src.corpora.indexer.IndexedDataset

class IndexedDataset(cache_dir, seq_len: int, stride: Optional[int] = None)[source]

Bases: IterDataPipe[BatchEncoding]

Methods

build_or_load

register_datapipe_as_function

register_function

set_getstate_hook

set_reduce_ex_hook

Attributes

functions

getstate_hook

reduce_ex_hook