ACE-Step 1.5: The Open-Source Model Challenging Suno in Music Generation

ACE-Step 1.5 is presented as an open-source alternative to Suno for music generation.

Key Features

• Quality: Claims to surpass Suno in standard evaluation scores.
• Speed: Generates a full song in under 2 seconds on an A100 GPU.
• Local Execution: Requires approximately 4GB of VRAM, with generation times under 10 seconds on an RTX 3090.
• LoRA: Allows customization of style with a few songs.
• License: MIT, allowing free commercial use.
• Data: Trained on authorized and synthetic data.

The training code, model weights, and LoRA code are available on GitHub. For those evaluating on-premise deployments, there are trade-offs to consider, and AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate the different options.