filliptm

filliptm / ComfyUI-FL-AceStep-Training

Public

FL AceStep Training - LoRA training nodes for ACE-Step 1.5 music generation in ComfyUI

81
8
100% credibility
Found Feb 19, 2026 at 43 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Custom nodes for ComfyUI enabling end-to-end LoRA training workflows for personalizing the ACE-Step music generation model directly in a visual node interface.

How It Works

1
🎨 Add music training tools

You drag special nodes into your visual canvas to teach AI your music style.

2
🧠 Load music model

Connect the base AI brain that understands music patterns.

3
📁 Scan your songs

Point to a folder of your favorite audio clips or recordings.

4
✨ Auto-describe clips

AI listens and writes fun descriptions, beats, and moods for each song.

5
🔧 Prepare for training

Chop clips into bite-sized pieces ready for the AI to learn from.

6
⚙️ Set training style

Pick how strong and long the learning session should be.

7
📊 Watch it learn live

See charts update in real-time as your custom style takes shape.

🎵 Your style is ready

Generate endless new music infused with your personal vibe!

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 43 to 81 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ComfyUI-FL-AceStep-Training?

This Python extension adds LoRA training nodes to ComfyUI for ACE-Step 1.5, the open-source music generation model from ace step github io. It lets you train custom LoRAs to inject your style, voice, or genre into music generation workflows, handling everything from scanning ace step training data to full end-to-end training right in the node graph. No need to switch tools—load ACE-Step models, preprocess long audio with tiled VAE encoding, and output ready-to-use LoRAs for inference.

Why is it gaining traction?

Unlike standalone ace-step training scripts, it runs natively in ComfyUI with real-time UI: live loss charts, progress bars, and stats via WebSocket. Auto-downloads models from Hugging Face, uses LLMs for dataset labeling (captions, BPM, key), and supports formats like WAV/MP3/FLAC. Developers love the seamless pipeline—scan directories, auto-label, preprocess to tensors, train—all without CLI hassle or VRAM crashes on long tracks.

Who should use this?

ComfyUI power users experimenting with music AI who want custom LoRAs for ACE-Step without Python scripting. Audio producers fine-tuning on personal datasets for generation of specific genres or voices. Teams building music gen pipelines needing ace step training data prep and LoRA export for production.

Verdict

Solid early pick for ComfyUI music workflows—excellent docs and MIT license make it easy to adopt despite 43 stars and 1.0% credibility score. Still beta (test coverage light), so test on small datasets first, but it delivers on ace-step LoRA training promise.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.