MuzWack

Inspiration

1. We are new friends who resonated over music.
2. Today is Father's Day.
3. Inspired when writing corny, "dad-humor" jokes into GPT-4.

What it does

1. Our project sings a song about a user-inputted topic entirely generated by AI, including lyrics and melody.
2. It can also generate humorously unexpected songs without input. 3.

How we built it

1. We used OpenAI's GPT-4 to write the lyrics and generate the melody.
2. We used Microsoft Azure for voice synthesis. 3.

Challenges we ran into

1. We tried to utilize all of the amazing tools we were given access to, but found ourselves with too many systems.
2. We learned about 3.

Accomplishments that we're proud of

1. 2. 3.

What we learned

1. We learned skills & techniques from the amazing mentors, and learned more about the capabilities of the emerging systems + APIs available from industry leaders like OpenAi.
2. We learned how to combine the different frameworks.
3. We learned how to be flexible and pivot from ideas.

What's next for MuzWack

1. We want to add in background tracks, such as percussion, guitar, and variable rhythms.
2. We want to emulate Cameo by adding customizability towards the singer's voice, for instance allowing the user to generate a song sung by Taylor Swift.
3. We could add a front-end interface that shows the animated face of the selected singer while they perform the song. https://huggingface.co/QuickWick/Music-AI-Voices/

Built With

• anyscale
• azure
• gpt-4