filtered_Orca / README.md
Sayan1997's picture
The Orca data is filtered out for responses less than 100 tokens
c8b7a19
metadata
dataset_info:
  features:
    - name: id
      dtype: string
    - name: system_prompt
      dtype: string
    - name: question
      dtype: string
    - name: response
      dtype: string
  splits:
    - name: train
      num_bytes: 2074578643.6622322
      num_examples: 1216347
  download_size: 1515594488
  dataset_size: 2074578643.6622322
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: mit
task_categories:
  - text-generation
language:
  - en
pretty_name: Orca_filtered
size_categories:
  - 10K<n<100K