README.md

metadata

license: cc-by-sa-4.0
task_categories:
  - text-classification
task_ids:
  - multi-class-classification
language:
  - cs
multilinguality:
  - monolingual
size_categories:
  - 1K<n<10K
pretty_name: Czech Verse Form Classification
tags:
  - poetry
  - czech
  - form-classification
  - poetic-form
  - text-classification
  - mteb
  - poetrymteb
  - embedding-evaluation
annotations_creators:
  - found
source_datasets:
  - riazhsks/multilingual-annotated-poems
  - versotym/corpusCzechVerse
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
    default: true
dataset_info:
  - config_name: default
    features:
      - name: id
        dtype: string
      - name: poem
        dtype: string
      - name: title
        dtype: string
      - name: author
        dtype: string
      - name: label
        dtype: int64
      - name: label_name
        dtype: string
    splits:
      - name: train
        num_examples: 7018
      - name: test
        num_examples: 1754

Czech Verse Form Classification

Single-label Czech poetic form classification for PoetryMTEB, using gold form labels from the Corpus of Czech Verse (CCV) subset packaged in multilingual-annotated-poems.

Only poems with a non-empty original form label are kept (most CCV poems are unlabeled for form). Rule-based auto-form labels are not used.

Dataset Card

Item	Description
Source	https://github.com/riazhsks/multilingual-annotated-poems → corpora/czech_poems_processed_ccv.jsonl; upstream CCV
Languages	Czech (cs)
Size	train=7018; test=1754 (no validation)
Classes	19 forms with frequency ≥ 10
Filtering	Gold form only; length ∈ [40, 8000]; rare forms dropped
Splits	Stratified by form ≈ 80% / 19%, seed=42
License	CC BY-SA 4.0 (packaging; cite CCV)
Dataset version	1.0.0
Evaluation metrics	accuracy, macro/weighted F1

Features

Field	Type	Description
id	string	Example id
poem	string	Poem text
title	string	Title
author	string	Author
label	int64	Form class index
label_name	string	Czech form name

Label inventory (kept)

id	label_name	gloss	train	test	source_count*
0	sonet	Sonnet (十四行)	4483	1121	5604
1	strofa venuše a adonise	Venus and Adonis stanza	578	145	723
2	ritornel	Ritornello	534	133	667
3	madrigal	Madrigal	472	118	590
4	hrdinský kuplet	Heroic couplet	170	43	213
5	stance	Stance / stanza form	169	42	211
6	siciliána	Siciliana	108	27	135
7	gazel	Ghazal	85	21	106
8	elegické distichon	Elegiac couplet	84	21	105
9	sapfická strofa	Sapphic stanza	73	18	91
10	rondó	Rondeau	69	17	86
11	tercína	Terza rima	43	11	54
12	rondel	Rondel	42	11	53
13	sestina	Sestina	31	8	39
14	sonet anglický	English sonnet	29	7	36
15	limerik	Limerick	20	5	25
16	alkajská strofa	Alcaic stanza	10	2	12
17	spenserova strofa	Spenserian stanza	10	2	12
18	chaucerova strofa	Chaucerian stanza / rhyme royal	8	2	10

Dropped forms (count < 10)

form	count
asklepiadská strofa 4	6
kasída	4
arte mayor	3
triolet	2
burnsova strofa	2
huitain	1

Construction method

1. Load CCV JSONL from multilingual-annotated-poems.
2. Keep rows with gold form + text (skip null form).
3. Length filter; keep classes ≥ 10.
4. Frequency-descending label ids; stratified train/test.

How to load

from datasets import load_dataset

ds = load_dataset("PoetryMTEB/CzechVerseFormClassification")
print(ds["test"][0]["label_name"])

Citation

@article{ccv2015,
  author  = {Plecháč, Petr and Kolár, Robert},
  title   = {The Corpus of Czech Verse},
  journal = {Studia Metrica et Poetica},
  volume  = {2},
  number  = {1},
  year    = {2015},
  pages   = {107--118},
  doi     = {10.12697/smp.2015.2.1.05}
}

@misc{riazhsks_multilingual_annotated_poems,
  title={Multilingual Annotated Poems},
  author={riazhsks},
  year={2025},
  url={https://github.com/riazhsks/multilingual-annotated-poems}
}

Hub packaging: PoetryMTEB/CzechVerseFormClassification (v1.0.0).