CrabNetModel

class mastml.models.CrabNetModel(composition_column, epochs, val_frac, drop_unary=False, savepath=None)[source]

Bases: BaseEstimator, TransformerMixin

Implementation of PyTorch-based CrabNet regressor model based on the following work:

Wang, A., Kauwe, S., Murdock, R., Sparks, T. “Compositionally restricted attention-based network for “ materials property predictions”, npj Computational Materials (2021) (https://www.nature.com/articles/s41524-021-00545-1)

The code to run CrabNet was integrated into MAST-ML based on source files available in this repository: https://github.com/anthony-wang/CrabNet

Args:

composition_column (str): string denoting the name of an input column containing materials compositions

epochs (int): number of training epochs

val_frac (float): fraction of input data to use as validation (0 to 1). Note this can be set to 0 for most MAST-ML

runs as the splitting is done outside of this model.

drop_unary (bool): whether or not to drop compositions containing one element. Should probably be False always. If true,

can cause some array length mismatches.

savepath (str): path to save the data to. This is set automatically in the data_splitters routine.

Methods:

get_model: method that prepares the CrabNet model. Called when the CrabNetModel class is instantiated.

fit: method that fits the model parameters to the provided training data
Args:
X: (pd.DataFrame), dataframe of X data, needs to contain at least a column of material compositions. Note

that featurization is done internally by the model.

y: (pd.Series), series of y target data

Returns:

fitted model

predict: method that evaluates model on new data to give predictions
Args:
X: (pd.DataFrame), dataframe of X data, needs to contain at least a column of material compositions. Note

that featurization is done internally by the model.

as_frame: (bool), whether to return data as pandas dataframe (else numpy array)

Returns:

series or array of predicted values

Methods Summary

fit(X, y)

get_model()

predict(X[, as_frame])

set_predict_request(*[, as_frame])

Configure whether metadata should be requested to be passed to the predict method.

Methods Documentation

fit(X, y)[source]
get_model()[source]
predict(X, as_frame=True)[source]
set_predict_request(*, as_frame: bool | None | str = '$UNCHANGED$') → CrabNetModel

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to predict.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters

as_framestr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for as_frame parameter in predict.

Returns

selfobject

The updated object.