jangmin's picture
Update README.md
00912d9
metadata
language:
  - ko
size_categories:
  - 10K<n<100K
dataset_info:
  features:
    - name: user_id
      dtype: int64
    - name: day
      dtype: string
    - name: order_ts
      dtype: string
    - name: positive_prod_id
      dtype: int64
    - name: negative_prod_id
      dtype: int64
    - name: negative_prod_ids
      sequence: int64
    - name: chosen
      dtype: string
    - name: rejected
      dtype: string
  splits:
    - name: train
      num_bytes: 122282877.9602969
      num_examples: 58535
    - name: test
      num_bytes: 52690471.08509643
      num_examples: 17332
    - name: rigorous_test
      num_bytes: 24661037.47070749
      num_examples: 8112
  download_size: 33220918
  dataset_size: 199634386.51610082
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
      - split: rigorous_test
        path: data/rigorous_test-*

Dataset Card for "ecommerce_purchase_history"

Dataset Description

Dataset Summary

이 데이터셋은 특정 이커머스 회사의 추천 시스템 연구 개발을 위한 데이터셋이다. 특정 기간에 대해 약 90일 동안의 구매 히스토리로부터 생성되었다. 구매 히스토리를 텍스트로 기술하였다.

llama2 토크나이저 기준 2,048 개의 토큰 미만의 예제 쌍만을 남기도록 수정하였다.

또한, test 스플릿의 경우 user_id, positive_prod_id 기준으로 train_split에 등장하지 않는 것만을 남겼다.

Supported Tasks and Leaderboards

Languages

This dataset is only made of ko(korean).

Dataset Structure