autogluon.core¶

Decorators for customizing AutoGluon to apply hyperparameter-tuning on arbitrary user-defined objects and functions.

Toy Example

Create class and function with searchable spaces for hyperparameters name and idx:

>>> import autogluon as ag
>>> @ag.obj(
...     name=ag.space.Categorical('auto', 'gluon'),
...     idx = ag.space.Int(0, 100),
... )
>>> class myobj:
...     def __init__(self, name, idx):
...         self.name = name
...         self.idx = idx
...
>>> @ag.func(framework=ag.space.Categorical('mxnet', 'pytorch'))
>>> def myfunc(framework):
...     return framework

Create the objects using decorated class and function as argument of autogluon.args

>>> @ag.args(
...     h=ag.space.Categorical('test', myobj()),
...     i = myfunc(),
...     )
>>> def train_fn(args, reporter):
...     h, i = args.h, args.i
...     assert hasattr(h, 'name') or h == 'test'
...     assert i in ['mxnet', 'pytorch']
...     reporter(epoch=1, accuracy=0)

Create a scheduler and run training trials to search for the best values of the hyperparameters:

>>> scheduler = ag.scheduler.FIFOScheduler(train_fn,
...                                        resource={'num_cpus': 2, 'num_gpus': 0},
...                                        num_trials=20,
...                                        reward_attr='accuracy',
...                                        time_attr='epoch')
>>> scheduler.run()

Core APIs¶

AutoGluon: AutoML Toolkit for Deep Learning

`args`	Decorator for a Python training script that registers its arguments as hyperparameters.
`obj`	Decorator for a Python class that registers its arguments as hyperparameters.
`func`	Decorator for a function that registers its arguments as hyperparameters.

args¶

autogluon.args(default=None, **kwvars)¶

Decorator for a Python training script that registers its arguments as hyperparameters.: Each hyperparameter takes fixed value or is a searchable space, and the arguments may either be: built-in Python objects (e.g. floats, strings, lists, etc.), AutoGluon objects (see autogluon.obj()), or AutoGluon search spaces (see autogluon.space.Int, autogluon.space.Real, etc.).

Examples

>>> import autogluon as ag
>>> @ag.args(batch_size=10, lr=ag.Real(0.01, 0.1))
>>> def train_func(args):
...     print('Batch size is {}, LR is {}'.format(args.batch_size, arg.lr))

obj¶

autogluon.obj(**kwvars)¶

Decorator for a Python class that registers its arguments as hyperparameters.: Each hyperparameter may take a fixed value or be a searchable space (autogluon.space).

Returns

Instance of autogluon.space.AutoGluonObject:: A lazily initialized object, which allows distributed training.

Examples

>>> import autogluon as ag
>>> from mxnet import optimizer as optim
>>> @ag.obj(
>>>     learning_rate=ag.space.Real(1e-4, 1e-1, log=True),
>>>     wd=ag.space.Real(1e-4, 1e-1),
>>> )
>>> class Adam(optim.Adam):
>>>     pass

func¶

autogluon.func(**kwvars)¶

Decorator for a function that registers its arguments as hyperparameters.: Each hyperparameter may take a fixed value or be a searchable space (autogluon.space).

Returns

Instance of autogluon.space.AutoGluonObject:: A lazily initialized object, which allows for distributed training.

Examples

>>> import autogluon as ag
>>> from gluoncv.model_zoo import get_model
>>> 
>>> @ag.func(pretrained=ag.space.Categorical(True, False))
>>> def cifar_resnet(pretrained):
...     return get_model('cifar_resnet20_v1', pretrained=pretrained)