Before benchmarking a likelihood minimization, the model parameters need to be reset to something away from the parameters that were used to generate the toy dataset. Otherwise, the fit is starting already at the minimium and we won't benchmark the actual minimization path.
Furthermore, the parameters need to be reset after each benchmark iteration.
Before benchmarking a likelihood minimization, the model parameters need to be reset to something away from the parameters that were used to generate the toy dataset. Otherwise, the fit is starting already at the minimium and we won't benchmark the actual minimization path.
Furthermore, the parameters need to be reset after each benchmark iteration.