from pydantic.dataclasses import dataclass
from pyraug.config import BaseConfig
[docs]@dataclass
class BaseModelConfig(BaseConfig):
"""This is the base configuration instance of the models deriving from
:class:`~pyraug.config.BaseConfig`.
Parameters:
input_dim (int): The input_data dimension
latent_dim (int): The latent space dimension. Default: None.
default_encoder (bool): Whether the encoder default. Default: True.
default_decoder (bool): Whether the encoder default. Default: True.
"""
input_dim: int = None
latent_dim: int = 10
uses_default_encoder: bool = True
uses_default_decoder: bool = True
[docs]@dataclass
class BaseSamplerConfig(BaseConfig):
"""
This is the base configuration of a model sampler
Parameters:
samples_number (int): The number of samples to generate
batch_size (int): The number of samples to generate in each batch
samples_per_save (int): The number of samples to be saved together.
By default, when generating, the generated data is saved in ``.pt`` format
in several files. This specifies the number of samples to be saved in these
files. Amend this argument if you deal with particularly large data. Default: 500.
no_cuda (bool): Disable `cuda`. Default: False
"""
output_dir: str = None
batch_size: int = 50
samples_per_save: int = 500
no_cuda: bool = False