IntegrateLayers() function cannot handle non-zero counts larger than 2^31-1

[Qinmengge]

Dear Seurat team,
I have been trying to use the sketch-based analysis on a 149 sample study with around 1.5 million cells. I know that the sparse matrix objects in R cannot handle non-zero values larger than 2^31-1, and I thought the sketch-based framework had a way of walking around this problem. However I observed the following error when performing IntegrateLayers() function:

My guess from the error message is that the problem is still from the R sparse matrix overflow. I am curious is there a way to solve this issue by sketch-based analysis?
Thanks a lot!