multimodal-art-projection / YuE

TL;DR

YuE2 introduces a novel approach to music generation that combines symbolic planning with zero-shot cover creation and agentic music editing, enabling users to create diverse musical pieces efficiently.

✦ Why It Matters

Explore YuE2 today to enhance your music creation process with AI-driven tools.

Key Takeaways

Explore YuE2's capabilities for generating songs from lyrics and style prompts.
Utilize the white-box interface to edit melodies and chords before rendering.
Leverage zero-shot cover functionality to reinterpret existing songs in new styles.
Integrate YuE2 into your projects using the provided Python API.

How It Works

YuE2 employs a Mixture-of-Transformers architecture to predict musical scores and semantic tokens, generating acoustic latents through flow matching. A Variational Autoencoder (VAE) then decodes these latents into high-fidelity stereo audio.

Users can interact with the model through a Python API, allowing for detailed control over the music generation process.

multimodal-art-projection / YuE

BerriAI / litellm

tensorflow / tensorflow

open-webui / open-webui

Agentic data science with the Data Agent Kit

BerriAI / litellm

tensorflow / tensorflow

open-webui / open-webui

Agentic data science with the Data Agent Kit