---
dataset_info:
  features:
  - name: path
    dtype: string
  - name: sentence
    dtype: string
  - name: duration
    dtype: float64
  - name: audio
    dtype: audio
  splits:
  - name: train
    num_bytes: 299832461.0
    num_examples: 637
  download_size: 296243761
  dataset_size: 299832461.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

## TVShow dataset

This dataset contains more than 600 audio samples from italian TV shows, manually annotated. The annotation process was performed using [this repository](https://github.com/alessiodevoto/custom-asr-dataset)