Biomod2 r package: selecting different cross validation and PA datasets for each type of modeling algorithm

Question

I am trying to create a species distribution model using Random Forest and Maxent algoritms in biomod2.

Based on Barbet-Massin et al 2012, I want to use a balanced number of pseudo-absence (PA) points for the RF model (which would be about 65 PA = to my presence points) with 10 runs and 10000 PA for the maxent model.

Additionally, is there a way to create seperate cross validation procedures depending on the type of model I am running? Since I am using two different model types in biomod2, I want a geographically structured cross validation model for maxent, but not for RF, since as far as I can tell cross validation such as this is unnessary for RF -- though I may totally be wrong and would appreciate further thoughts on this!

I am hoping to ensemble these models together, and I have not found a sound way of doing this when creating seperate biomod datasets using BIOMOD_FormatingData().

myBiomodData <- BIOMOD_FormatingData(
    resp.var = myResp.bij,
    expl.var = myExpl.bij,
    resp.xy = RespXY.bij,
    resp.name = RespName.bij,
    PA.strategy = 'disk',
    PA.dist.min = 5,  
    PA.dist.max = 222000, # ~2 degrees
    PA.nb.absence = c(10000, 65), # RF = PA:P
    PA.nb.rep = 10,
    na.rm = TRUE
)

myBiomodModelOut <- BIOMOD_Modeling(
    myBiomodData,
    models = c('RF', 'MAXENT'),   
    bm.options = myBiomodOption,
    prevalence = 0.5,
    var.import = 3, 
    metric.eval = c('TSS', 'ROC'),
    CV.nb.rep = 2, # Two different types of CV datasets
    CV.perc = 0.8,
    CV.do.full.models = FALSE,
    modeling.id = paste('spname', "bmout", sep = "")
)