Tag
VIBE is an AI system that generates background music for videos by aligning with text instructions, using continuous latent modeling and reinforcement learning for better synchronization with video mood, timing, and genre.
HeartMuLa is an open-source family of AI models for music generation that converts lyrics and tags into full songs, including a language model, music codec, lyrics transcriber, and audio-text alignment model, with demos available on Hugging Face Spaces and ModelScope.
I trained a 1.2B DiT model for instrumental game music generation, using Stable Audio's VAE and aiming to cover diverse styles. The project is open-source with a WebUI and samples available on HuggingFace.
An open-source, self-hosted AI music generation radio tool called TAPEDECK that uses MiniMax Music 3 to create endless, personalized radio stations from a user's music library on a local GPU.
Qwen-Music is a new paper that splits music generation into two stages: planning with compact semantic tokens and CoT reasoning over melody, then rendering high-fidelity audio, achieving state-of-the-art results.
Google introduces Lyria 3 Pro, an advanced music generation model capable of creating tracks up to 3 minutes long with greater structural control. The model is integrated into multiple Google products including Vertex AI, Google AI Studio, Gemini, and Google Vids.
Google Senior Product Manager Joel Yawili discusses the features and philosophy behind Lyria 3 and Lyria 3 Pro, highlighting improvements in song length, structure control, and multimodal integration with Gemini.