def fit(self, pipeline_config, run_result_dir):
instance_file_config_parser = ConfigFileParser([
ConfigOption(name='path', type='directory', required=True),
ConfigOption(name='is_classification', type=to_bool,
required=True),
ConfigOption(name='is_multilabel', type=to_bool, required=True),
ConfigOption(name='num_features', type=int, required=True),
ConfigOption(name='categorical_features',
type=bool,
required=True,
list=True),
ConfigOption(
name='instance_shape',
type=[ast.literal_eval, lambda x: isinstance(x, tuple)],
required=True)
])
instance_info = instance_file_config_parser.read(
os.path.join(run_result_dir, 'instance.info'))
instance_info = instance_file_config_parser.set_defaults(instance_info)
dm = DataManager()
if instance_info["is_multilabel"]:
dm.problem_type = ProblemType.FeatureMultilabel
elif instance_info["is_classification"]:
dm.problem_type = ProblemType.FeatureClassification
else:
dm.problem_type = ProblemType.FeatureClassification
return {'instance_info': instance_info, 'data_manager': dm}
def fit(self, pipeline_config, instance):
assert pipeline_config['problem_type'] in [
'feature_classification', 'feature_multilabel',
'feature_regression'
]
dm = DataManager(verbose=pipeline_config["data_manager_verbose"])
dm.read_data(instance,
is_classification=(pipeline_config["problem_type"] in [
'feature_classification', 'feature_multilabel'
]),
test_split=pipeline_config["test_split"])
return {"data_manager": dm}
__version__ = "0.0.1"
__license__ = "BSD"
import os, sys
sys.path.append(os.path.abspath(os.path.join(__file__, "..", "..", "..")))
import logging
from autoPyTorch import AutoNetClassification, AutoNetMultilabel
import autoPyTorch.pipeline.nodes as autonet_nodes
from autoPyTorch.components.metrics.additional_logs import test_result
import autoPyTorch.components.metrics as autonet_metrics
from autoPyTorch.data_management.data_manager import DataManager
dm = DataManager(verbose=1)
dataset_dir = os.path.abspath(os.path.join(os.path.dirname( __file__ ), '..', '..', 'datasets'))
# choose between the 5 defined testcases
TEST_CASE = 1
""" TEST CASE 1: Sparse data """
if TEST_CASE == 1:
dm.read_data(os.path.join(dataset_dir, "automl/newsgroups/newsgroups_public.info"), is_classification=True)
metric = "pac_metric"
additional_metrices = ["accuracy"]
""" TEST CASE 2: Sparse binary data """
if TEST_CASE == 2:
dm.read_data(os.path.join(dataset_dir, "automl/dorothea/dorothea_public.info"), is_classification=True)
metric = "auc_metric"
__author__ = "Max Dippel, Michael Burkart and Matthias Urban" __version__ = "0.0.1" __license__ = "BSD" import os import sys import logging from autoPyTorch import AutoNetClassification from autoPyTorch.data_management.data_manager import DataManager from autoPyTorch.utils.mem_test_thread import MemoryLogger dm = DataManager(verbose=1) dataset_dir = os.path.abspath(os.path.join(os.path.dirname( __file__ ), '..', 'datasets')) dm.read_data(os.path.join(dataset_dir, "classification/dataset_28_optdigits.csv"), is_classification=True) # 5620 samples, 10 classes, 65 features ---> 98% validation accuracy mem_logger = MemoryLogger() mem_logger.start()
__author__ = "Max Dippel, Michael Burkart and Matthias Urban"
__version__ = "0.0.1"
__license__ = "BSD"
import os, sys
sys.path.append(os.path.abspath(os.path.join(__file__, "..", "..", "..")))
from autoPyTorch import AutoNetClassification
from autoPyTorch.data_management.data_manager import DataManager
# Note: You can write your own datamanager! Call fit with respective train, valid data (numpy matrices)
dm = DataManager()
dm.generate_classification(num_classes=3, num_features=21, num_samples=1500)
# Note: every parameter has a default value, you do not have to specify anything. The given parameter allow a fast test.
autonet = AutoNetClassification(budget_type='epochs',
min_budget=1,
max_budget=9,
num_iterations=1,
log_level='info')
res = autonet.fit(X_train=dm.X,
Y_train=dm.Y,
X_valid=dm.X_train,
Y_valid=dm.Y_train)
print(res)
print("Score:", autonet.score(X_test=dm.X_train, Y_test=dm.Y_train))
__author__ = "Max Dippel, Michael Burkart and Matthias Urban"
__version__ = "0.0.1"
__license__ = "BSD"
import os, sys
sys.path.append(os.path.abspath(os.path.join(__file__, "..", "..", "..")))
from autoPyTorch import AutoNetRegression
from autoPyTorch.data_management.data_manager import DataManager
# Note: You can write your own datamanager! Call fit train, valid data (numpy matrices)
dm = DataManager()
dm.generate_regression(num_features=21, num_samples=1500)
X_train = dm.X
Y_train = dm.Y
X_valid = dm.X_train
Y_valid = dm.Y_train
# Note: every parameter has a default value, you do not have to specify anything. The given parameter allow a fast test.
autonet = AutoNetRegression(budget_type='epochs',
min_budget=1,
max_budget=9,
num_iterations=1,
log_level='info')
res = autonet.fit(X_train=X_train,
Y_train=Y_train,
X_valid=X_valid,
Y_valid=Y_valid)
print(res)
os.mkdir(outputs_folder)
if not os.path.exists(output_base_dir):
os.mkdir(output_base_dir)
if not os.path.exists(result_dir):
os.mkdir(result_dir)
# iterate over all runs
for run_number in runs_range:
replacement_dict["RUN_NUMBER"] = run_number
for config_id in configs_range:
replacement_dict["CONFIG_ID"] = config_id
replacement_dict["CONFIG_FILE"] = all_configs[config_id]
# get autonet
dm = DataManager()
dm.problem_type = {
"feature_classification": ProblemType.FeatureClassification,
"feature_multilabel": ProblemType.FeatureMultilabel,
"feature_regression": ProblemType.FeatureRegression
}[benchmark_config["problem_type"]]
autonet = CreateAutoNet().fit(benchmark_config, dm)["autonet"]
autonet_config_file = benchmark_config["autonet_configs"][
config_id]
for instance_id in instances_range:
replacement_dict["INSTANCE_ID"] = instance_id
replacement_dict["INSTANCE_FILE"] = all_instances[instance_id]
# read autonet config
SetAutoNetConfig().fit(benchmark_config, autonet,
def fit(self, pipeline_config, instance):
# Get data manager for train, val, test data
if pipeline_config['problem_type'] in ['feature_classification', 'feature_multilabel', 'feature_regression']:
dm = DataManager(verbose=pipeline_config["data_manager_verbose"])
if pipeline_config['test_instances'] is not None:
dm_test = DataManager(verbose=pipeline_config["data_manager_verbose"])
else:
dm = ImageManager(verbose=pipeline_config["data_manager_verbose"])
if pipeline_config['test_instances'] is not None:
dm_test = ImageManager(verbose=pipeline_config["data_manager_verbose"])
# Read data
if pipeline_config['test_instances'] is not None:
# Use given test set
dm.read_data(instance,
is_classification=(pipeline_config["problem_type"] in ['feature_classification', 'feature_multilabel', 'image_classification']),
test_split=0.0)
dm_test.read_data(pipeline_config['test_instances'],
is_classification=(pipeline_config["problem_type"] in ['feature_classification', 'feature_multilabel', 'image_classification']),
test_split=0.0)
dm.X_test, dm.Y_test = dm_test.X_train, dm_test.Y_train.astype(np.int32)
else:
# Use test split
dm.read_data(instance,
is_classification=(pipeline_config["problem_type"] in ['feature_classification', 'feature_multilabel', 'image_classification']),
test_split=pipeline_config["test_split"])
return {"data_manager": dm}
__author__ = "Max Dippel, Michael Burkart and Matthias Urban"
__version__ = "0.0.1"
__license__ = "BSD"
import os, sys
sys.path.append(os.path.abspath(os.path.join(__file__, "..", "..", "..")))
import logging
from autoPyTorch import AutoNetClassification
from autoPyTorch.data_management.data_manager import DataManager
dm = DataManager(verbose=1)
dataset_dir = os.path.abspath(
os.path.join(os.path.dirname(__file__), '..', '..', 'datasets'))
# choose between the 10 classification testcases on real data.
TEST_CASE = 4
if TEST_CASE == 1:
dm.read_data(os.path.join(dataset_dir,
"classification/dataset_22_mfeat-zernike.csv"),
is_classification=True)
# 2000 samples, 10 classes, 48 features
if TEST_CASE == 2:
dm.read_data(os.path.join(dataset_dir, "classification/phpbL6t4U.csv"),
is_classification=True)
# 13910 samples, 6 classes, 128 features
if TEST_CASE == 3:
__author__ = "Max Dippel, Michael Burkart and Matthias Urban"
__version__ = "0.0.1"
__license__ = "BSD"
import os, sys
sys.path.append(os.path.abspath(os.path.join(__file__, "..", "..", "..")))
import logging
from autoPyTorch import AutoNetClassification
from autoPyTorch.data_management.data_manager import DataManager
dm = DataManager(verbose=1)
dataset_dir = os.path.abspath(
os.path.join(os.path.dirname(__file__), '..', '..', 'datasets'))
# choose between the 10 classification testcases on real data.
TEST_CASE = 4
if TEST_CASE == 1:
dm.read_data("openml:22", is_classification=True)
# 2000 samples, 10 classes, 48 features
if TEST_CASE == 2:
dm.read_data("openml:1476", is_classification=True)
# 13910 samples, 6 classes, 128 features
if TEST_CASE == 3:
dm.read_data("openml:1464", is_classification=True)
# 748 samples, 2 classes, 4 features
