|
{ |
|
"sources": { |
|
"the_pile": { |
|
"provider": "local", |
|
"file_type": "json", |
|
"files": [ |
|
"/fs/cml-datasets/Pile/train/00.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/01.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/02.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/03.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/04.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/05.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/06.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/07.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/08.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/09.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/10.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/11.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/12.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/13.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/14.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/15.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/16.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/17.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/18.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/19.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/20.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/21.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/22.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/23.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/24.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/25.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/26.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/27.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/28.jsonl.zst", |
|
"/fs/cml-datasets/Pile/train/29.jsonl.zst" |
|
], |
|
"filter": null, |
|
"split": "train", |
|
"streaming": true, |
|
"remove_columns": null, |
|
"concatenate_successive_entries": 0 |
|
} |
|
}, |
|
"name": "the_pile", |
|
"normalizer": { |
|
"force_lowercase": true, |
|
"strip_accents": true, |
|
"force_english_keyboard": true, |
|
"whitespace_escape": false |
|
}, |
|
"tokenizer": "WordPiece", |
|
"vocab_size": 32768, |
|
"seq_length": 128, |
|
"include_cls_token_in_corpus": false, |
|
"include_sep_token_in_corpus": true, |
|
"use_type_ids": false, |
|
"max_entries_in_raw_dataset": 8000000.0, |
|
"max_seq_in_tokenized_dataset": 85000000.0, |
|
"named_entity_simplification": false, |
|
"remove_whitespaces": false, |
|
"remove_trash": true, |
|
"trash_cutoff": 0.25, |
|
"deduplicate_entries": false, |
|
"deduplication_threshold": 75, |
|
"ordering": "sentence-length-curriculum" |
|
} |