common/data_loaders.py

• 18 E501: line too long (107 > 79 characters) (and 14 similar)

  • Line too long (107 > 79 characters):

    9 uniform_over_groups=None, grouper=None, distinct_groups=True, n_groups_per_batch=None,

  • Line too long (99 > 79 characters):

    14 - loader (str): Loader type. 'standard' for standard loaders and 'group' for group loaders,

  • Line too long (104 > 79 characters):

    15 which first samples groups and then samples a fixed number of examples belonging

  • Line too long (97 > 79 characters):

    19 - uniform_over_groups (None or bool): Whether to sample the groups uniformly or according

  • Line too long (108 > 79 characters):

    21 Setting to None applies the defaults for each type of loaders.

  • Line too long (108 > 79 characters):

    22 For standard loaders, the default is False. For group loaders,

  • Line too long (91 > 79 characters):

    24 - grouper (Grouper): Grouper used for group loaders or for uniform_over_groups=True

  • Line too long (108 > 79 characters):

    25 - distinct_groups (bool): Whether to sample distinct_groups within each minibatch for group loaders.

  • Line too long (99 > 79 characters):

    26 - n_groups_per_batch (int): Number of groups to sample in each minibatch for group loaders.

  • Line too long (94 > 79 characters):

    48 # Replacement needs to be set to True, otherwise we'll run out of minority samples

  • Line too long (84 > 79 characters):

    49 sampler = WeightedRandomSampler(weights, len(dataset), replacement=True)

  • Line too long (126 > 79 characters):

    65 f'n_groups_per_batch was set to {n_groups_per_batch} but there are only {grouper.n_groups} groups specified.')

  • Line too long (80 > 79 characters):

    84 def get_eval_loader(loader, dataset, batch_size, grouper=None, **loader_kwargs):

  • Line too long (116 > 79 characters):

    117 f'batch_size ({batch_size}) must be evenly divisible by n_groups_per_batch ({n_groups_per_batch}).')

  • Line too long (169 > 79 characters):

    120 f'The dataset has only {len(group_ids)} examples but the batch size is {batch_size}. There must be enough examples to form at least one complete batch.')

  • Line too long (92 > 79 characters):

    123 self.unique_groups, self.group_indices, unique_counts = split_into_groups(group_ids)

  • Line too long (81 > 79 characters):

    135 self.group_prob = unique_counts.numpy() / unique_counts.numpy().sum()

  • Line too long (86 > 79 characters):

    149 replace=len(self.group_indices[group]) <= self.n_points_per_group,