vak.config.dataset.DatasetConfig¶
- class vak.config.dataset.DatasetConfig(path, splits_path=None, name=None, params={})[source]¶
Bases:
object
Class that represents dataset table in configuration file.
- path¶
Path to the directory that contains the dataset. Equivalent to the root parameter of
torchvision
datasets.- Type:
- splits_path¶
Path to file representing splits. Default is None.
- Type:
pathlib.Path, optional
- name¶
Name of dataset. Only required for built-in datasets from the
datasets
module. Default is None.- Type:
str, optional
- params¶
Parameters for dataset class, passed in as keyword arguments. E.g.,
window_size=2000
. Default is None.- Type:
dict, optional
- __init__(path, splits_path=None, name=None, params={}) None ¶
Method generated by attrs for class DatasetConfig.
Methods
__init__
(path[, splits_path, name, params])Method generated by attrs for class DatasetConfig.
asdict
()Convert this
DatasetConfig
instance to adict
that can be passed into functions that take adataset_config
argument, likevak.train()
andvak.predict()
.from_config_dict
(config_dict)Attributes
- asdict()[source]¶
Convert this
DatasetConfig
instance to adict
that can be passed into functions that take adataset_config
argument, likevak.train()
andvak.predict()
.