README.md

---
dataset_info:
- config_name: corpus
  features:
  - name: docid
    dtype: string
  - name: title
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 9772890386
    num_examples: 3185312
  download_size: 4854439740
  dataset_size: 9772890386
- config_name: reddit_tomt
  features:
  - name: query_id
    dtype: string
  - name: query
    dtype: string
  - name: positive_passages
    list:
    - name: docid
      dtype: string
    - name: text
      dtype: string
    - name: title
      dtype: string
  - name: hard_negative_passages
    sequence: 'null'
  - name: negative_passages
    list:
    - name: docid
      dtype: string
    - name: text
      dtype: string
    - name: title
      dtype: string
  splits:
  - name: train
    num_bytes: 8080604019
    num_examples: 9455
  - name: val
    num_bytes: 995934250
    num_examples: 1186
  - name: test
    num_bytes: 12058548
    num_examples: 1180
  download_size: 4197412538
  dataset_size: 9088596817
- config_name: trec_tot
  features:
  - name: query_id
    dtype: string
  - name: query
    dtype: string
  - name: positive_passages
    list:
    - name: docid
      dtype: string
    - name: text
      dtype: string
    - name: title
      dtype: string
  - name: hard_negative_passages
    sequence: 'null'
  - name: negative_passages
    list:
    - name: docid
      dtype: string
    - name: text
      dtype: string
    - name: title
      dtype: string
  splits:
  - name: train
    num_bytes: 266489376
    num_examples: 150
  - name: dev1
    num_bytes: 261848666
    num_examples: 150
  - name: dev2
    num_bytes: 1539542
    num_examples: 150
  download_size: 231138975
  dataset_size: 529877584
configs:
- config_name: corpus
  data_files:
  - split: train
    path: corpus/train-*
- config_name: reddit_tomt
  data_files:
  - split: train
    path: reddit_tomt/train-*
  - split: val
    path: reddit_tomt/val-*
  - split: test
    path: reddit_tomt/test-*
- config_name: trec_tot
  data_files:
  - split: train
    path: trec_tot/train-*
  - split: dev1
    path: trec_tot/dev1-*
  - split: dev2
    path: trec_tot/dev2-*
---