def train_breast_cancer(config):
params = LGBMEstimator(**config).params
X_train = ray.get(X_train_ref)
train_set = lgb.Dataset(X_train, label=y_train)
gbm = lgb.train(params, train_set)
preds = gbm.predict(X_test)
pred_labels = np.rint(preds)
tune.report(mean_accuracy=accuracy_score(y_test, pred_labels), done=True)
def evaluate_config(config):
"""evaluate a hyperparameter configuration"""
# we uss a toy example with 2 hyperparameters
metric = (round(config["x"]) - 85000)**2 - config["x"] / config["y"]
# usually the evaluation takes an non-neglible cost
# and the cost could be related to certain hyperparameters
# in this example, we assume it's proportional to x
time.sleep(config["x"] / 100000)
# use tune.report to report the metric to optimize
tune.report(metric=metric)
def simple_obj(config, resource=10000):
config_value_vector = np.array([config["x"], config["y"], config["z"]])
score_sequence = []
for i in range(resource):
a = rand_vector_unit_sphere(3)
a[2] = abs(a[2])
point_projection = np.dot(config_value_vector, a)
score_sequence.append(point_projection)
score_avg = np.mean(np.array(score_sequence))
score_std = np.std(np.array(score_sequence))
score_lb = score_avg - 1.96 * score_std / np.sqrt(resource)
tune.report(samplesize=resource, sphere_projection=score_lb)
def obj_w_intermediate_report(resource, config):
config_value_vector = np.array([config["x"], config["y"], config["z"]])
score_sequence = []
for i in range(resource):
a = rand_vector_unit_sphere(3)
a[2] = abs(a[2])
point_projection = np.dot(config_value_vector, a)
score_sequence.append(point_projection)
if (i + 1) % 100 == 0:
score_avg = np.mean(np.array(score_sequence))
score_std = np.std(np.array(score_sequence))
score_lb = score_avg - 1.96 * score_std / np.sqrt(i + 1)
tune.report(samplesize=i + 1, sphere_projection=score_lb)
def cost(param):
tune.report(loss=(param["height"] - 14) ** 2 - abs(param["width"] - 3))
def simple_func(config):
tune.report(metric=(config["cost_related"]["a"] - 4)**2 *
(config["b"] - 0.7)**2)
def simple_func(config):
obj = (config["cost_related"]["a"] - 4)**2 \
+ (config["b"] - config["cost_related"]["a"])**2
tune.report(obj=obj)
tune.report(obj=obj, ab=config["cost_related"]["a"] * config["b"])