yta-tts-kokoro-onnx 0.0.16

Youtube Autonomous TTS Kokoro Module

Youtube Autonomous TTS Kokoro Onnx Module

This project is to include the TTS functionality using the official kokoro-onnx project (https://github.com/thewh1teagle/kokoro-onnx) which is apparently supporting more languages than the original project.

The models and voices needed are downloaded automatically by the kokoro-onnx dependency from Hugginface. But here are the links just in case you need to download them manually.

Models and voices

Name	Description	Link
kokoro-v1.0.onnx	Model based on Onnx	https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/kokoro-v1.0.onnx
voices-v1.0.bin	Voices	https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/voices-v1.0.bin

Supported languages

• American English
• British English
• Spanish
• French
• Italian
• Japanese
• Chinese
• Hindi
• Portuguese

Supported voices

The voices are specifically created for a language and have a gender, and their name is based on this premise, but can be used in any other different language. Two valid voices are jf_tebukuro and im_nicola.

The first letter of the voice is the language (j for Japanese and i for Italian), and the second is for the gender (f is female and m is male). The rest is just the real name of that voice.

To check all the voices available, check the yta_tts_kokoro_voices dependency in https://github.com/Implosiv3/yta-tts-kokoro-voices.

Models

You need to have the kokoro-v1.0.onnx model file and the voices-v1.0.bin voices file and set the MODEL_PATH and VOICES_PATH set in the .env.

Some results

Here you have one example using the Spanish language: https://www.dailymotion.com/video/xa87qpy