PiotrSty's picture
Upload folder using huggingface_hub
359ceb1 verified
|
Raw History Blame Contribute Delete
5.35 kB
---
license: cc-by-4.0
language: pl
pretty_name: polocrbench-train-pool-v1
---
# polocrbench-train-pool-v1
A single self-contained **training pool** for Polish document understanding,
assembled for work toward [PolOCRBench](https://poleval.pl/tasks/task1)
(PolEval 2026 — transcription, table extraction, KIE). Community-assembled
training material — **not** the official organizer release.
**3 610 pages** (3 405 train / 205 development) from seven open-licensed sources:
| source | pages | license | labels |
|---|---:|---|---|
| `docatlas-pl-clean` ([repo](https://huggingface.co/datasets/PiotrSty/docatlas-pl-clean-v1)) | 1 713 | CC BY 4.0 (Heakl et al. 2026, arXiv:2605.12623) | full-page Markdown |
| `impact-hist-train` ([repo](https://huggingface.co/datasets/PiotrSty/impact-print-v2-train-pages)) | 298 | CC BY 3.0 (IMPACT Centre of Competence / PSNC) | page texts joined from 1 987 line transcriptions |
| `polocrbench-synth` (project-authored, seed 20260922) | 1 200 | CC BY 4.0 | Markdown + HTML tables + KIE fields (faktura/umowa/pismo_urzedowe/formularz) |
| `polocrbench-synth-photo` (project-authored, seed 20261018, **photo recipe**) | 300 | CC BY 4.0 | same, with phone-photo degradations: perspective warp, directional shadow ramp, blur 0.8–2.0, downscale + JPEG |
| `handwriting-notes` (author self-licensed: Piotr Styła, 2026-10-17) | 40 | CC BY 4.0 | full-page transcription incl. bracketed drawing descriptions; 16 pages orientation-normalized (rot. 180°) |
| `handwriting-mednotes` (self-licensed: Piotr Styła, 2026-10-23) | 49 | CC BY 4.0 | study notes with medical knowledge (dosage schedules, disease patterns as taught) — no personal health records; mostly page fragments with cut edges (`[ucięte]` markers) |
| `handwriting-gardnotes` (self-licensed: Piotr Styła, 2026-10-23) | 10 | CC BY 4.0 | garden journal 1988–89 on yellowed paper: plant/variety lists with strike-outs (`[skreślone]`), weather table, gardening log; one re-scan pair kept as rescanning variety |
Synthetic pages carry the subtask-B and C annotations: `tables.jsonl`
(1 072 HTML tables) and `kie.jsonl` (1 500 field-level JSON records), keyed by
`page_id`. `kie-real.jsonl` adds precision-first KIE on **45 real contemporary
pages** (29 checksum-valid NIP values, 6 labeled dates, 20 labeled amounts;
deterministic extraction, ISO dates / dot-decimal amounts, spec-aligned) —
the contemporary corpus is genuinely sparse in these identifiers, so this is
an honest thin auxiliary layer, not a substitute for the synthetic KIE mass.
Historical pages are out of KIE scope (the spec's KIE types are
invoice/contract/official letter/form). Together the synthetic sources cover the task's planned training
composition (~1 200 synthetic) including the photo-degradation family the spec
asks for; real pages (2 011) exceed the planned ~800 — an intentional,
documented deviation.
## Quality gates
- Cross-union dedup on file-sha256 and text-sha256: 1 dropped (ledger
`dropped-dedup.jsonl`).
- Deterministic PII label scan (e-mail, +48 phone, checksummed PL PESEL,
mod-97 PL accounts) on real sources only — synthetic identities are
generated data and PESEL/NIP fields are legitimate training signal:
0 removed (`dropped-pii.jsonl`).
- Overlap screening vs four frozen local eval manifests: **0** against
IMPACT Test A staged, full-page dev v4 and EHRI printed replay. The 208
flags against `qwen3vl-hist-corpus-v2` hit **only its TRAIN pages** (same
source works; 65 pages intentionally shared via the psnc subset) — zero
flags against its validation/test splits. See `overlap-report.json`.
- Splits are document-consistent: docatlas development rows (205 pages)
carried over from the source manifests; synthetic and historical are
train-only; frozen eval collections are excluded by construction.
## Boundaries
- **Nothing is human-verified or gold.** Labels are automated (DocAtlas,
IMPACT ground truth joined per page), generator-produced, or AI
pre-annotations of self-licensed handwritten notes (drawings described in
brackets; `[ucięte]` marks scan-cut edges). The medical notes source is
study/training material with medical knowledge (dosage schedules, disease
patterns as taught); it contains no personal health records — lecturer
surnames appear as professional references. Deterministic PII patterns
(e-mail/phone/PESEL/account) were not found.
- The official PolOCRBench Train/Test A/Test B, evaluator and schemas come
from the organizers ([poleval/2026-polocrbench](https://github.com/poleval/2026-polocrbench));
this pool is independent training material.
- Historical page JPEGs are format-converted derivatives of the IMPACT
dLibra TIFFs (see the source repo for the acquisition audit).
## Files
- `images/` — 3 511 page images, named by pool id (`da<batch>_<row>`,
`hist_<work>__<doc>`, `syn_<id>`, `synph_<id>`).
- `manifest.jsonl` — one row per page: id, source, image/text hashes, split,
license, provenance flags.
- `tables.jsonl`, `kie.jsonl` — subtask B/C annotations for synthetic pages.
- `kie-real.jsonl`, `kie-real-report.json` — deterministic KIE mining on real
pages (`mine_kie_real.py`).
- `dropped-*.jsonl`, `overlap-report.json`, `report.json`, `checksums.json` —
evidence; `assemble_train_pool_v1.py` — deterministic assembly script.