Willy030125/whisper_large_v3_turbo_noise_redux_v1
Automatic Speech Recognition
•
Updated
•
34
Dataset Viewer
The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
Ambient Noise Collection dataset
This dataset is useful for reducing ASR model Hallucinations (especially Whisper), by default Whisper often transcribing non-speech audio as hallucination transcriptions.
This dataset attempts to improve ASR model that have hallucinations on non-speech audio.
Collected several audio from source:
- https://www.kaggle.com/datasets/nafin59/hospital-ambient-noise
- https://www.kaggle.com/datasets/solorzano/ambient-noise
- https://www.kaggle.com/datasets/minsithu/audio-noise-dataset
- https://www.kaggle.com/datasets/ivanj0/audiodata
- https://huggingface.co/datasets/Myrtle/CAIMAN-ASR-BackgroundNoise
Dataset info:
num_rows: 147
Format
Each example is a dictionary with the following fields:
{
"path": "audio/ambient_noise_1.wav",
"audio": {
"path": "audio/ambient_noise_1.wav",
"array": [...],
"sampling_rate": 16000
},
"sentence": "Some transcription"
}
Load dataset
Use HuggingFace datasets v2.18:
pip install datasets==2.18.0
Use HuggingFace datasets to load:
from datasets import load_dataset, Audio
dataset = load_dataset("Willy030125/ambient_noise_audio")
- Downloads last month
- 55