README.md

metadata

dataset_info:
  - config_name: cs-en
    features:
      - name: translation
        struct:
          - name: cs
            dtype: string
          - name: en
            dtype: string
    splits:
      - name: train
        num_bytes: 3432181
        num_examples: 12076
      - name: validation
        num_bytes: 318813
        num_examples: 1002
    download_size: 0
    dataset_size: 3750994
  - config_name: de-en
    features:
      - name: translation
        struct:
          - name: de
            dtype: string
          - name: en
            dtype: string
    splits:
      - name: train
        num_bytes: 4108729
        num_examples: 14211
      - name: validation
        num_bytes: 329855
        num_examples: 1002
    download_size: 0
    dataset_size: 4438584
  - config_name: is-en
    features:
      - name: translation
        struct:
          - name: is
            dtype: string
          - name: en
            dtype: string
    splits:
      - name: train
        num_bytes: 554190
        num_examples: 2009
    download_size: 0
    dataset_size: 554190
  - config_name: ru-en
    features:
      - name: translation
        struct:
          - name: ru
            dtype: string
          - name: en
            dtype: string
    splits:
      - name: train
        num_bytes: 5427552
        num_examples: 15000
      - name: validation
        num_bytes: 442271
        num_examples: 1002
    download_size: 0
    dataset_size: 5869823
  - config_name: zh-en
    features:
      - name: translation
        struct:
          - name: zh
            dtype: string
          - name: en
            dtype: string
    splits:
      - name: train
        num_bytes: 4700299
        num_examples: 15406
      - name: validation
        num_bytes: 285969
        num_examples: 1002
    download_size: 0
    dataset_size: 4986268
configs:
  - config_name: cs-en
    data_files:
      - split: train
        path: cs-en/train-*
      - split: validation
        path: cs-en/validation-*
  - config_name: de-en
    data_files:
      - split: train
        path: de-en/train-*
      - split: validation
        path: de-en/validation-*
  - config_name: is-en
    data_files:
      - split: train
        path: is-en/train-*
  - config_name: ru-en
    data_files:
      - split: train
        path: ru-en/train-*
      - split: validation
        path: ru-en/validation-*
  - config_name: zh-en
    data_files:
      - split: train
        path: zh-en/train-*
      - split: validation
        path: zh-en/validation-*

Dataset Card for "ALMA-Human-Parallel"

This is human-written parallel dataset used by ALMA translation models.

@misc{xu2023paradigm,
      title={A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models}, 
      author={Haoran Xu and Young Jin Kim and Amr Sharaf and Hany Hassan Awadalla},
      year={2023},
      eprint={2309.11674},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{xu2024contrastive,
      title={Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation}, 
      author={Haoran Xu and Amr Sharaf and Yunmo Chen and Weiting Tan and Lingfeng Shen and Benjamin Van Durme and Kenton Murray and Young Jin Kim},
      year={2024},
      eprint={2401.08417},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}