Potential Parallelism for Hyperparameter Tuning
cw-tan opened this issue · comments
This post is in reference to the function to optimize hyperparameters:
flare/flare/bffs/sgp/sparse_gp.py
Line 627 in 700f2c8
Since scipy.minimize is used for the optimization, we could potentially use the minimize_parallel function from this project for a parallel version of scipy.minimize's L-BFGS-B option.