Icelandic TTS Model - Talr贸mur3
This repository hosts an Icelandic text-to-speech (TTS) model finetuned from fish-speech using the Talr贸mur3 dataset. The model is designed to generate high-quality Icelandic speech.
Model Details
- Base Model: fishaudio/fish-speech-1.5
- Dataset: Talr贸mur3 from Clarin-IS
- Repository: https://github.com/sigurdurhaukur/tts-icelandic
Installation
To use this model, install the required dependencies:
# Create a python 3.10 virtual environment, you can also use virtualenv
conda create -n fish-speech python=3.10
conda activate fish-speech
# Install pytorch
pip3 install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1
# (Ubuntu / Debian User) Install sox + ffmpeg
apt install libsox-dev ffmpeg
# (Ubuntu / Debian User) Install pyaudio
apt install build-essential \
cmake \
libasound-dev \
portaudio19-dev \
libportaudio2 \
libportaudiocpp0
# Install fish-speech
pip3 install -e .[stable]
Download the model
huggingface-cli download Sigurdur/fish-speech-1.5-icelandic --local-dir checkpoints/fish-speech-1.5
Download the fish-speech repo and follow the instructions from here https://speech.fish.audio/inference/
Data
The model was trained on 0.45 hours of speaker F01
with the happy
emotion.
@misc{20.500.12537/344,
title = {Talr{\'o}mur 3 v0.1 (24.09)},
author = {{\"O}rn{\'o}lfsson, Gunnar Thor and Sigurgeirsson, Atli {\TH}{\'o}r and Nikul{\'a}sd{\'o}ttir, Anna Bj{\"o}rk and Schnell, Daniel},
url = {http://hdl.handle.net/20.500.12537/344},
note = {{CLARIN}-{IS}},
copyright = {Creative Commons - Attribution 4.0 International ({CC} {BY} 4.0)},
year = {2024}
}
Model Performance
The model performs horribly on Icelandic, and has lost all previous language skills. The model is probably over-fitting.
Examples
Example 1
Text: "einn, tveir, 镁r铆r, fj贸rir, fimm, dimmalimm."
Example 2
Text: "G贸冒an daginn, 茅g heiti Anna."
Acknowledgments
Special thanks to Clarin-IS for providing the Talr贸mur3 dataset.
Contact
For any questions, please reach out via LinkedIn: Sigur冒ur Haukur Birgisson
- Downloads last month
- 11
Model tree for Sigurdur/fish-speech-icelandic
Base model
fishaudio/fish-speech-1.5