Voice Sample & Text to Voice [Dub3]

This module is contributed by Fabri from Dub3


This module takes a voice sample as an input and can then output another voice sample for a given text prompt - in multiple languages.

OSS Model

The OSS Model is based on this model: https://github.com/coqui-ai/TTS

Lilypad Module

Run in CLI

First upload a voice .wav file to IPFS

Use https://web3.storage is an easy way to upload files to IPFS. Use https://cloudconvert.com/ to convert recordings to .wav files

Here's are two voice recording examples:


Then use the command below to run the module. Note the parameters for changing the language and adding a prompt.

lilypad run github.com/dub3-space/lily-synthetizoor:b4d2e078c3447e1df12588ee854360087ce7f081 -i URL=https://bafybeibrum5bqhslpoemqiyvikbg3b5ujte6qppvk2cbdstwxnhejypxl4.ipfs.w3s.link/ally-voice.wav -i SENTENCE='Hello! Youre listening to voice translation on Lilypad with dub3!' -i LANGUAGE=en -i INPUT_FOLDER=/inputs -i OUTPUT_FOLDER=/outputs

You should see Lilypad run something like this:

You'll find your results under the output folder:

Last updated