hylcool's picture
Update README.md
409d491
metadata
license: apache-2.0
task_categories:
  - text-generation
language:
  - en
tags:
  - data-juicer
  - pretraining
size_categories:
  - 10K<n<100K

The Pile -- EuroParl (refined by Data-Juicer)

A refined version of EuroParl dataset in The Pile by Data-Juicer. Removing some "bad" samples from the original dataset to make it higher-quality.

This dataset is usually used to pretrain a Large Language Model.

Notice: Here is a small subset for previewing. The whole dataset is available here (About 2.2GB).

Dataset Information

  • Number of samples: 61,601 (Keep ~88.23% from the original dataset)

Refining Recipe

# global parameters
# global parameters
project_name: 'Data-Juicer-recipes-EuroParl'
dataset_path: '/path/to/your/dataset'  # path to your dataset directory or file
export_path: '/path/to/your/dataset.jsonl'

np: 50  # number of subprocess to process your dataset
open_tracer: true

# process schedule
# a list of several process operators with their arguments
process:
  - clean_email_mapper:
  - clean_links_mapper:
  - fix_unicode_mapper:
  - punctuation_normalization_mapper:
  - whitespace_normalization_mapper:

  - alphanumeric_filter:
      tokenization: false
      min_ratio: 0.75  # <3sigma (0.779)
      max_ratio: 0.90 # >3sigma(0.878)
  - average_line_length_filter:  # for code
      max_len: 588  # 3sigma
  - character_repetition_filter:
      rep_len: 10
      max_ratio: 0.16  # >3sigma (0.114)
  - flagged_words_filter:
      lang: en
      tokenization: true
      max_ratio: 0.0007  # 3sigma
  - language_id_score_filter:
      min_score: 0.7
  - maximum_line_length_filter:  # for code
      max_len: 4000  # >3sigma (3104)
  - perplexity_filter:
      lang: en
      max_ppl:  7596 #(3sigma)
  - special_characters_filter:
      max_ratio: 0.3  # > 3sigma (0.243)
  - text_length_filter:
      max_len: 2e5
  - words_num_filter:
      tokenization: true
      min_num: 20
      max_num: 1e5  # 3sigma
  - word_repetition_filter:
      lang: en
      tokenization: true
      rep_len: 10
      max_ratio: 0.2  # > 3sigma (0.185)

  - document_simhash_deduplicator:
      tokenization: space
      window_size: 6
      lowercase: true
      ignore_pattern: '\p{P}'
      num_blocks: 6
      hamming_distance: 4