cuDNN.jl was split into a separate package (that lives in the CUDA.jl repository) so to use it you must now use using cuDNN after adding it to your envoronment just like you would with CUDA.jl
2 Likes
cuDNN.jl was split into a separate package (that lives in the CUDA.jl repository) so to use it you must now use using cuDNN after adding it to your envoronment just like you would with CUDA.jl