Predicting on a glmmLasso model

I'm trying to use the glmmLasso pacakage and am able to run a model and get a summary output and extract the fitted values, but I'm wondering if/how I can make predictions with it?

Am I missing a step?

For example:

library(glmmLasso)
library(tidyverse)

mt_tbl <- mtcars %>% as_tibble() %>% 
  mutate(cyl = factor(cyl))

glm_cars <- glmmLasso(mpg ~  hp + drat + wt,
                        data=mt_tbl,
                        rnd = list(cyl=~1),
                        family = gaussian(link = "identity"),
                        lambda = .9,
                        switch.NR = TRUE,
                        final.re = TRUE)

# These work
summary(glm_cars)
glm_cars$fitted.values

# I want this to work
predict(glm_cars, mt_tbl)

If you do not convert mtcars to tibble, the code works fine with data.frame like

library(glmmLasso)
library(tidyverse)

#Convert cyl to factor
mtcars$cyl <- factor(mtcars$cyl)

#Run the model using 'data.frame'
glm_cars <- glmmLasso(mpg ~  hp + drat + wt,
                      data=mtcars,
                      rnd = list(cyl=~1),
                      family = gaussian(link = "identity"),
                      lambda = .9,
                      switch.NR = TRUE,
                      final.re = TRUE)

# These work
summary(glm_cars)
glm_cars$fitted.values

# Predict using 'data.frame' works
predict(glm_cars, mtcars)