Commit graph

4405 commits

Author SHA1 Message Date
James Hensman
e1d0965866 changed version number in setup.py 2013-04-10 12:15:23 +01:00
James Hensman
98ade75661 changed version number in setup.py 2013-04-10 12:14:58 +01:00
James Hensman
c4b2caa367 Merge branch 'devel' 2013-04-10 12:12:51 +01:00
Nicolo Fusi
807e85123e Merge branch 'master' of github.com:SheffieldML/GPy 2013-04-10 11:54:55 +01:00
James Hensman
eb1d8f211f stability improvements in sparse_GP 2013-04-10 11:04:49 +01:00
James Hensman
45403105f9 changed X_uncertainty for X_variance (in the code) for consistency with actual naming (in the printing) 2013-04-10 09:34:15 +01:00
James Hensman
fd0b172e81 various work on BGPLVM oil demo 2013-04-10 09:28:58 +01:00
James Hensman
bb734a6dd7 added simple BGPLVM_oil demo 2013-04-09 15:43:06 +01:00
Alan Saul
e0c1e4a4df Fixed laplace approximation and made more numerically stable with cholesky decompositions, and commented 2013-04-08 19:58:54 +01:00
Alan Saul
431f93ef23 Stabalised most of the algorithm (apart from the end inversion which is impossible) 2013-04-08 18:09:07 +01:00
Alan Saul
31d8faecf8 Added timing and realised mdot can be faster as its almost always a diagonal matrix its multiplying with 2013-04-05 17:56:02 +01:00
Alan Saul
4a14a82dfb Got the mode finding without computing Ki 2013-04-05 17:34:11 +01:00
Ricardo Andrade
0af208c1b6 Merge branch 'devel' of github.com:SheffieldML/GPy into devel 2013-04-05 17:28:38 +01:00
Ricardo Andrade
22ad5d94e9 Changes in FITC approximation computation 2013-04-05 17:27:44 +01:00
James Hensman
8e4d839b5d yak shaving 2013-04-05 12:08:21 +01:00
James Hensman
97704d9928 added the rbfcos kernel
ARD seems to work

dK_dX still todo
2013-04-05 11:09:14 +01:00
Alan Saul
7d5342ffd6 Merge branch 'master' of github.com:SheffieldML/GPy 2013-04-03 14:25:02 +01:00
Alan Saul
c7a58acd8f Added testing to modules 2013-04-03 14:24:55 +01:00
Alan Saul
2006a94caa Fixed broadcasting bug, rasm now appears to work 2013-04-03 10:55:58 +01:00
Alan Saul
0312f319ad Still working on rasmussen, link function needs vectorizing I think 2013-04-02 20:00:31 +01:00
Neil Lawrence
97db2a5bd7 Edit to linalg.py PCA function to stop it changing data matrix. 2013-04-02 16:53:42 +02:00
Alan Saul
afa5b1f956 Tidying up 2013-04-02 12:39:57 +01:00
Alan Saul
ffc168c1d2 Added predicted values for student t, works well 2013-04-02 12:33:01 +01:00
Neil Lawrence
9422c603b1 Minor modifications to visualization routines and examples. 2013-04-02 10:49:09 +02:00
Neil Lawrence
fce4dd7fde Further edits on visualization code for faces example. 2013-04-02 02:20:53 +02:00
Neil Lawrence
3fd0672092 Added base implementation of data visualization framework for use with GP-LVM. 2013-04-01 20:20:01 +01:00
Neil Lawrence
0f168a1667 Added mocap.py for loading in motion capture data. 2013-04-01 11:01:16 +01:00
Alan Saul
15d5c2f22d Working laplace, just needs predictive values 2013-03-28 17:42:42 +00:00
Ricardo Andrade
d42b731146 Merge branch 'fitc' into devel 2013-03-27 15:10:02 +00:00
James Hensman
85810030c5 small efficiency changes in rbf 2013-03-27 15:08:25 +00:00
Ricardo Andrade
bf9e6cbd5f Not needed 2013-03-27 15:06:21 +00:00
Ricardo Andrade
67d3ba6700 small changes 2013-03-27 15:05:50 +00:00
James Hensman
adfa6de1d8 rbf now works in a more memory friendly fashion 2013-03-27 15:03:46 +00:00
Ricardo Andrade
942e00d9a9 Merge branch 'em_fix' into fitc 2013-03-27 14:56:11 +00:00
Ricardo Andrade
c72d2b99d9 print iteration number 2013-03-27 13:52:51 +00:00
Nicolas
f2b49fe69a Bug fixed in periodic kernels: Warning were not handled properly 2013-03-27 12:45:08 +00:00
James Hensman
3546a2a729 Merge branch 'master' into devel 2013-03-27 12:28:51 +00:00
Nicolo Fusi
db895209ca added (optional) iter param dump 2013-03-26 11:49:20 +00:00
Nicolo Fusi
c9709cf4da fixed small bug in SGD 2013-03-25 18:02:42 +00:00
Ricardo Andrade
be828ad38b Insignificant but annoying bug corrected 2013-03-22 19:04:32 +00:00
Ricardo Andrade
35a099f4ed minor changes 2013-03-22 19:03:46 +00:00
Ricardo Andrade
b007f7cda6 Merge branch 'master' into fitc 2013-03-22 18:44:54 +00:00
Ricardo Andrade
8dd33615c1 pseudo EM algorithm for EP and maybe Laplace. 2013-03-22 18:38:24 +00:00
Ricardo Andrade
a375cb9a1b generalized fitc + examples 2013-03-22 16:04:51 +00:00
Nicolo Fusi
55ad96f38b made BGPLVM oil flow demo work, added ARD weights plot 2013-03-22 15:58:02 +00:00
Alan Saul
7b0d0550cb Seemed to be working, now its not 2013-03-22 12:50:47 +00:00
Nicolo Fusi
eeb965d136 added BGPLVM oil flow demo and changed default X_variance init 2013-03-21 15:28:46 +00:00
Alan Saul
474d5484b0 Changing definitions again... 2013-03-21 14:00:22 +00:00
Nicolò Fusi
8ab1cfaf65 changes in GPLVM plotting 2013-03-21 11:05:17 +00:00
Ricardo Andrade
0c0c7b000c Generalized FITC is back 2013-03-19 19:08:25 +00:00