Show HN: Generate beats for your song in seconds using Latent Diffusion Models

Showing this on behalf of a friend here. A friend has built a copilot for music creation - CassetteAI. It generates music from text using the inputs you provide to it. It uses a Latent Diffusion Model trained on just instrumental audio to "imagine" music in its latent space to generate beats. Would love to hear your feedback!