from deephyper.problem import NaProblem
from nas_big_data.combo.load_data import load_data_cache
from nas_big_data.combo.search_space_shared import create_search_space
Problem = NaProblem(seed=2019)
Problem.load_data(load_data_cache)
Problem.search_space(create_search_space, num_layers=5)
# schedules: https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/schedules
Problem.hyperparameters(
lsr_batch_size=True,
lsr_learning_rate=True,
batch_size=Problem.add_hyperparameter((16, 2048, "log-uniform"),
"batch_size"),
learning_rate=Problem.add_hyperparameter(
(1e-4, 0.01, "log-uniform"),
"learning_rate",
),
optimizer=Problem.add_hyperparameter(
["sgd", "rmsprop", "adagrad", "adam", "adadelta", "adamax", "nadam"],
"optimizer"),
patience_ReduceLROnPlateau=Problem.add_hyperparameter(
(3, 30), "patience_ReduceLROnPlateau"),
patience_EarlyStopping=Problem.add_hyperparameter(
(3, 30), "patience_EarlyStopping"),
num_epochs=100,
verbose=0,
callbacks=dict(
ReduceLROnPlateau=dict(monitor="val_r2",
from deephyper.benchmark.nas.linearReg.load_data import load_data
from deephyper.problem import NaProblem
from deepspace.tabular import OneLayerSpace
Problem = NaProblem()
Problem.load_data(load_data)
Problem.search_space(OneLayerSpace)
Problem.hyperparameters(
batch_size=Problem.add_hyperparameter((1, 100), "batch_size"),
learning_rate=Problem.add_hyperparameter((1e-4, 1e-1, "log-uniform"),
"learning_rate"),
optimizer=Problem.add_hyperparameter(["adam", "nadam", "rmsprop"],
"optimizer"),
num_epochs=1,
)
Problem.loss("mse")
Problem.metrics(["r2"])
Problem.objective("val_r2")
# Just to print your problem, to test its definition and imports in the current python environment.
if __name__ == "__main__":
print(Problem)
model = Problem.get_keras_model([1])
