This vignette covers common problems that occur while using `glmmTMB`

. The contents will expand with experience.

If your problem is not covered below, try updating to the latest version of `glmmTMB`

on GitHub. The developers might have solved the problem in a newer version.

You may see the same warning as in the following example:

`zinbm0 = glmmTMB(count~spp + (1|site), zi=~spp, Salamanders, family="nbinom2")`

```
## Warning in fitTMB(TMBStruc): Model convergence problem; non-positive-
## definite Hessian matrix. See vignette('troubleshooting')
```

This often occurs when a model is overparameterized (i.e. the data does not contain information to estimate the parameters). In the example above, the problem is that zero observations do not depend on species as in `zi=~spp`

; rather, they depend on another covariate that is missing from the model `minded`

. Plotting the data against potential covariates should help to avoid fitting unreasonable models that don’t converge.

In some cases, scaling predictor variables may help.

Models with non-positive definite Hessian matricies should be excluded from further consideration, in general.

`m1 = glmmTMB(count~spp + mined + (1|site), zi=~spp + mined, Salamanders, family="genpois")`

```
## Warning in fitTMB(TMBStruc): Model convergence problem; extreme or very
## small eigen values detected. See vignette('troubleshooting')
```

In this example, the fixed effect covariance matrix is `NaN`

. It may have to do with the generalized Poisson (`genpois`

) distribution, which is known to have convergence problems; luckily, the negative binomial (`nbinom1`

and `nbinom2`

) and/or Conway-Maxwell Poisson (`compois`

) are good alternatives.

Models with convergence problems should be excluded from further consideration, in general.

```
Warning in nlminb(start = par, objective = fn, gradient = gr) :
NA/NaN function evaluation
```

This warning often occurs when the optimizer wanders into a region of parameter space that is invalid. It is not a problem as long as the optimizer has left that region of parameter space upon convergence, which is indicated by an absence of the model convergence warnings described above.

The following warnings can be treated in the same way as an NA/NaN function evaluation:

`Cholmod warning 'matrix not positive definite'`

`Warning in f(par, order = order, ...) : value out of range in 'lgamma'`

```
dat1 = expand.grid(y=-1:1, rep=1:10)
m1 = glmmTMB(y~1, dat1, family="nbinom2")
```

`## Error in nlminb(start = par, objective = fn, gradient = gr, control = control$optCtrl): NA/NaN gradient evaluation`

`## Timing stopped at: 0.003 0 0.003`

The error occurs here because the negative binomial distribution is not appropriate for data with negative values.

If you see this error, check that the response variable meets the assumptions of the specified distribution.

```
Error in nlminb(start = par, objective = fn, gradient = gr) :
gradient function must return a numeric vector of length x
```

```
Error in optimHess(par.fixed, obj$fn, obj$gr):
gradient in optim evaluated to length x
```

Try rescaling predictor variables. Try a simpler model and build up.