---
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: subtitle
    list:
    - name: duration
      dtype: float64
    - name: start
      dtype: float64
    - name: text
      dtype: string
  - name: title
    dtype: string
  - name: phrase_index
    dtype: string
  - name: id
    dtype: string
  - name: language
    dtype: string
  splits:
  - name: train
    num_bytes: 1117623136.0
    num_examples: 50
  download_size: 1092785047
  dataset_size: 1117623136.0
configs:
- config_name: default
  data_files:
  - split: train
    path: de/train-*
---
# Large-scale Joint Weakly-Supervised and Instruction Learning for Medical Speech Translation

This repo is under development. Stay tuned!