c4-ja-5k-metrics / README.md
minato-ryan's picture
Upload dataset
cd1b52a verified
|
raw
history blame
3.28 kB
metadata
dataset_info:
  - config_name: all-nlp
    features:
      - name: text
        dtype: string
      - name: language
        dtype: string
      - name: confidence
        struct:
          - name: ARABIC
            dtype: float64
          - name: CHINESE
            dtype: float64
          - name: DUTCH
            dtype: float64
          - name: ENGLISH
            dtype: float64
          - name: FRENCH
            dtype: float64
          - name: GERMAN
            dtype: float64
          - name: ITALIAN
            dtype: float64
          - name: JAPANESE
            dtype: float64
          - name: POLISH
            dtype: float64
          - name: PORTUGUESE
            dtype: float64
          - name: RUSSIAN
            dtype: float64
          - name: SPANISH
            dtype: float64
          - name: TURKISH
            dtype: float64
      - name: lines_with_no_ending_punctuation
        struct:
          - name: label
            struct:
              - name: no_ending
                dtype: int64
          - name: mask
            sequence:
              sequence: int64
      - name: lines_with_too_few_words
        struct:
          - name: label
            struct:
              - name: too_few
                dtype: int64
          - name: mask
            sequence:
              sequence: int64
      - name: has_naughty_word
        dtype: bool
      - name: naughty_words
        sequence: string
      - name: has_javascript
        dtype: bool
      - name: has_lorem_ipsum
        dtype: bool
      - name: has_curly_brace
        dtype: bool
      - name: line_count
        dtype: int64
      - name: character_count
        dtype: int64
      - name: token_count
        dtype: int64
      - name: text_count
        dtype: int64
      - name: sentence_counts
        sequence:
          sequence: int64
    splits:
      - name: train
        num_bytes: 60566536
        num_examples: 5000
    download_size: 27094942
    dataset_size: 60566536
  - config_name: dsir
    features:
      - name: text
        dtype: string
      - name: tokens
        sequence: string
      - name: weight
        dtype: float32
      - name: prob_dists
        sequence: float32
    splits:
      - name: train
        num_bytes: 484592779
        num_examples: 5000
      - name: percent_0_5
        num_bytes: 24229638.95
        num_examples: 250
      - name: percent_20_25
        num_bytes: 24229638.95
        num_examples: 250
      - name: percent_40_45
        num_bytes: 24229638.95
        num_examples: 250
      - name: percent_60_65
        num_bytes: 24229638.95
        num_examples: 250
      - name: percent_80_85
        num_bytes: 24229638.95
        num_examples: 250
      - name: percent_95_100
        num_bytes: 24229638.95
        num_examples: 250
    download_size: 331722089
    dataset_size: 629970612.7000002
  - config_name: dsir-domain
    features:
      - name: text
        dtype: string
      - name: tokens
        sequence: string
      - name: weight
        dtype: float32
      - name: prob_dists
        sequence: float32
    splits:
      - name: train
        num_bytes: 484592779
        num_examples: 5000
    download_size: 74755598
    dataset_size: 484592779
configs:
  - config_name: all-nlp
    data_files:
      - split: train
        path: all-nlp/train-*
  - config_name: dsir
    data_files:
      - split: train
        path: dsir/train-*
      - split: percent_0_5
        path: dsir/percent_0_5-*
      - split: percent_20_25
        path: dsir/percent_20_25-*
      - split: percent_40_45
        path: dsir/percent_40_45-*
      - split: percent_60_65
        path: dsir/percent_60_65-*
      - split: percent_80_85
        path: dsir/percent_80_85-*
      - split: percent_95_100
        path: dsir/percent_95_100-*
  - config_name: dsir-domain
    data_files:
      - split: train
        path: dsir-domain/train-*