dataset_info:
features:
- name: audio
dtype: audio
- name: text
dtype: string
splits:
- name: train
num_bytes: 66172545707.104
num_examples: 226252
- name: validation
num_bytes: 3377143244.636
num_examples: 11908
- name: test
num_bytes: 3428390565.124
num_examples: 11908
download_size: 73808778897
dataset_size: 72978079516.86398
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
task_categories:
- automatic-speech-recognition
- text-to-speech
language:
- de
size_categories:
- 100K<n<1M
Dataset Card for Bundestag ASR Dataset
This dataset consists of ~610 hours of audio-transcripts from political speeches of the German Bundestag, intended for use in Automatic Speech Recognition (ASR) systems.
Dataset Details
Dataset Description
This dataset contains extensive audio recordings and corresponding transcripts from political speeches delivered in the German Bundestag. It provides a rich resource for training and evaluating ASR systems, particularly in the context of formal and political speech in the German language.
- Language(s) (NLP): German
Dataset Sources
- Repository: https://opendata.iisys.de/dataset/asr-bundestag/
Direct Use
The dataset is suitable for developing, training, and evaluating ASR models, particularly those focusing on German language and political discourse. It can also be used for linguistic research and natural language processing tasks related to speech recognition.
Source Data
The source data consists of recorded speeches from sessions of the German Bundestag. https://www.bundestag.de/mediathek
Bias, Risks, and Limitations
The dataset may reflect biases present in the speeches of politicians, such as political leanings, socio-economic perspectives, and other biases inherent in political discourse. It also focuses on formal speech, which may not generalize to informal speech contexts.