Edit model card

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

Overview

This model is a state-of-the-art text-to-speech system designed for the Central Kurdish (ckb) language. This model converts written text into natural and intelligible spoken words, providing an essential tool for various applications such as accessibility, language learning, and content creation.

Datasets

  • Dataset Name: PawanKrd/tts-ckb
  • Description: This dataset comprises a diverse collection of Central Kurdish texts and corresponding audio recordings, meticulously curated to train and evaluate this model.

Model Performance

This model has been evaluated on the PawanKrd/tts-ckb dataset, achieving the following performance metrics:

  • Loss: 19.1

    • Explanation: This metric measures the discrepancy between the predicted audio and the actual audio. A lower loss indicates better performance.
  • SER (Sentence Error Rate): 34.66%

    • Explanation: This metric quantifies the percentage of sentences with errors in the generated speech. A lower SER indicates more accurate and reliable speech synthesis.

Applications

This model can be utilized in various scenarios, including but not limited to:

  • Accessibility: Enhancing the accessibility of written content for individuals with visual impairments.
  • Language Learning: Assisting learners in acquiring proper pronunciation and listening skills in Central Kurdish.
  • Content Creation: Enabling the creation of audio content such as podcasts, audiobooks, and announcements in Central Kurdish.
Downloads last month
56
Safetensors
Model size
36.3M params
Tensor type
F32
·

Dataset used to train PawanKrd/tts-ckb

Evaluation results