I trained a causual forest with 1 million * 115 samples, then I saved model with saveRDS. However, the model file's size is about 5G.When I predicted, reading modelfile and predicting needed 20G memory.
Is there any wan to reduce the size of model which can support forecast?
I trained a causual forest with 1 million * 115 samples, then I saved model with
saveRDS
. However, the model file's size is about 5G.When I predicted, reading modelfile and predicting needed 20G memory.Is there any wan to reduce the size of model which can support forecast?