Datasets:

Icold
/

kron-corpus

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
source: string
licence: string
url: string
language: string
domain: string
sha256: string
bytes: int64
docs: int64
path: string
note: string
text: string
title: string
to
{'title': Value('string'), 'text': Value('string'), 'language': Value('string'), 'domain': Value('string'), 'source': Value('string'), 'licence': Value('string'), 'url': Value('string'), 'note': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              source: string
              licence: string
              url: string
              language: string
              domain: string
              sha256: string
              bytes: int64
              docs: int64
              path: string
              note: string
              text: string
              title: string
              to
              {'title': Value('string'), 'text': Value('string'), 'language': Value('string'), 'domain': Value('string'), 'source': Value('string'), 'licence': Value('string'), 'url': Value('string'), 'note': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

title string	text string	language string	domain string	source string	licence string	url string	note string
A Rickconvenient Mort/Transcript	Ah, I love that new T-shirt smell. Chicks are gonna be all up in our biz when they see us wearing these Rick and Morty custom tees. I bet we could sell these. Hells, yeah. I fucking love merch! The rain! lt-it burns! Oh, shit. Quick, Morty, protect the shirts! The door! Open the door! Rick! Is it-Morty, will yo...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/A Rickconvenient Mort/Transcript	easter egg: series dialogue
Air Force Wong/Transcript	Wong: Before we move off this topic, can I ask if you're currently a hologram? Rick: Yeah, yeah, h-hold on, one sec, left a bag of funyuns down there last week. Didn't want to waste our time. Okay therapy me Wong: Do you need to take that? Rick: No just the president Wong: Maybe we should turn off the phone Rick: ...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Air Force Wong/Transcript	easter egg: series dialogue
Amortycan Grickfitti/Transcript	Beth: Jerry, can you watch the kids tonight? Someone played Barry White over the racetrack speakers, and now there's a sevenhorse pregnancy. Jerry: Oh, um-- Rick: No, he can't watch the kids, it's guys' night. Beth: How did you two start having guys' nights? Rick: So now, you're complaining that I like your stupid ...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Amortycan Grickfitti/Transcript	easter egg: series dialogue
Analyze Piss/Transcript	(Rick and Morty are on an island firing at small creatures) Morty: I-I can't hold them off, Rick! Th-There's too many! ( Rick gets slapped by a creature's arm) Mr. Nimbus: Richard! Your fortress is a salty waste of coral and clam. Relinquish it at once! (Rick runs up to him and punches him. He breaks a hole in a shi...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Analyze Piss/Transcript	easter egg: series dialogue
Anatomy Park (episode)/Transcript	[Open Ext. Smith residence] (Outside, there is snow on the ground. The Smith house is decorated for Christmas, and Jerry can be heard singing Last King Christmas.) Jerry: singing Last King Christmas last arrived! [Trans Int. Kitchen] (The kitchen is a mess of food in various stages of preparation. Jerry pulls a h...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Anatomy Park (episode)/Transcript	easter egg: series dialogue
Auto Erotic Assimilation/Transcript	This article is a transcript of the Season 2 episode Auto Erotic Assimilation. ‎Open Ext. Rick's Ship in flying through a planet's rings (Morty is in the backseat, Summer is in the starboard seat, Rick is driving) All: (singing) Love. Connection. Experience. (Morty claps) Rick: Yeah Morty All: (singing) Come together....	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Auto Erotic Assimilation/Transcript	easter egg: series dialogue
Bethic Twinstinct/Transcript	Morty: Oh, my God, I cannot fucking believe this! Here we fucking go! Oh! Jeez! Rick: Pardon the interruption. We are once again in the federal clear. Morty: Rick, what the hell?! A-Are you a Turkey Dracula now? Rick: Morty, is that a Pooplickian GamePod? Morty: Actually it's the GamePod XL. They say it's the most ...	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Bethic Twinstinct/Transcript	easter egg: series dialogue
Big Trouble in Little Sanchez/Transcript	"This article is a transcript of the Rick and Morty episode Big Trouble In Little Sanchez. It aired (...TRUNCATED)	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Big Trouble in Little Sanchez/Transcript	easter egg: series dialogue
Childrick of Mort/Transcript	"Summer: What kind of parents drag their kids away for the entire week of spring break? I'm missing (...TRUNCATED)	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Childrick of Mort/Transcript	easter egg: series dialogue
Claw and Hoarder: Special Ricktim's Morty/Transcript	"Rick, we did it!\nWe -- we got it!\n\nHurry up, morty.\nGet in!\n\nChachi!\n[bleep] chachi, morty.\(...TRUNCATED)	en	humour	Fandom wiki (api.php)	CC-BY-SA-3.0	https://rickandmorty.fandom.com/wiki/Claw and Hoarder: Special Ricktim's Morty/Transcript	easter egg: series dialogue

End of preview.

KronCorpus

A bilingual English–Russian pretraining corpus with a science accent.

KronCorpus is the training corpus behind KronLLM, a compact language model whose weights are Kronecker products. It is published separately because the data is useful on its own: a small, fully documented, science-heavy bilingual corpus for anyone training a model in the 10M–100M parameter range.

What is in it

property	value
tokens	2,183,929,941 train tokens (8K BPE, 3.80 bytes/token)
size	8.61 GB, 945,234 documents
languages	EN 46.2 % / RU 53.8 % of bytes, which is EN 62.1 % / RU 37.9 % of tokens
documents	787,004 after MinHash (20,022 dropped) + 289,000 Russian bulk articles (title-deduplicated)
tokenizer	8K byte-level BPE, bilingual, trained on the mix
held-out	2 % per language, disjoint by document
provenance	source, licence, URL and SHA-256 for every shard in `manifest.jsonl`

Why bytes and not tokens. The quota is fixed in bytes because that is the only quantity that does not depend on the tokenizer, and the token share is measured and published next to it. The measurement is blunt: Russian costs 5.12 bytes/token against 3.00 for English, so a byte quota is not a token quota. The quota build gave Russian 15.7 % of the tokens at 29 % of the bytes; the fix was more Russian data, not a different formula. The full Russian Wikipedia (2.0 GB of articles) and the full Russian Wikisource (1.1 GB of public-domain literature) were imported with curated titles skipped, taking Russian to 37.9 % of tokens.

Composition

Domain quotas are the design contract; the "measured" column is the actual mix after dedup, as a share of all documents.

domain	quota	measured	main sources
mathematics	25 %	19.1 %	OpenWebMath, Wikipedia math
science (physics, astronomy, cosmology)	14 %	11.1 %	arXiv, Wikipedia, curated cosmology articles, theoretical-physics categories
code	15 %	7.5 %	CodeParrot Clean, CPython stdlib
machine learning	12 %	5.4 %	arXiv cs.LG/CS.CL abstracts, Cosmopedia, framework docs
literature	11 %	21.2 %	Project Gutenberg, Wikisource (1.1 GB Russian), FineWeb-2
logic	10 %	3.2 %	Wikipedia logic, Lean 4 mathlib proofs
encyclopedic	9 %	30.3 %	Wikipedia EN/RU, full Russian Wikipedia import
CS / software engineering / engineering	3 %	2.3 %	IETF RFCs, Wikipedia (microprocessors, VLSI, XP, agile)
humour / satire	1 %	<0.1 %	Project Gutenberg, Fandom transcripts
self-identity	1 %	0.09 %	synthetic: 800 plain + 850 pastiche per language, no numeric facts

The table is honest about the two hard domains: logic and ML are short. Logic text at scale barely exists outside formal-proof repositories, and ML prose is mostly behind publisher paywalls. The mixer reports both as shortfalls instead of padding them with mislabelled text.

How sources were selected (the priority order)

Verifiable provenance beats volume. A source is admitted only if its licence is stated and a URL can be recorded. Every shard carries source / licence / url / sha256 in manifest.jsonl; nothing is anonymous.
Primary sources over scrapes. Wikipedia dumps, arXiv, Project Gutenberg, OpenWebMath and official repositories come first; web text (FineWeb-2) is used only where a domain has no primary source, and its shards are marked as keyword-routed.
Science and mathematics first. The quotas are not balanced for "general English": a quarter of the corpus is mathematics, and physics, astronomy and cosmology get an explicit slice of the science budget.
Quality filters before volume. Minimum length, symbol ratio, repetition ratio, script check per language, and a MinHash near-duplicate pass (Jaccard 0.8) that removes translated mirrors and boilerplate.
Domain labels are declared as heuristics where they are heuristics. Wikipedia and FineWeb shards are routed by subject keywords, and the manifest says so. Curated lists (cosmology, chips/XP) are explicit.
Leave room for the model's own name. 1 % is reserved for a synthetic self-identity slice; it is generated from a fixed fact sheet, not scraped.

Sources

source	licence	role
Wikipedia (HF dump 2023-11-01, EN/RU)	CC BY-SA	encyclopedic, domains via keywords
OpenWebMath	ODC-By 1.0	mathematics
arXiv (abstracts + full text)	arXiv non-exclusive	science, ML
Project Gutenberg (pglaf mirror)	public domain	literature, satire
Wikisource	CC BY-SA	Russian literature
CodeParrot Clean	permissive, per file	code
FineWeb-2 `rus_Cyrl`	ODC-By 1.0	Russian domains without a primary source
Lean 4 `mathlib4`	Apache-2.0	formal proofs (logic)
Cosmopedia v2	Apache-2.0	ML/science textbook-style prose
Fandom wiki API	CC BY-SA 3.0	easter-egg slice only

Loading

The tokenised bins (uint16, 4.4 GB) are distributed outside git — GitHub is not a data host.

# Kaggle (private dataset with the tokenised bins, uploaded after the build)
#   kaggle datasets download -d icold21/kron-corpus-2b
import numpy as np
tokens = np.memmap("train.bpe.bin", dtype=np.uint16, mode="r")

# Hugging Face (card, manifest, identity slice, easter eggs — no bins)
#   from datasets import load_dataset
#   ds = load_dataset("Icold/kron-corpus")

The raw, deduplicated text shards and the exact mix are reproducible from this repository: scripts/ contains the fetchers, the mixer, the tokenizer training and the bin preparation. See docs/composition.md for measured numbers and docs/rebuild.md for the command sequence.

Reproducing the build

python scripts/build_bilingual.py fetch-hf --source all      # Wikipedia, arXiv, OpenWebMath, FineWeb-2
python scripts/build_bilingual.py fetch-gutenberg-ids ...    # curated public-domain books
python scripts/build_bilingual.py fetch-tarball ...          # Lean mathlib4
python scripts/gen_identity.py --samples-per-language 800    # self-identity slice
python scripts/build_bilingual.py build --budget-en-mb 4300 --budget-ru-mb 2870
python scripts/prepare_bilingual_bin.py --vocab 8192

Fetches are append-only and sharded; build is deterministic given the same shards (dedup walks documents in SHA-256 order).

Licence

The corpus is a collection: each shard keeps the licence of its source, as recorded in manifest.jsonl. CC BY-SA text (Wikipedia, Wikisource, Fandom) requires attribution and share-alike; Gutenberg texts are public domain in the US; OpenWebMath, FineWeb-2 and peS2o are ODC-By 1.0; mathlib4 is Apache-2.0. The build scripts themselves are MIT, like KronLLM.

If you redistribute a mix, keep manifest.jsonl next to it — that file is the attribution.

_{Two slices in data/easter_eggs/ and data/identity/ are easter eggs: a
synthetic self-identity set where the model calls itself Kron (800 plain +
850 pastiche per language), and 2 MB of animated sci-fi dialogue (81 episode
transcripts plus 11 season-9 plot summaries). Both are under 0.1 % of the corpus and marked in the
manifest.}

Downloads last month: 102

Total file size:

5.12 MB