ACE-Step Repaint

A ComfyUI node in Audio/ACE-Step with 14 inputs and 3 outputs.

By kana112233·Created 6 months ago·Updated 6 months ago· 27

ACE-Step Repaint

◄caption►

◄repainting_start0.00►

◄repainting_end-1.00►

◄checkpoint_dirAce-Step1.5►

◄config_pathacestep-v15-turbo►

◄lm_model_pathacestep-5Hz-lm-1.7B►

◄seed-1►

◄inference_steps8►

◄deviceauto►

◄thinkingtrue►

◄quantizationNone►

◄compile_modelfalse►

◄audio_formatflac►

CategoryAudio/ACE-Step

Inputs (14)

Name	Type	Default	Description
src_audio	AUDIO		The original audio signal to be repainted.
caption	STRING		Style description prompt for the repainted region.
repainting_start	FLOAT	0.000–600	Start time for the repainting region in seconds.
repainting_end	FLOAT	-1.00-1–600	End time for the repainting region in seconds. -1 means until the end of the audio.
checkpoint_dir	COMBO	Ace-Step1.5	Directory containing ACE-Step model weights (DiT model).
config_path	COMBO	acestep-v15-turbo	Specific model configuration to use (e.g., v1.5 turbo).
lm_model_path	COMBO	acestep-5Hz-lm-1.7B	Path to the language model used for processing metadata.
seed	INT	-1-1–18446744073709550000	Random seed for reproducibility. Set to -1 for random generation.
inference_steps	INT	81–64	Number of diffusion steps. Higher values (e.g., 25-50) improve quality but are slower.
device	COMBO	auto	Computing platform to run the model on.
thinkingopt	BOOLEAN	true	Whether to show the language model's Chain-of-Thought reasoning.
quantizationopt	COMBO	None	Model quantization (e.g., int8). Reduces VRAM usage but requires torchao and compile_model=True. Incompatible with LoRA.
compile_modelopt	BOOLEAN	false	Whether to use torch.compile to optimize the model. Required for quantization. Slow on first run but faster afterwards.
audio_formatopt	COMBO	flac	Output audio file format.