed. [Max Zwiessele]
* New GP_classification model. [Ricardo]
* FITC test not skipped any more. [Ricardo]
* Unit_tests corrected. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Re-merged. only RA's errors (probit?) remain. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* MERGE. [Max Zwiessele]
* Domains added and class names in priors capitalized. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
Conflicts:
GPy/models/GPLVM.py
* Remove copies (they are now in visualize code ...). [Neil Lawrence]
* Neil's flailing attempts to update the flailing stick man. [Neil Lawrence]
* Match_moments function passes transformed values. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Added domain matching in set_prior. [Max Zwiessele]
* Added domains to priors. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Examples changed to use new link_functions. [Ricardo]
* Merge branch 'devel' into link_functions. [Ricardo]
* Link functions defined. [Ricardo]
* Added domains in transformatins and priors. [Max Zwiessele]
* Tidied the computation of A in sparse_GP. [James Hensman]
* Enabled DSYR on feeble machines. [James Hensman]
i.e. where numpy is compiled without proper blas linkage
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Cross-terms. [Nicolo Fusi]
* Test. [Nicolo Fusi]
* Minor SGD changes. [Nicolo Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Changes in SGD. [Nicolo Fusi]
* SCG printing prettyfied. [Max Zwiessele]
* Merge branch 'mrd' into devel. [Max Zwiessele]
* Last changes on BGPLVM stable version. [Max Zwiessele]
* Newline only on display. [Max Zwiessele]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Andreas]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Teo de Campos]
* Readded float variance. [Max Zwiessele]
* Max_bound adjust. [Max Zwiessele]
* Swiss_roll example changes. [Max Zwiessele]
* Catch beta > 0 not needed anymore. [Max Zwiessele]
* Swiss_roll adjustments. [Max Zwiessele]
* Stability enhancing clipping in logexp_clipped and reverse of stability clipping of parameters. [Max Zwiessele]
* Added propper clipping, before overflow happens. [Max Zwiessele]
* Added proper linebeak for SCG printing. [Max Zwiessele]
* Merge branch 'mrd' into devel. [Max Zwiessele]
* Swiss_roll example added, BGPLVM_oil now working. [Max Zwiessele]
* Clipping bounds released to 1e6. [Max Zwiessele]
* Clipping now upper bound adjustable. [Max Zwiessele]
* Added .orig and noseid files to be ignored. [Max Zwiessele]
* Minor formatting changes. [Max Zwiessele]
* WARNING: added parameter clipping for catching infinity problems [p in (-1e300, 1e300)] [Max Zwiessele]
* Printing improved. [Max Zwiessele]
* Merge branch 'devel' into mrd. [Max Zwiessele]
* Swiss_roll example. [Max Zwiessele]
* Added swissroll example. [Max Zwiessele]
* Making clipping adjustable. [Max Zwiessele]
* Modified: GPy/models/GPLVM.py Using the following kernel by default: kernel = kern.rbf(Q, ARD=Q>1) + kern.bias(Q, np.exp(-2)) + kern.white(Q, np.exp(-2)) [Teo de Campos]
* Modified: GPy/util/visualize.py Added the functionality of showing a mosaic of NxN reconstructed images when the size of the number of elements in the feature vector greater than dimensions[0]*dimensions[1]. [Teo de Campos]
* Modified: GPy/examples/dimensionality_reduction.py brendan_faces(): normalizing the feature vectors w.r.t. the global mean and standard deviation. Changed optimisation constraints because it was never converging. [Teo de Campos]
* Need to commit to resolve a conflict ... [Neil Lawrence]
* Overwrite my changes with James's. [Neil Lawrence]
* Merge changes. [Neil Lawrence]
* Separated out untransform_params, enabling flexibility downstream. [James Hensman]
* Docstrings in SCG. [Andreas]
* Prediction returns Nx1 vectors. [Ricardo]
* Chol_inv added somewhere. [Ricardo]
* Chol_inv added somewhere. [Ricardo]
* Convenient but not important changes. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Changed prod_orthogonal in tests. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Allowed GP models to plot multiple outputs (in 1D at least) [James Hensman]
* Test corrected, by Nicolas. [Ricardo]
* EP case is working fine. [Ricardo]
* Some changes to speed up... just a little. [Ricardo]
* Explanation added to DSYR. [Ricardo]
* Some changes. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Teo de Campos]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Speeding up FITC. [Ricardo]
* GPy/util/visualize.py: fixed conflict. [Teo de Campos]
* Ricardo told me to do this. [Teo de Campos]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Removed RA's profile deco. [James Hensman]
* Generalised backsub_both_sides. [James Hensman]
* Remove profile. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolo Fusi]
* Moved linalg function to GPy.linalg. [Nicolo Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolò Fusi]
* Trying to follow changes in likelihood. [Nicolò Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolò Fusi]
* Small changes. [Nicolò Fusi]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolò Fusi]
* Better f_inv. [Nicolò Fusi]
* Small change to GPLVM param indexing. [Nicolò Fusi]
* Minor changes to the symmetrify function. [James Hensman]
* DSYR is being used now. [Ricardo]
* Gradients are working now. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Added DSYR for ricardo. [James Hensman]
* Gradietns check :) [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Using weaved symmetrify in pdinv now. [James Hensman]
* Fixed the bug in stick. [James Hensman]
* New functions for EP-matching moments. [Ricardo]
* Minor change. [Ricardo]
* Change in gradients computation. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Ricardo]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Nicolas]
* Merge conflict in transformations. [James Hensman]
* Minor modifications to the constraint behaviour. [James Hensman]
* Unified framework for addition and product of kernels, with a tensor flag (boolean) instead of and. [Nicolas]
* MRD updates merge. [Max Zwiessele]
- termination rule changes for SCG
- oil flow updates
* Example update to run oil dataset. [Max Zwiessele]
* MRD updates and minor changes. [Max Zwiessele]
* Termination rule update. [Max Zwiessele]
* More changes. [Ricardo]
* Adding FITC to the list. [Ricardo]
* Some changes. [Ricardo]
* Changes to FITC. [Ricardo]
* New FITC... again :( [Ricardo]
* Reverted linalg lapack.flapack for the poor NIPS deadline people. [Alan Saul]
* Image_show() is now able to use a palette to map quantized images to their original color. This uses PIL (import Image). I also enabled the image to be normalised from 0 to 255 in a more robust way (before this, it was assuming images ranged from 0 to 1). [Teo de Campos]
* Merged. [Teo de Campos]
* Merge MRD stability and scg termination rule into devel. [Max Zwiessele]
* Last changes for mrd stability. [Max Zwiessele]
* Scale Factor removed and moved V=Y*beta into likelihoods. [Max Zwiessele]
* BGPLVM MRD Examples and plotting adjustments. [Max Zwiessele]
* Conjugate gradient optimizer without callback (no c.join) [Max Zwiessele]
* SCG convergence tolerance increased, will now converge more easily. [Max Zwiessele]
* Fixed a bug in all_constrained_indices. [James Hensman]
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [James Hensman]
* Typo corrected for negative constrains. [Nicolas]
* Fixed a tie-bug for ND. [James Hensman]
* Added file:transformations. [James Hensman]
* Eigenvalue decomposition of psi2. [James Hensman]
* Merge branch 'devel' into new_constraints. [James Hensman]
Conflicts:
GPy/core/model.py
GPy/core/parameterised.py
* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [Max Zwiessele]
* Minor tidy up of names in visualize (replace histogram with bar chart in lvm_dimselect). [Neil Lawrence]
* Merge branch 'devel' of https://github.com/SheffieldML/GPy into devel. [Neil Lawrence]
* Trying to upgrade numpy version to 1.7.1 as there was an error introduced for weave on 1.7.0 causing tests to fail. [Alan Saul]
* Unification of the visualize object hierarchy and standardization of the click and move behaviour of lvm and lvm_dimselect. Set colours of input sensitivity histogram to red for left (port) and green for right (starboard). [Neil Lawrence]
* Readded parameterized changes. [Max Zwiessele]
* Model re compilation added. [Max Zwiessele]
* Whitespace. [James Hensman]
* Fixed bug in constrain_fixed where soem values weren't deleted. [James Hensman]
* Improved stability of sparse GP for certain-input case. [James Hensman]
* First attemot at the new constraint framework. [James Hensman]
* Yak shaving. [James Hensman]
* More stabilisation of sparse GP. [James Hensman]
* Removed uncollapsed sparse GP. superceeded by the forthcoming svigp package. [James Hensman]
* Reimplemented caching in prod_orthogonal... [James Hensman]
* Weaved coregionalise. much performance gained. [James Hensman]
* Remo0ved slices from models. [James Hensman]
slices are now handles by special indexing kern parts, such as
coregionalisation, independent_outputs. The old slicing functionality
has been removed simply to clean up the code a little.
Now that input_slices still exist (and will continue to be useful) in
kern.py. They do need a little work though, for the psi-statistics
* Cmu_mocap() example mostly working except some fiddling with axes for visualization. Also changes to naming of scaling and offset parameters in GP.py and deal with the case where the scale parameter is zero. [Neil Lawrence]
* Added m['ard'] gives all parameters matching 'ard', as well as setting m['ard'] = x to set all mrd parameters. [Max Zwiessele]
* Stupid kern stash merge. [Max Zwiessele]
* Kern conflict. [Max Zwiessele]
* Kern stash conflict. [Max Zwiessele]
* Psi_stat_test stash. [Max Zwiessele]
* Merge branch 'devel' into mrd. [Max Zwiessele]
* Bugfix: cross term psi1 bias + linear. [Max Zwiessele]
* Psi1 not working (strange transposes) [Max Zwiessele]
* Moved *2. of psi2 statistics into kern and corrected bias+linear cross term. [Max Zwiessele]
* Merge branch 'devel' into mrd. [Max Zwiessele]
* Psi stat tests done and failing gracefully. [Max Zwiessele]
* Changed the name of GP_EP (from simple) in the unit test, added a messages option for full EP. [James Hensman]
* Changes the naming of kern components. [James Hensman]
Kern components now only get a number if their name is duplicated
* Fixed small bugs in linalg, setup.py. [James Hensman]
* Changed the behaviour of pdinv: now returns L and Li as well as the inverse. [James Hensman]
* More tidying in EP, removed examples from _module_ ( and opened discussion on github. [James Hensman]
* Trivial merge resolution. [James Hensman]
* Some tidying in the EP code. [James Hensman]
* Removed uncertain gp regression from the model __init__, since it's now just a switch in the sparse GP. [James Hensman]
* Fixed SGD to work with new interface. [Nicolo Fusi]
* Fixed merge. [Nicolo Fusi]
* Models are now pickleable. [Nicolo Fusi]
* Working implementation of SGD. [Nicolo Fusi]
* GPLVM accepts an initial value for X (in case you don't want to use the default random/PCA init) [Nicolo Fusi]
* Now passing a reference of the model to the optimizer (used in SGD) [Nicolo Fusi]
* Fixed bug in rbf_ARD kernel (dK_dX) [Nicolo Fusi]
* Fixed import error for examples and import error for uncertain inputs GP. [Nicolo Fusi]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolo Fusi]
* Merge changes for model.py and optimization.py on comments. [Neil Lawrence]
* Minor edits. [James Hensman]
* Models can now specify a preferred optimser (defaults to tnc) [James Hensman]
* Some tidying in the uncollapsed GP. [James Hensman]
* Made uncertain inputs a simple swith in the sparse GP class. This simplifies the inherritance structure. [James Hensman]
* Merge branch 'master' of github.com:SheffieldML/GPy. [James Hensman]
* Rbf_ARD now in the updated format for the computation of the derivatives (included for the psi-statistics, but not tested) [Nicolas]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolas]
* GPy: Some rewriting for the exponential and Matern kernels. They now pass the unit test. [Nicolas]
* Added a skeleton of the uncollapsed sparse GP. [James Hensman]
* Minor commenting. [James Hensman]
* General tidying in sparse_GP_regression. [James Hensman]
Have also made a small ampount of headway in
enabling heteroscedatic noise.
* Saved a little computation by exploiting the symmetry of a matrix. [James Hensman]
* I think the gradients bug in the sparse GP model is due to Kmm being unstable to invert. REducing M in some of the examples really helps. [James Hensman]
* Added datasets.py back in and minor changes. [Neil Lawrence]
* Fixed bug in GP_regression. [Nicolo Fusi]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolo Fusi]
* Derivatives of the exponential kernel in the right format. [Nicolas]
* Added precomputation of linear kernel, changed the logic a bit. [Nicolo Fusi]
* Fixed bug in linear_ARD. [Nicolo Fusi]
* Merge branch 'master' of github.com:SheffieldML/GPy. [Nicolo Fusi]
* Added vim's swp files to gitignore. [James Hensman]
* Demo for GP regressio with uncertain inputs. [James Hensman]
