Use correct output folder.
Nevermind don't include trunk. Silly idea...
Replace 10.1 with trunk for windows builds.
Fix broken build.
Remove confusing 10.x version number
Re-add adlink and adcomp with correct permissions.
Remove adcomp and adlink. Will re-add them with correct permissions.
fix build for Open MPI
Remove example documentation files that are redundant or not useful
Merge from trunk -r 72:262 into branch
change to configure script, adcomp, adlink
Minor improvements in adcomp/adlink/admb for Linux
Change to adlink, adcomp
Change to configure
Change to configure and make file to work with more OSes
Just to test branches on buildbot.
[Issue #26] Added adstring_array equal operator.
[Issue #27] Added missing vc10 scripts and improve installation.
[Issue #2] Allow users in Visual C++ to keypress q,c and n without first pressing Ctrl-C.
Opps added orange twice...
[Bug #22] Use Hugo Mildenbergers fixes that removed warnings.
[Support #13] Location of developers' documentation for coding standards.
[Feature #2] Ctrl-C now working with Microsoft C++ compiler.
[Bug #6] Hard code the command to build for macos10.5 with snow leopard.
Get ready for 10.2
Add XCode 4 to list of supported compilers.
[Bug #9] Should ignore line feed character '\r' when reading data.
Update build instructions.
Fix broken builds from previous commit.
[Bug #10] uses compiler defaults and clean up admb script.
Minor makefile chnage
Configure changes. Small change so hessian comp for non-RE works.
Forgot to add two files.
Changes to configure to work with Open MPI.
Do not use svnversion in makefile because it may not be installed.
Bug fix by fournier see http://lists.admb-project.org/pipermail/developers/2012-January/000507.html
update glmmadmb.tpl; further robustification, maxphase -> 5 from DF
added phi (fixed-effect orthogonalization matrix) output to glmmadmb.tpl
code clean up
Opps forgot to clean statslib.
Issue 3: added contrib to the distribution and added included them in the admb scripts.
Profile likelihood for RE models from Dave
Change to mod_sd.cpp required to remove warning from glmmadmb.tpl
Changes to glmmadmb.tpl
Combined dnbinom.cpp and dnbinom_eco.cpp, all now named dnbinom. Added dzinbinom.cpp zero inflated negative binomial negative log likelihood.
Changes to df1b2fnl.cpp xmodelm3.cpp needed for new robust glmmadmb.tpl
Fournier fixes for ivsort.cpp.
Update of glmmadmb.tpl
fix latex error in dox for dnorm.cpp and dlnorm.cpp
update of glmmadmb.tpl
Fix valgrind reported uninitialised variable.
Fix valgrind reported memory leak.
Minor corrections in ADMB-RE bibliography
Add how to cite to -info and README.txt.
set svn:keywords Id on .tex files
Changes to ADMB-Re manual
Fix const compile errors.
Fixed statsLib.h so that it contains the function and class declarations for the implemented functions. Removed redundant #include statements. Added some functions including vcubic_spline, dunif, dinvgamma, dmvlogistic, eplogis, mutifan, and rmvlogistic. Created makefiles for g++ and Visual studio (cl) compilers. These have both been tested on Windows. g++ works through mingw comand shell and cl works through the Visual studio command prompt. Still need to get dbinom_eco and dbinom_eco_zi working.
addedthe truncated negative binomial distribution
Update glmmadmb tpl.
fix 4Gb limit for -cbs in df_file.cpp
Fix nested4 for parallel.
Added nested4 example.
Set output from master only.
RE Hessian and Covariance now correct. Output files now correct. Only minor changes to follow.
Improved separable model. RE now converges for multiple slaves.
consolidated mpi examples
Added another sparable example. Added code to coordinate convergance within get_uhat_quasi_newton_block_diagonal
slight change so hessian comp of non-RE models still works
made changes to the hessian for RE. Fixes hessian complaint for 2 slaves. Still doesn't converge for 2 slaves.
df1b2gradlist suffixes. non-RE gradient sync.
Fix Bug#11 tpl2rem should work with CR+LF line endings.
Prepend 10.1 instead of 10.0
Fix typo Documentation#13
Update 10.0 to 10.1
Added function definition for solve(dvar_matrix,dvar_matrix) to fvar.hpp such that the function can be called from outside expm.cpp - Useful for inverting matrices in random effects programs.
Don't run nbmm for vc10-32bit.
Update hard coded version number.
Changed name of separable_bounds class to ad_separable_manager. Improved so the bounds are calculated only once through the life of the program.
Update vc build path.
Fixes for borland build.
sycronized inner maxg amoung slaves and master. Slaves don't print inner maxg
« Previous 1 ... 11 12 13 14 15 Next » (1201-1300/1474) | Per page: 25, 50, 100
Also available in: Atom