README.md

metadata

annotations_creators:
  - Duygu Altinok
language:
  - tr
license:
  - cc-by-sa-4.0
multilinguality:
  - monolingual
size_categories:
  - 200K<n<300K
source_datasets:
  - original
task_categories:
  - morphology
pretty_name: TrMorphTestser
config_names:
  - lemma
  - cekimli
  - common_lemma
  - common_noun
  - common_verb
dataset_info:
  - config_name: lemma
    features:
      - name: word
        dtype: string
      - name: lemma
        dtype: string
      - name: suffixes
        dtype: string
    splits:
      - name: validation
        num_bytes: 2562460
        num_examples: 42406
    download_size: 2562460
  - config_name: common_lemma
    features:
      - name: word
        dtype: string
      - name: lemma
        dtype: string
      - name: suffixes
        dtype: string
    splits:
      - name: validation
        num_bytes: 32960
        num_examples: 584
  - config_name: cekimli
    features:
      - name: word
        dtype: string
      - name: lemma
        dtype: string
      - name: suffixes
        dtype: string
    splits:
      - name: validation
        num_bytes: 14026516
        num_examples: 163112
  - config_name: common_noun
    features:
      - name: word
        dtype: string
      - name: lemma
        dtype: string
      - name: suffixes
        dtype: string
    splits:
      - name: validation
        num_bytes: 97043
        num_examples: 1396
  - config_name: common_verb
    features:
      - name: word
        dtype: string
      - name: lemma
        dtype: string
      - name: suffixes
        dtype: string
    splits:
      - name: validation
        num_bytes: 251269
        num_examples: 3474
configs:
  - config_name: cekimli
    data_files:
      - split: validation
        path: data/cekimli/validation-*
  - config_name: lemma
    data_files:
      - split: validation
        path: data/lemma/validation-*
  - config_name: common_lemma
    data_files:
      - split: validation
        path: data/common_lemma/validation-*
  - config_name: common_noun
    data_files:
      - split: validation
        path: data/nouns/validation-*
  - config_name: common_verb
    data_files:
      - split: validation
        path: data/verbs/validation-*

Dataset Card for TrMorphTester

This dataset is a testing dataset for Turkish morphology, aiming to calculate how other subword strategies aligns with morphological segmentation of Turkish.

The data is automatically generated from Turkish morpoholigical lexicon. For each row, we offer a surface form, then lemma and all suffixes, separated by a + character.

The dataset has several splits for several purposes:

• lemma: Surface form is same with lemma, no suffixes at all. Nouns, verbs, adjectives, adverbs together.
• common_lemma: Some most common lemmas of Turkish. Nouns, verbs, adjectives, adverbs together.
• common_noun: Inflected nouns.
• common_verb: Inflected verbs.
• cekimli: All sort of inflected words.

This dataset is a part of the Turkish Subwords research collection.