Victorian tier-2 corpus, 1837-1901
Additional English books in the 1837-1901 window, deduplicated against the existing British Library + Project Gutenberg corpus, for extending pretraining data beyond the 3.35B-token tier-1 set.
Contents
loc-txt.tar- one.txtper book, flat, UTF-8, no nested directoriestier2_loc_manifest.csv-lccn,title,author,year,pages,words,chars,sha256tier2_loc_drop.csv- dropped items and the reason (year,short,low_alpha,low_type_token,dup_bl:0.61, ...)tokenizer-32k.json- the tokenizer the corpus was measured against
Build
| Books kept | 40,912 |
| Characters | 11,444,117,354 |
| Tokens (est., 4.27 chars/token) | ~2.68B |
| Years | 1837-1901 |
| Source rows scanned | 131,885 |
| Dropped: out of range | 131,885 |
| Dropped: OCR quality | 11,737 low type/token, 7,913 short, 321 low alpha, 214 word length |
| Dropped: duplicates | 2,496 intra-tier-2, 1,300 vs British Library, 729 vs Project Gutenberg |
Total corpus after merge: 3.35B (tier-1) + 2.68B (tier-2) = ~6.0B unique tokens.
Provenance
Source: Library of Congress Selected Digitized Books, via the
storytracer/LoC-PD-Books dataset (CC0). Full text is LoC OCR, US public
domain.
Cleaning (applied to every book)
- Unicode NFC, ligature expansion (fi, fl, ffi, ffl), soft-hyphen removal.
- Running heads dropped: lines under 70 chars repeating 10+ times.
- Page numbers and roman-numeral folios dropped.
- Line-break dehyphenation (
govern-\nment->government). - Quality gate: >=20k chars, >=55% alphabetic, mean word length 3.5-8.0, type/token ratio >=0.12.
Deduplication
Char-12 MinHash, 64 permutations, 32-band LSH, Jaccard >=0.45 against:
- every tier-1 British Library book (28,097 texts)
- every Project Gutenberg text (10,256)
- every earlier tier-2 book in the same run
Threshold rationale: at 5% OCR character error rate the same book scores ~0.39 on char-12 shingles while unrelated text sits at ~0.018. 0.45 is conservative: it drops clear re-scans and multi-volume repeats, and keeps partial retellings.
Not included
- HathiTrust (814k public-domain volumes, 480GB+): requires a research application, and Google-digitised volumes need an institutional agreement.
- Chronicling America newspapers: ~1 TB of batch tarballs for a few billion usable tokens; poor download-to-text ratio.
Use
hf download raspbfox/victoria-tier2 loc-txt.tar --repo-type dataset --local-dir /tmp/t2
mkdir -p data/tier2/loc && tar -xf /tmp/t2/loc-txt.tar -C data/tier2/loc
python bin/11_build_bl_corpus.py --phase tokenize \
--text-dir data/tier2/loc --token-dir data/tokens-tier2
- Downloads last month
- -