Hugging Face's logo

Datasets:
disco-jack-basement
/
byob-pd-book-corpus

Tasks:
Text Generation
Modalities:
Text
Formats:
text
Languages:
English
French
German
Size:
100M - 1B
Tags:
text
literature
public-domain
project-gutenberg
wikisource
character-level
Libraries:
Datasets
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
text
string
Dedication
You gave me Christopher Robin, and then
You breathed new life in Pooh.
Whatever of each has left my pen
Goes homing back to you.
My book is ready, and comes to greet
The mother it longs to see-
It would be my present to you, my sweet,
If it weren't your gift to me.
Contradiction
An introduction is to introduce people, but Christopher Robin and his friends, who have already been introduced to you, are now going to say Goodbye. So this is the opposite. When we asked Pooh what the opposite of an Introduction was, he said "The what of a what?" which didn't help us as much as we had hoped, but luck...
Why we are having a Contradiction is because last week when Christopher Robin said to me, "What about that story you were going to tell me about what happened to Pooh when-" I happened to say very quickly, "What about nine times a hundred and seven?" And when we had done that one, we had one about cows going through a ...
I
In Which a House Is Built at Pooh Corner for Eeyore
One day when Pooh Bear had nothing else to do, he thought he would do something, so he went round to Piglet's house to see what Piglet was doing. It was still snowing as he stumped over the white forest track, and he expected to find Piglet warming his toes in front of his fire, but to his surprise he saw that the door...
"He's out," said Pooh sadly. "That's what it is. He's not in. I shall have to go a fast Thinking Walk by myself. Bother!"
But first he thought that he would knock very loudly just to make quite sure ... and while he waited for Piglet not to answer, he jumped up and down to keep warm, and a hum came suddenly into his head, which seemed to him a Good Hum, such as is Hummed Hopefully to Others.
The more it snows
(Tiddely pom),
The more it goes
(Tiddely pom),
The more it goes
(Tiddely pom),
On snowing.
And nobody knows
(Tiddely pom),
How cold my toes
(Tiddely pom),
How cold my toes
(Tiddely pom),
Are growing.
"So what I'll do," said Pooh, "is I'll do this. I'll just go home first and see what the time is, and perhaps I'll put a muffler round my neck, and then I'll go and see Eeyore and sing it to him."
He hurried back to his own house; and his mind was so busy on the way with the hum that he was getting ready for Eeyore that, when he suddenly saw Piglet sitting in his best armchair, he could only stand there rubbing his head and wondering whose house he was in.
"Hallo, Piglet," he said. "I thought you were out."
"No," said Piglet, "it's you who were out, Pooh."
"So it was," said Pooh. "I knew one of us was."
He looked up at his clock, which had stopped at five minutes to eleven some weeks ago.
"Nearly eleven o'clock," said Pooh happily. "You're just in time for a little smackerel of something," and he put his head into the cupboard. "And then we'll go out, Piglet, and sing my song to Eeyore."
"Which song, Pooh?"
"The one we're going to sing to Eeyore," explained Pooh.
The clock was still saying five minutes to eleven when Pooh and Piglet set out on their way half an hour later. The wind had dropped, and the snow, tired of rushing round in circles trying to catch itself up, now fluttered gently down until it found a place on which to rest, and sometimes the place was Pooh's nose and ...
"Pooh," he said at last, and a little timidly, because he didn't want Pooh to think he was Giving In, "I was just wondering. How would it be if we went home now and practised your song, and then sang it to Eeyore tomorrow-or-or the next day, when we happen to see him?"
"That's a very good idea, Piglet," said Pooh. "We'll practise it now as we go along. But it's no good going home to practise it, because it's a special Outdoor Song which Has To Be Sung In The Snow."
"Are you sure?" asked Piglet anxiously.
"Well, you'll see, Piglet, when you listen. Because this is how it begins. The more it snows, tiddely pom-"
"Tiddely what?" said Piglet.
"Pom," said Pooh. "I put that in to make it more hummy. The more it goes, tiddely pom, the more-"
"Didn't you say snows?"
"Yes, but that was before."
"Before the tiddely pom?"
End of preview. Expand in Data Studio

Dataset Card for BYOB LM on Steroids - Public-Domain Book Training Corpus

A multilingual, public-domain corpus of literary, philosophical, and scientific works, published for character-level language-model pre-training. It ships in two forms: raw per-author .txt files (one file per author) and a pre-tokenized, memmap-ready cache (train.bin / val.bin / meta.json) for fast training.

Dataset Sources

All texts are public domain in their source jurisdiction. Project Gutenberg boilerplate headers/footers and trademark notices are stripped, and text is NFKD-normalized. Verify public-domain status in your own jurisdiction.

Dataset Structure

Nested public-domain tiers, each a subset of the next (medium subset of large subset of xlarge subset of 2xlarge subset of 4xlarge):

tier characters authors vocab
medium 523,561,434 109 97
large 1,079,432,702 248 198
xlarge 2,050,985,583 431 199
2xlarge 4,262,779,281 1524 204
4xlarge 10,184,454,167 4132 209

(Size is measured in characters; size_categories counts characters as units.)

  • Raw text: one .txt per author under <tier>/ - source-tagged (<author>-complete.txt) in the medium/large/xlarge tiers, or consolidated to a single <author>.txt per author in 2xlarge and 4xlarge.
  • Per-tier manifests: indices/<tier>.md (authors, works, char counts).
  • Prepared caches: prepared/<tier>_bin/{train.bin,val.bin,meta.json}, where meta.json = {vocab_size, stoi, itos, train_len, val_len, dtype:"uint8"}. These are a byob-internal convenience cache, fully reproducible from the raw .txt; pull them with hf_hub_download, not load_dataset.

Uses

Language-model pre-training, especially character-level GPTs. The prepared/<tier>_bin/ caches are flat uint8 streams (one char = one byte, vocab < 256) meant to be memmapped (see byob_llm CharDataset.from_bin). Not annotated; not intended for supervised tasks needing labels.

from datasets import load_dataset
ds = load_dataset("<repo-id>", "4xlarge")   # a tier config: medium / large / xlarge / 2xlarge / 4xlarge

Dataset Creation

Curation Rationale

PUBLIC DOMAIN ONLY. The corpus enforces, in code, a strict rule: NO Russian content anywhere - no Russian authors, no Russian-language works, no Russia-themed material. This is a deliberate, code-enforced curation policy.

Source Data

Harvested from Project Gutenberg (via Gutendex), Standard Ebooks, the Internet Archive, and Wikisource; Gutenberg boilerplate stripped; cleaned, de-duplicated, and NFKD-normalized. The medium/large/xlarge tiers are frozen and source-tagged; 2xlarge and 4xlarge are consolidated to one merged .txt per author (2xlarge the curated, slimmed derivative - Gutenberg-preferred where an author appears in more than one source; 4xlarge the full master - content-preserving, every source file per author merged so nothing harvested is dropped). See indices/<tier>.md for the full per-author manifest. The corpus spans English, American, French, German, Italian, Spanish, Portuguese, Dutch, the Nordic languages (Swedish/Danish/Norwegian/ Finnish), Polish, Hungarian, Czech, Latin, Greek, and Ukrainian literature and philosophy, among others.

Bias, Risks, and Limitations

Historical texts reflect the dated and biased language of their eras. The corpus is multilingual but English-dominant. No content filtering beyond the curation rules.

License and Attribution

Released under CC0-1.0 (public-domain dedication). Note: Wikisource editorial apparatus may be CC-BY-SA, and the NFKD normalization is a derivative. Please credit Project Gutenberg and Wikisource.
Downloads last month
3,605
Number of rows:
155,640,356
Total file size:
36.3 GB

Models trained or fine-tuned on disco-jack-basement/byob-pd-book-corpus