Hello, I am using L-BFGS to solve a maximum likelihood problem. I provide analytical gradients to the solver which is nice. However, many steps I use in the function calculation can be repeated in the gradient calculation. In fact, it’s possible to calculate these within one function call which ret…

I think this is what you need data = DataFrame(x = repeat([1,2], 500), y = sample([0,1], 1000)) function fg!(F, G, beta) Prob = SharedArray{T}(1000, 1) z = @distributed (+) for i in 1:1000 if data[!, :y][i] .== 1 Prob[i] = exp(beta[1] .* data[!, :x][i])/(1 + exp(beta[1] .* data[…

Possible to used cached values in Optim.optimize?

General Usage

bashonubuntu November 30, 2019, 11:23am 18

Perfect, thanks.

Topic		Replies	Views
Is there an equivalent to Optim.jl's only_fg! in Optimization.jl? Optimization (Mathematical) question	1	111	December 8, 2025
Gradient-based optimization when gradient is only a function of y = f(x) Optimization (Mathematical) optim , optimization	3	276	May 1, 2024
Optim.jl - computing value, gradient and hessian simultaneously Optimization (Mathematical)	17	3994	August 28, 2018
Optim user-supplied gradient not working Optimization (Mathematical) optim	6	1095	May 21, 2019
Providing Objective, Gradient and Hessian to the Ipopt wrapper using a single function? Optimization (Mathematical) optimization	12	1341	April 22, 2021

Possible to used cached values in Optim.optimize?

Related topics