metadata

license: cc-by-4.0
task_categories:
  - automatic-speech-recognition
  - text-to-speech
language:
  - vi
pretty_name: VietMed unlabeled set
size_categories:
  - 100K<n<1M
dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: Metadata ID
      dtype: string
  splits:
    - name: train
      num_bytes: 57670081699.38
      num_examples: 230516
  download_size: 51899577807
  dataset_size: 57670081699.38
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

unofficial mirror of VietMed (Vietnamese speech data in medical domain) unlabeled set

official announcement: https://arxiv.org/abs/2404.05659

official download: https://huggingface.co/datasets/leduckhai/VietMed

this repo contains the unlabeled set: 966h - 230k samples

i also gather the metadata: see info.csv

my extraction code: https://github.com/phineas-pta/fine-tune-whisper-vi/blob/main/misc/vietmed-unlabeled.py

need to do: check misspelling, restore foreign words phonetised to vietnamese

usage with HuggingFace:

# pip install -q "datasets[audio]"
from datasets import load_dataset
from huggingface_hub import hf_hub_download
from pandas import read_csv

repo_id = "doof-ferb/VietMed_unlabeled"
dataset = load_dataset(repo_id, split="train", streaming=True)
info_file = hf_hub_download(repo_id=repo_id, filename="info.csv", repo_type="dataset")
info_dict = read_csv(info_file, index_col=0).to_dict("index")

def merge_info(batch):
    meta = info_dict.get(batch["Metadata ID"], "")
    if meta != "":
        batch["Domain"]      = meta["Domain"]
        batch["ICD-10 Code"] = meta["ICD-10 Code"]
        batch["Accent"]      = meta["Accent"]
    else:
        batch["Domain"]      = ""
        batch["ICD-10 Code"] = ""
        batch["Accent"]      = ""
    return batch
dataset = dataset.map(merge_info)