megatron.data.indexed_dataset#

Description

Classes

IndexedCachedDataset(path)

IndexedDataset(path)

Loader for IndexedDataset

IndexedDatasetBuilder(out_file[, dtype])

MMapIndexedDataset(path[, skip_warmup])

MMapIndexedDatasetBuilder(out_file[, dtype])

Functions

code(dtype)

create_doc_idx(sizes)

data_file_path(prefix_path)

dataset_exists(path, impl)

get_available_dataset_impl()

index_file_path(prefix_path)

infer_dataset_impl(path)

make_builder(out_file, impl[, vocab_size])

make_dataset(path, impl[, skip_warmup])

read_longs(f, n)

write_longs(f, a)