README.md

metadata

dataset_info:
  features:
    - name: src_lang
      dtype: int64
    - name: tgt_lang
      dtype: int64
    - name: src_text
      dtype: string
    - name: tgt_text
      dtype: string
    - name: score
      dtype: float64
  splits:
    - name: train
      num_bytes: 2734319005
      num_examples: 5669543
  download_size: 1362042181
  dataset_size: 2734319005
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Description

General and Technical Domain Data with LaBSE Score >= 0.8 (Bidirectional only with Indic Languages). Includes Samanantar by AI4Bharat and NPTEL Data by Speech Lab.