README.md

metadata

license: afl-3.0
task_categories:
  - text2text-generation
language:
  - en
tags:
  - recipe
size_categories:
  - 100K<n<1M
source_datasets:
  - lishuyang/recipepairs
configs:
  - config_name: default
    data_files:
      - split: train
        path: p_train.parquet
      - split: test
        path: p_test.parquet
      - split: validation
        path: p_val.parquet
  - config_name: mini
    data_files:
      - split: train
        path: new_train.parquet
      - split: test
        path: new_test.parquet
      - split: validation
        path: new_val.parquet

train : 64K pairs test&validate : 8K pairs 8 : 1 : 1

mini train : 8K pairs t&v : 800 pairs