def run_bohamiann(self, iterations=20):
cs = self.__class__.get_configspace()
lower, upper = get_upper_lower(cs)
results = bohamiann(self.compute, lower, upper, num_iterations=iterations, cs=cs)
x_best = results["x_opt"]
config = get_config_dictionary(x_best, cs)
print(config)
def run_es(self, iterations=20):
cs = self.__class__.get_configspace()
lower, upper = get_upper_lower(cs)
if not os.path.exists(self.run_dir):
os.mkdir(self.run_dir)
log_dir = os.path.join(self.run_dir, f'EXP{self.experiment_no}')
if not os.path.exists(log_dir):
os.mkdir(log_dir)
results = entropy_search(self.compute, lower, upper, num_iterations=iterations, cs=cs, min_budget=self.min_budget, max_budget=self.max_budget)
save_results_optimisation(results, log_dir)
x_best = results["x_opt"]
self.experiment_no += 1
print(x_best)
def run_es(self, iterations=20):
cs = self.__class__.get_configspace()
lower, upper = get_upper_lower(cs)
results = entropy_search(self.compute,
lower,
upper,
num_iterations=iterations,
cs=cs,
min_budget=2,
max_budget=5)
x_best = results["x_opt"]
config = get_config_dictionary(x_best, cs)
print(config)
def run_bohamiann(self, iterations=20):
cs = self.__class__.get_configspace()
lower, upper = get_upper_lower(cs)
results = bohamiann(self.compute,
lower,
upper,
num_iterations=iterations,
cs=cs)
if not os.path.exists(self.run_dir):
os.mkdir(self.run_dir)
log_dir = os.path.join(self.run_dir, f'EXP{self.experiment_no}')
if not os.path.exists(log_dir):
os.mkdir(log_dir)
save_results_optimisation(results, log_dir)
x_best = results["x_opt"]
self.experiment_no += 1
print('Best run', x_best)
return x_best