chop.datasets

chop.dataset.nerf

chop.dataset.nerf.get_nerf_dataset(name: str, path: PathLike, split: str)

chop.dataset.nerf.get_nerf_dataset_cls(name: str)

chop.dataset.nlp

chop.dataset.nlp.get_nlp_dataset(name: str, split: str, tokenizer, max_token_len: int, num_workers: int, load_from_cache_file: bool = True, auto_setup: bool = True)

chop.dataset.nlp.get_nlp_dataset_cls(name: str)

chop.dataset.physical

chop.dataset.physical.get_physical_dataset(name: str, path: Path, split: str)

chop.dataset.physical.get_physical_dataset_cls(name: str)

chop.dataset.vision

chop.dataset.vision.get_vision_dataset(name: str, path: PathLike, split: str, model_name: str)

Parameters:

• name (str) – name of the dataset

• path (str) – path to the dataset

• train (bool) – whether the dataset is used for training

• model_name (Optional[str, None]) – name of the model. Some pretrained models have

• evaluation. (model-dependent transforms for training and)

Returns:

dataset (with transforms)

Return type:

dataset (torch.utils.data.Dataset)

chop.dataset.vision.get_vision_dataset_cls(name: str)