Skip to content

Dataset · Hebrish

Noisy condition speech recordings in English and Hebrew

Hebrish/Noisy-Condition-Sentences

An audio dataset of read-aloud sentences in noisy environments in English and Hebrew for ASR tasks.

What's inside

The collection contains between 10k and 100k audio and text records in English and Hebrew. It is a dedicated subset of the master dataset danielrosehill/Read-Aloud-Voice-Sentences, selecting only duplicates recorded under noisy conditions (outdoors, with background speech, etc.). The authors note that the dataset includes recordings from Hebrish/Background-Speech and is designed for convenient inspection of specific tasks, while the main repository is recommended for training.

What you can do with it

  • Evaluate ASR model robustness to background noise in English and Hebrew
  • Fine-tune speech recognition systems for real-world outdoor environments
Tasks
Automatic speech recognition
Languages
English, Hebrew
Size
10K–100K rows
Downloads this month
67.8K
Likes
0
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("Hebrish/Noisy-Condition-Sentences")