741,000+ skills indexed

CHENHUI-XCOMMUNITY

Explore all skills by this author →

Verified:Website GitHub

heartmula

Generates full songs from lyrics and style tags using open-source HeartMuLa music models.

Updated Jul 3, 2026

Search Tags:#gpu #music generation #audio synthesis #open-source #ai music #heartmula #lyrics to song

One-click install

npx skills add https://github.com/CHENHUI-X/toolbox --skill heartmula-chenhui-x

Or copy as Structured Prompt for Agent▼

Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/CHENHUI-X/toolbox/tree/main/custom-skills/media/heartmula
Command: npx skills add https://github.com/CHENHUI-X/toolbox --skill heartmula-chenhui-x

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Creating original music typically requires composition skills, instruments, or expensive commercial services like Suno. This Skill lets you generate complete songs locally from just lyrics and style tags using the open-source HeartMuLa models, with no cloud subscription required. ## Core Features & Use Cases - Lyrics-to-Song Generation: Produces full MP3 songs (48kHz stereo) from bracketed lyrics and comma-separated style tags using 3B/7B music language models. - Local and Offline Operation: Runs entirely on your own GPU with lazy loading to fit in as little as 8GB VRAM, or split across multiple GPUs. - Multilingual Support: Includes HeartCodec for high-fidelity audio reconstruction, HeartTranscriptor for lyrics transcription, and HeartCLAP for audio-text alignment. - Use Case: A content creator needs a custom background song for a video. They write lyrics with [Verse]/[Chorus] structure, specify tags like "piano,happy,wedding", and generate a 4-minute MP3 in about 4 minutes on a single GPU. ## Quick Start Ask the AI to clone the heartlib repository, install dependencies with the documented patches, download the HeartMuLa checkpoints, and generate a song from your lyrics and tags files.

Frequently Asked Questions about heartmula

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I generate a song from lyrics with HeartMuLa?▼

Write lyrics with bracketed structural tags like [Verse] and [Chorus], create a comma-separated tags file, then run examples/run_music_generation.py with --model_path, --lyrics, --tags, and --save_path arguments. Output is a 48kHz stereo MP3.

What GPU do I need for local AI music generation?▼

HeartMuLa requires a minimum of 8GB VRAM using --lazy_load true, which peaks at about 6.2GB for the 3B model. 16GB+ VRAM is recommended, and you can split models across GPUs with --mula_device and --codec_device flags.

Can HeartMuLa run on CPU or macOS?▼

CPU mode works with --mula_device cpu but is extremely slow, taking 30-60+ minutes per song versus about 4 minutes on GPU. Triton is unavailable on macOS, so GPU acceleration is Linux/CUDA only.

Why does HeartMuLa fail after installing dependencies?▼

The pinned dependencies conflict with newer packages, requiring upgrades to datasets and transformers plus two source patches: RoPE cache reinitialization in modeling_heartmula.py and ignore_mismatched_sizes=True in HeartCodec loading calls.

What are the limitations of HeartMuLa song generation?▼

Tags may be ignored since lyrics tend to dominate generation, a known upstream issue. RTX 5080 incompatibility has been reported, and using bfloat16 for HeartCodec degrades audio quality, so keep the codec at float32.

Related Skills

Explore More Design & Creative Skills→

suno

Write Suno AI music prompts with style tags, metatag song structures, and lyrics.

KaiserWhoLearns

FFmpeg Audio Processing

Extract, normalize, and mix audio from video files using FFmpeg commands.

music-lyrics

Extract timestamped lyrics from audio using faster-whisper.

ai-music-generation

Generates music and songs with AudioCraft and HeartMuLa, and analyzes audio via spectrograms.