The cumprod
rule with CUDA hits scalar indexing error
#1447
Labels
ChainRules
adjoint -> rrule, and further integration
Hi, it seems like the rule is not properly kicking in for CUDA.
The text was updated successfully, but these errors were encountered: