ILSpeech

Hebrew Single-Speaker TTS Dataset 🗣️

Dataset content

This dataset contains around 40 minutes of studio-quality speech from a single male speaker.

It includes clean text and phonemes. The audio is in 22.04 kHz mono format with clean recordings.

The total dataset size is approximately 120 MB.

Sample

Format

Phoneme annotations are provided in LJSpeech format file_id|phonemes inside metadata.csv file (LJSpeech format).

Example: 1|ma hamatsav where 1 will point to ./wav/1.wav. each wav length is 2 to 25 seconds.

and the phonemization was done using https://github.com/thewh1teagle/mishkal, with manual corrections for accuracy.

License

The dataset is licensed under MIT, allowing for commercial use, provided the use is appropriate and does not harm anyone.

See License.pdf

Setup

wget https://huggingface.co/datasets/thewh1teagle/ILSpeech/resolve/main/ilspeech_2025_04_v1.zip
unzip ilspeech_2025_04_v1.zip

Training

You can find training receipes for the dataset in github.com/thewh1teagle/ilspeech-train

Phonemize

The dataset comes with phonemes already so you don't need this for train,

However you do need it after you train the model.

you can phonemize the text using mishkal:

from mishkal import phonemize

phonemes = phonemize("שָׁלוֹם עוֹלָם", schema='modern')
print(phonemes)