I can’t seem to figure out how to use the CuArray to compute the gradient of the same function f at many points x. E.g. using Zygote:gradient f(x) = x^2 x = rand(100) gradient.(f, x) the above works fine becuase it’s CPU code But the below doesn’t work using CuArrays allowscalar(false) x = CuA…

How compute `gradient.(f, w)` on GPU?

xiaodai November 16, 2019, 8:06am 3

Can’t quite get this to work on th GPU unless I differentiate the function myself.

Topic		Replies	Views
Gradient of `sum(f, x)` using Zygote, when `f` is real valued and `x isa CuArray{<:Complex}` GPU question , differentiation , zygote	1	746	January 8, 2020
Mapping functions using CuArrays GPU cuarrays	2	1830	September 9, 2020
Gradient of the einsum not working with CuArrays, when it's output is assigned to the field of a mutable struct GPU machine-learning	0	512	February 10, 2020
CuArray and Optim GPU optim , optimization	9	3318	September 4, 2018
Gradient of sum of singular values of a matrix with CUDA.jl GPU question , zygote	7	1005	June 8, 2022