Models

class TabularModel

TabularModel.build

@classmethod
def build(cls: type["TabularModel"],
          preproc: TabularPreproc,
          size: str | Size | TabularModelSize,
          block: str | None = None,
          dropout: float | None = 0.12) -> "TabularModel"

Tabular model to generate synthetic tabular relational data.

Arguments:

• preproc - A TabularPreproc object.
• size - The size configuration of the model. Could be either a TabularModelSize object, a Size object, or a string representation of the latter.
• block - The block type. The possible values depend on whether the data is single table or multi table. For a single table, either ‘free’ (default), ‘causal’, or ‘lstm’. For multi table data, either ‘free’ (default) or ‘lstm’.
• dropout - The dropout probability.

Returns:

A TabularModel instance.

TabularModel.generate

def generate(n_samples: int | None = None,
             ctx: dict[str, pd.DataFrame] | None = None,
             batch_size: int = 0,
             max_block_size: int = 0,
             temp: float = 1.0) -> RelationalData

Generate synthetic relational data.

Arguments:

• n_samples - Desired number of samples in the root table. Must be given if and only if ctx is not provided.
• ctx - The columns of the context from where to start a conditional generation. If provided, n_samples should not be given. The content of the pd.DataFrame’s must match the context columns provided to the TabularPreproc:
  1. If only a subset of the root table’s columns are provided, the model will generate the foreign keys.
  2. If a subset of columns for each table are provided, each pd.DataFrame must also contain the primary and foreign keys, and the generated synthetic data will have the same keys provided by the context. The foreign keys referring to lookup tables should be treated as feature columns, not as foreign keys.
• batch_size - Batch size used during generation. If 0, all data is generated in a single batch.
• max_block_size - Maximum length for each generated sample. Active only for multi-table datasets and for generation from the root table (denoted above as 1.). If 0, no limit is enforced.
• temp - Temperature parameter for sampling.

Returns:

A RelationalData object with the generated synthetic tabular data.

TabularModel.save

def save(path: Path | str) -> None

Save the model to a checkpoint at the given path.

TabularModel.load

@classmethod
def load(cls: type["TabularModel"], path: Path | str) -> "TabularModel"

Load the model from the checkpoint at the given path.

class TextModel

TextModel.build

@classmethod
def build(cls: type["TextModel"],
          preproc: TextPreproc,
          size: str | Size | TextModelSize,
          block_size: int,
          dropout: float | None = 0.12) -> "TextModel"

Text model to generate synthetic text columns of a table which is part of a relational structure.

Arguments:

• preproc - A TextPreproc object.
• size - The size configuration of the model. Could be either a Size object, a TextModelSize object or a string representation of such objects.
• block_size - Maximum text sequence length that the model can process.
• dropout - The dropout probability.

Returns:

A TextModel instance.

TextModel.build_from_pretrained

@classmethod
def build_from_pretrained(cls: type["TextModel"],
                          preproc: TextPreproc,
                          path: Path | str,
                          block_size: int | None = None) -> "TextModel"

Build a text model from a pretrained model.

Arguments:

• preproc - A TextPreproc object.
• path - The path to the checkpoint of the pre-trained model.
• block_size - Maximum text sequence length that the model can process during fine-tuning.

Returns:

A TextModel instance with the weights loaded from the pre-trained model.

TextModel.generate

def generate(data: RelationalData,
             batch_size: int = 0,
             max_text_len: int = 0,
             temp: float = 1.0) -> RelationalData

Generate text columns in the current table.

Arguments:

• data - A RelationalData object containing synthetic data.
• batch_size - Batch size used during generation. If 0, generate all data in a single batch.
• max_text_len - Maximum length for the generated text. If 0, the maximum possible value is used, namely the value of the TabularModel.max_block_size attribute.
• temp - Temperature parameter for sampling.

Returns:

A RelationalData object with the generated synthetic text data.

TextModel.save

def save(path: Path | str) -> None

Save the model to a checkpoint at the given path.

TextModel.load

@classmethod
def load(cls: type["TextModel"], path: Path | str) -> "TextModel"

Load the model from the checkpoint at the given path.

class Size

Enumeration class representing different model sizes. Supported sizes are: SMALL, MEDIUM and LARGE.

class TabularModelSize

Model size for TabularModel objects.

Arguments:

• n_layers - Number of internal layers.
• h - Number of heads.
• d - Size of the internal dimension.

TabularModelSize.from_size

@classmethod
def from_size(cls, size: Size | str) -> "TabularModelSize"

Create an instance based on a given Size or its string representation.

Arguments:

• size - A Size object.

class TextModelSize

Model size for TextModel objects.

Arguments:

• n_layers - Number of internal layers.
• h - Number of heads.
• d - Size of the internal dimension.

TextModelSize.from_size

@classmethod
def from_size(cls, size: Size | str) -> "TextModelSize"

Create an instance based on a given Size or its string representation.

Arguments:

• size - A Size or a str representing a Size.