split_sizes_from_probabilities

lightning_pose.data.utils.split_sizes_from_probabilities(total_number: int, train_probability: float, val_probability: float | None = None, test_probability: float | None = None) → list[int]

Returns the number of examples for train, val and test given split probs.

Parameters:

• total_number – total number of examples in dataset

• train_probability – fraction of examples used for training

• val_probability – fraction of examples used for validation

• test_probability – fraction of examples used for test. Defaults to None. Can be computed as the remaining examples.

Returns:

[num training examples, num validation examples, num test examples]