[FORGED] Re: Error from glmmTMB() --- no joy.
Yeah, I realized today that I was wrong about this fixing the problem. I'm not stuck yet/want to work on this, but I will pursue it in the context of a glmmTMB issue. (I like everyone else am a little busier than usual these days ...) Ben
On 2020-03-19 8:17 p.m., Rolf Turner wrote:
On 19/03/20 3:52 pm, Ben Bolker wrote: <SNIP>
I made a small change that appears to allow the
original model to run correctly.? If you want to try it out, and have
development tools (compilers etc.) and the remotes package installed,
remotes::install_github("glmmTMB/glmmTMB/glmmTMB at cloglog_fix")
?? should install the patched version (I think).
This will presumably be merged with the master branch sometime soon but
I want to add some tests etc.
I've finally got around to trying out your patch.? No problems with the install_github() process, but after doing the install I executed:
library(glmmTMB)
fmla <- cbind(Dead, Alive) ~ (Trt + 0)/Dose + (Dose | Rep)
X??? <- dget("X.txt")
fit? <- glmmTMB(fmla, data = X, family = betabinomial(link = "cloglog"),
?????????????? dispformula = ~1
and got:
Error in (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN gradient evaluation In addition: Warning messages: 1: In (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN function evaluation 2: In (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN function evaluation 3: In (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN function evaluation 4: In (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN function evaluation 5: In (function (start, objective, gradient = NULL, hessian = NULL,? : ? NA/NaN function evaluation Timing stopped at: 0.572 0 0.572
So my Sadim touch persists in its effect. cheers, Rolf P. S. With the logit link I got warnings
1: In (function (start, objective, gradient = NULL, hessian = NULL,? :
? NA/NaN function evaluation
2: In fitTMB(TMBStruc) :
? Model convergence problem; non-positive-definite Hessian matrix. See
vignette('troubleshooting')
but no errors as such. R.