README.md

metadata

dataset_info:
  features:
    - name: source_url
      dtype: string
    - name: data_type
      dtype: string
    - name: product_id
      dtype: string
    - name: product_name
      dtype: string
    - name: brand
      dtype: string
    - name: rating
      dtype: float64
    - name: text
      dtype: string
    - name: messages
      list:
        - name: role
          dtype: string
        - name: content
          dtype: string
  splits:
    - name: train
      num_bytes: 31900429
      num_examples: 35898
    - name: test
      num_bytes: 3557471
      num_examples: 3987
  download_size: 28210533
  dataset_size: 35457900
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
task_categories:
  - question-answering
language:
  - en
tags:
  - Cosmetic
  - Chemistry
pretty_name: Soma - Cosmetic Q&A
size_categories:
  - 10K<n<100K