def _construct_loaders(self, X_train, y_train, eval_set):
"""Generate dataloaders for train and eval set.
Parameters
----------
X_train : np.array
Train set.
y_train : np.array
Train targets.
eval_set : list of tuple
List of eval tuple set (X, y).
Returns
-------
train_dataloader : `torch.utils.data.Dataloader`
Training dataloader.
valid_dataloaders : list of `torch.utils.data.Dataloader`
List of validation dataloaders.
"""
# all weights are not allowed for this type of model
y_train_mapped = self.prepare_target(y_train)
for i, (X, y) in enumerate(eval_set):
y_mapped = self.prepare_target(y)
eval_set[i] = (X, y_mapped)
train_dataloader, valid_dataloaders = create_dataloaders(
X_train,
y_train_mapped,
eval_set,
self.updated_weights,
self.batch_size,
self.num_workers,
self.drop_last,
self.pin_memory,
)
return train_dataloader, valid_dataloaders
def construct_loaders(self, X_train, y_train, X_valid, y_valid, weights,
batch_size):
"""
Returns
-------
train_dataloader, valid_dataloader : torch.DataLoader, torch.DataLoader
Training and validation dataloaders
-------
"""
train_dataloader, valid_dataloader = create_dataloaders(
X_train, y_train, X_valid, y_valid, 0, batch_size)
return train_dataloader, valid_dataloader
def construct_loaders(self, X_train, y_train, X_valid, y_valid, weights,
batch_size, num_workers, drop_last):
"""
Returns
-------
train_dataloader, valid_dataloader : torch.DataLoader, torch.DataLoader
Training and validation dataloaders
-------
"""
y_train_mapped = np.vectorize(self.target_mapper.get)(y_train)
y_valid_mapped = np.vectorize(self.target_mapper.get)(y_valid)
train_dataloader, valid_dataloader = create_dataloaders(
X_train, y_train_mapped, X_valid, y_valid_mapped, weights,
batch_size, num_workers, drop_last)
return train_dataloader, valid_dataloader
def construct_loaders(self, X_train, y_train, X_valid, y_valid, weights,
batch_size, num_workers, drop_last):
"""
Returns
-------
train_dataloader, valid_dataloader : torch.DataLoader, torch.DataLoader
Training and validation dataloaders
-------
"""
if isinstance(weights, int):
if weights == 1:
raise ValueError(
"Please provide a list of weights for regression.")
if isinstance(weights, dict):
raise ValueError(
"Please provide a list of weights for regression.")
train_dataloader, valid_dataloader = create_dataloaders(
X_train, y_train, X_valid, y_valid, weights, batch_size,
num_workers, drop_last)
return train_dataloader, valid_dataloader
