cotengra.hyperoptimizers.hyper_nevergrad

Hyper optimization using nevergrad.

Classes

NevergradOptLib

Hyper-optimization using nevergrad.

Functions

convert_param_to_nevergrad(param)
get_methods_space(methods)
convert_to_nevergrad_space(method, space)

Module Contents

cotengra.hyperoptimizers.hyper_nevergrad.convert_param_to_nevergrad(param)

cotengra.hyperoptimizers.hyper_nevergrad.get_methods_space(methods)

cotengra.hyperoptimizers.hyper_nevergrad.convert_to_nevergrad_space(method, space)

class cotengra.hyperoptimizers.hyper_nevergrad.NevergradOptLib

Bases: cotengra.hyperoptimizers.hyper.HyperOptLib

Hyper-optimization using nevergrad.

setup(methods, space, optimizer=None, sampler='NaiveTBPSA', method_sampler=None, budget='auto', num_workers=1, method_budget='auto', method_num_workers=1, sampler_opts=None, method_sampler_opts=None)

Initialize the nevergrad optimizer.

Parameters:

• methods (list[str]) – The list of contraction methods to optimize over.

• space (dict[str, dict[str, dict]]) – The search space.

• optimizer (HyperOptimizer, optional) – The parent optimizer instance, used for max_repeats and _num_workers when budget or num_workers are 'auto'.

• sampler (str, optional) – The optimizer to use to search each method’s search space, see nevergrad docs.

• method_sampler (str, optional) – The meta-optimizer to use to select overall methods.

• budget (int, optional) – Supplied to optimizer.

• num_workers (int, optional) – Supplied to optimizer.

• method_budget (int, optional) – Supplied to meta-optimizer.

• method_num_workers (int, optional) – Supplied to meta-optimizer.

get_setting()

Get a setting to trial from one of the nevergrad optimizers.

report_result(setting, trial, score)

Report the result of a trial to the nevergrad optimizers.