Inspiration
Took Inspiration from the real pain of content creators and limitations of viewership due to language barrier.
What it does
It translates/dubs the content from one language into another, in the original voice of the content creator using voice cloning with the emotions & excitement that the original voice has.
How we built it
It runs on powerful GPUs, I have been using google Colab. Using two models CosyVoice and Sense vocie, it extracts texts from audio file with emotions mentioned in the text, then reads the text in the sample voice of the original creator.
Challenges we ran into
Due to limited access to T4 GPUs, I had to change the accounts I use to run colab and that change costs a lot of time. I had to run the same colab cells into different notebooks to make sure everything works.
Accomplishments that we're proud of
I am proud that I have created a real solution to a real problem.
What we learned
I learn about Ngrok more, since I had to create a UI frontend local and plug the backend of colab with it.
What's next for YouDub
I would like to win this hackathon and use the winning amount to run this project on a privately hosted GPU, outreaching content creators on YouTube to be my initial clients.
Log in or sign up for Devpost to join the conversation.