trainers.ForceMatching.set_datasets

trainers.ForceMatching.set_datasets#

ForceMatching.set_datasets(dataset, train_ratio=0.7, val_ratio=0.1, shuffle=False, include_all=True)#

Sets the datasets for training, testing and validation.

Parameters:

  • dataset – Dictionary containing input and target data as numpy arrays.

  • train_ratio – Percentage of dataset to use for training.

  • val_ratio – Percentage of dataset to use for validation.

  • shuffle – Whether to shuffle data before splitting into train-val-test.

  • include_all – Compute the loss for all samples of the splits by padding the last batch and masking out double samples. Not applied to the training split.

Contents