You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Although it's great to have ML-specific functionality held separately from base CuArrays in this package, I'm wondering if we should keep convolutions in CuArrays itself (effectively making it part of the standard library). Image processing is a big use case for GPUs, and it'd be particularly nice if CuArrays and ImageFiltering.jl worked out of the box together.
This wouldn't imply any hard dependency on CUDNN for CuArrays, only the soft dep that we already have (and we could have slower backup kernels when CUDNN isn't available).
Although it's great to have ML-specific functionality held separately from base CuArrays in this package, I'm wondering if we should keep convolutions in CuArrays itself (effectively making it part of the standard library). Image processing is a big use case for GPUs, and it'd be particularly nice if CuArrays and ImageFiltering.jl worked out of the box together.
This wouldn't imply any hard dependency on CUDNN for CuArrays, only the soft dep that we already have (and we could have slower backup kernels when CUDNN isn't available).