filtered-python-edu / README.md
styal's picture
Update README.md
4b40ea7 verified
metadata
dataset_info:
  features:
    - name: blob_id
      dtype: string
    - name: repo_name
      dtype: string
    - name: path
      dtype: string
    - name: length_bytes
      dtype: int64
    - name: score
      dtype: float64
    - name: int_score
      dtype: int64
    - name: text
      dtype: string
    - name: is_english
      dtype: bool
  splits:
    - name: train
      num_bytes: 1226088650
      num_examples: 1018270
  download_size: 626200552
  dataset_size: 1226088650
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Dataset created with this filter from Avelina/python-edu.

def my_filter(example):
    score = example["score"] > 4.1
    lengh = example["length_bytes"] < 3000 and example["length_bytes"] > 200
    return score and lengh

A is_english boolean have been added with the model papluca/xlm-roberta-base-language-detection.