Skip to content

README.md

Latest commit

 

History

History
42 lines (40 loc) · 1013 Bytes

README.md

File metadata and controls

42 lines (40 loc) · 1013 Bytes
Raw

Gradient-GP-Regression based optimizer

install:

git clone https://github.com/SingletC/gpopt.git
cd gpopt
python -m pip install .

Verifation of installation (optional):

python -m gpopt.test

usage:

from gpopt.optimizer import GPOPT
from gpopt.utils import func_with_grad # your function here
import numpy as np
# Sig: func_with_grad: Callable[[np.ndarray], Tuple[float, np.ndarray]]
x0 = np.random.rand(100)
opt = GPOPT(func_with_grad, x0,tol = 1e-2)
xf = opt.optimize()
# dump opt for debug
import pickle
with open('opt.pkl', 'wb') as f:
    pickle.dump(opt, f)

advanced usage:

step by step optimization:

from gpopt.optimizer import GPOPT
from gpopt.utils import func_with_grad
import numpy as np
import pickle
x0 = np.random.rand(100)
opt = GPOPT(func_with_grad, x0,tol = 1e-2)
for i in range(5):
    opt.step()
    ### do thing here. e.g., dump the opt for debug.
    with open(f'opt_{i}.pkl', 'wb') as f:
        pickle.dump(opt, f)