How to Generate Music Using the Ace-Step 1.5 XL Interface
Press play on the video. It'll jump straight to the section that answers the
title above — no need to watch the full video.
Ace-Step 1.5
Audio Generation
Customize
Learn how to configure settings like CPU offload and int8 quantization, and use prompts and lyric tags to create EDM tracks with Ace-Step 1.5 XL.
Turbo vs. SFT Model Differences
When using a Turbo model, you only need 4-8 steps. However, if you use an SFT model, you need to set it between 30-50 steps, which will take significantly longer to complete.
Memory (VRAM) Saving Tips
Disabling the 'Language Model' and enabling 'Int8 Quantization' are the primary keys to success if you have a GPU with less than 16GB-20GB of VRAM.
Model Compilation Reminder
Enabling 'Compile Model' will make subsequent song generations faster, but the initial generation process will be very slow as PyTorch needs to perform the compilation first.
More from Create AI Voice & Music
View All
How to Install Ace-Step 1.5 XL Locally for AI Music
Ace-Step 1.5
Git
Generating full AI songs with lyrics using Ace-Step 1.5
Ace-Step 1.5
Editing lyrics and melodies with the Repaint feature
Ace-Step 1.5
Remixing songs into different genres using AI Cover
Ace-Step 1.5
Build an Interactive Music Sequencer and Synthesizer with Kimi K2 Thinking
Kimi K2 Thinking
Compose Music Tracks and Arrange Loops with Fryderyk AI
Fryderyk