Struct ai_dataloader::iterable::DataLoader

pub struct DataLoader<D, C> { /* private fields */ }

For iterable dataset, the datalaoder will yield until the underlying iterator is None. As the iteration over the dataset can be done mutliple time, depending if the underlying dataset iterator consume the dataset or not.

Implementations

impl<D> DataLoader<D, DefaultCollate>where
    D: IntoIterator,
    DefaultCollate: Collate<D::Item>,

pub fn builder(dataset: D) -> Builder<D, DefaultCollate>

return a DataLoader builder.

impl<'dataset, D, C> DataLoader<D, C>where
    D: 'dataset,
    &'dataset D: IntoIterator,
    C: Collate<<&'dataset D as IntoIterator>::Item>,

pub fn iter(&'dataset self) -> Iter<<&'dataset D as IntoIterator>::IntoIter, C>

Iterate over the dataloader without consuming the underlying dataset. As it make no sens to collate reference into a tensor, by default element are copied.

Trait Implementations

impl<D: Debug, C: Debug> Debug for DataLoader<D, C>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'dataset, D, C> IntoIterator for &'dataset DataLoader<D, C>where
    D: 'dataset,
    &'dataset D: IntoIterator,
    C: Collate<<&'dataset D as IntoIterator>::Item>,

type Item = <C as Collate<<&'dataset D as IntoIterator>::Item>>::Output

The type of the elements being iterated over.

type IntoIter = Iter<<&'dataset D as IntoIterator>::IntoIter, C>

Which kind of iterator are we turning this into?

fn into_iter(self) -> Self::IntoIter

Creates an iterator from a value.

impl<D, C> IntoIterator for DataLoader<D, C>where
    D: IntoIterator,
    C: Collate<<D as IntoIterator>::Item>,

type Item = <C as Collate<<D as IntoIterator>::Item>>::Output

The type of the elements being iterated over.

type IntoIter = IntoIter<<D as IntoIterator>::IntoIter, C>

Which kind of iterator are we turning this into?

fn into_iter(self) -> Self::IntoIter

Creates an iterator from a value.