Tensorflow-backed Dataloader
Tensorflow
-backed Dataloader¤
to_tf_dataset¤
to_tf_dataset (dataset:jax_dataloader.datasets.Dataset)
to_tf_dataset¤
to_tf_dataset (dataset:jax_dataloader.datasets.Dataset)
to_tf_dataset¤
to_tf_dataset (dataset:jax_dataloader.datasets.Dataset)
DataLoaderTensorflow¤
DataLoaderTensorflow (dataset:Union[jax_dataloader.datasets.Dataset,tenso rflow.python.data.ops.dataset_ops.DatasetV2,Annotat ed[Union[datasets.arrow_dataset.Dataset,datasets.da taset_dict.DatasetDict,datasets.dataset_dict.Iterab leDatasetDict,datasets.iterable_dataset.IterableDat aset],beartype.vale.Is[lambda_:hf_datasetsisnotNone ]]], batch_size:int=1, shuffle:bool=False, drop_last:bool=False, **kwargs)
Tensorflow Dataloader
Type | Default | Details | |
---|---|---|---|
dataset | typing.Union[jax_dataloader.datasets.Dataset, tensorflow.python.data.ops.dataset_ops.DatasetV2, typing.Annotated[typing.Union[datasets.arrow_dataset.Dataset, datasets.dataset_dict.DatasetDict, datasets.dataset_dict.IterableDatasetDict, datasets.iterable_dataset.IterableDataset], beartype.vale.Is[lambda _: hf_datasets is not None]]] | ||
batch_size | int | 1 | Batch size |
shuffle | bool | False | If true, dataloader shuffles before sampling each batch |
drop_last | bool | False | Drop last batch or not |
kwargs |
test_dataloader(DataLoaderTensorflow, samples=20, batch_size=12)
test_dataloader(DataLoaderTensorflow, samples=20, batch_size=10)
test_dataloader(DataLoaderTensorflow, samples=11, batch_size=10)
An NVIDIA GPU may be present on this machine, but a CUDA-enabled jaxlib is not installed. Falling back to cpu.
test_dataloader(DataLoaderTensorflow, ds_type='tf', samples=40, batch_size=11)
test_dataloader(DataLoaderTensorflow, ds_type='hf', samples=40, batch_size=11)