I wasn't sure if batch matmul was supported, as this seems to be documented nowhere except in pytorch's documentation. It seems to work fine with small tensors but breaks down past a certain size: So ...

cNMF is a pipeline for inferring gene expression programs from scRNA-Seq. It takes a count matrix (N cells X G genes) as input and produces a (K x G) matrix of gene expression programs (GEPs) and a (N ...