README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: string
    - name: audio
      dtype: audio
    - name: subtitle
      list:
        - name: duration
          dtype: float64
        - name: start
          dtype: float64
        - name: text
          dtype: string
    - name: language
      dtype: string
    - name: title
      dtype: string
  splits:
    - name: train
      num_bytes: 51921294003.95
      num_examples: 2675
  download_size: 50290433886
  dataset_size: 51921294003.95
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Large-scale Joint Weakly-Supervised and Instruction Learning for Medical Speech Translation

This repo is under development. Stay tuned!