Ace-Step 1.5 Concept Sliders Collection Released
Test the provided sliders on Ace‑Step 1.5 to verify compatibility.
Test the sliders on your Ace‑Step 1.5 setup and integrate them into your workflow.
Summary
Ace‑Step 1.5 now supports concept sliders that can be trained in 10 to 60 minutes without the need for dedicated datasets. The author has edited the AI Toolkit code to enable native support for Slider Loras, allowing users to train new sliders quickly. The collection includes 15 pre‑trained sliders: male‑to‑female voice, studio‑to‑lofi, Bass boost, Choir‑to‑solo vocalist, digital‑to‑acoustic sound, Aggressive‑to‑gentle, drum intensity, energetic‑to‑calm, happiness, soft‑to‑projected voice, talking‑to‑singing, tempo, and danceability. Each slider can be applied directly in Ace‑Step 1.5 pipelines, eliminating the need for custom datasets. The author plans to expand the collection based on community requests. This resource is valuable for developers looking to add dynamic audio transformations to their models.
The sliders are ready for immediate use and can be integrated into existing workflows with minimal effort. The training process is streamlined, making it accessible for both beginners and advanced users.
Overall, the release provides a practical solution for generating diverse audio concepts within Ace‑Step 1.5, enhancing creative possibilities.
Key changes
- Ace‑Step 1.5 concept sliders can be trained in 10‑60 minutes without specific datasets
- Collection includes 15 sliders such as male‑to‑female voice, studio‑to‑lofi, Bass boost, Choir‑to‑solo vocalist, digital‑to‑acoustic sound, Aggressive‑to‑gentle, drum intensity, energetic‑to‑calm, happiness, soft‑to‑projected voice, talking‑to‑singing, tempo, and danceability
- No dedicated concept datasets are required; training uses existing audio
- Sliders are ready for immediate use in Ace‑Step 1.5 pipelines