INNER CODE UNIT · Python
QGDataset
AMontgomerie/question_generator · training/dataset.py:10
class QGDataset(torch.utils.data.Dataset):
def __init__(
self,
data: datasets.Dataset,
max_length: int,
pad_mask_id: int,
tokenizer: AutoTokenizer
) -> None:
self.data = pd.DataFrame(data)
self.max_length = max_length
self.pad_mask_id = pad_mask_id
self.tokenizer = tokenizer
def __len__(self) -> int:
return len(self.data)
def __getitem__(self, index: int) -> Mapping[str, torch.Tensor]:
item = self.data.loc[index]