Nodes/ACE-Step-ComfyUI/ACE-Step Settings
ComfyUI Node

ACE-Step Settings

A ComfyUI node in api node/audio/ACE-Step with 11 inputs and 1 output.

By ace-step·Created 6 months ago·Updated 5 months ago· 71
ACE-Step Settings
    • settings
    seed-1
    thinkingtrue
    use_cot_captiontrue
    use_cot_languagetrue
    temperature0.85
    lm_cfg_scale2.0
    lm_top_p0.90
    lm_top_k0
    dit_guidance_scale7.0
    dit_inference_steps8
    dit_infer_methodode
    Categoryapi node/audio/ACE-Step

    Inputs (11)

    NameTypeDefaultDescription
    seedSTRING-1-1 = random
    thinkingBOOLEANtrueEnable 5Hz LM audio code generation (llm_dit mode)
    use_cot_captionBOOLEANtrueLLM rewrites/enhances caption via CoT
    use_cot_languageBOOLEANtrueLLM auto-detects vocal language
    temperatureFLOAT0.850–2LM sampling temperature
    lm_cfg_scaleFLOAT2.01–5LM classifier-free guidance scale
    lm_top_pFLOAT0.900–1LM nucleus sampling top-p
    lm_top_kINT00–200LM top-k sampling (0 = disabled)
    dit_guidance_scaleFLOAT7.00–20DiT classifier-free guidance scale
    dit_inference_stepsINT81–200DiT diffusion steps (turbo default = 8)
    dit_infer_methodCOMBOodeDiT ODE/SDE inference method

    Outputs (1)

    NameTypeDescription
    settingsACESTEP_SETTINGS