apizt 的 ACE 代理探测当前返回404，故本例按实际支持范围使用官方 ACE-Step 1.5 本地 Cover API：8秒原创程序旋律经4步CPU BF16推理得到48kHz WAV，耗时约308秒。输入是程序旋律，不能写成本人哼唱实测。
local-ace15-cover
remix.wav
source.wav
execution.json
{
"tool": "ACE-Step 1.5 official generate_music API",
"sourceCommit": "ca1e85fe9430179831e6bc6be790c332190a3866",
"modelRevision": "19671f406d603126926c1b7e2adc169acbcade22",
"runtimeAdaptations": [
"CPU DiT BF16 override to avoid FP32 memory expansion; model/algorithm unchanged",
"Validate only actual three core components for source-conditioned Cover; LM not downloaded or initialized",
"CPU two threads; offload flags enabled; no MLX/compile/quantization; own RSS stop at 9.25 GiB"
],
"params": {
"task_type": "cover",
"instruction": "Fill the audio semantic mask based on the given conditions:",
"reference_audio": null,
"src_audio": "original-eight-second-melody.wav",
"audio_codes": "",
"caption": "An original warm acoustic instrumental, plucked acoustic guitar carrying the clear C major motif, gentle bass and very light brushed percussion, no vocals, intimate dry studio sound.",
"global_caption": "",
"lyrics": "[Instrumental]",
"instrumental": true,
"vocal_language": "unknown",
"bpm": 60,
"keyscale": "C Major",
"timesignature": "4",
"duration": 8.0,
"enable_normalization": true,
"normalization_db": -1.0,
"fade_in_duration": 0.0,
"fade_out_duration": 0.0,
"latent_shift": 0.0,
"latent_rescale": 1.0,
"inference_steps": 4,
"seed": 20261024,
"guidance_scale": 1.0,
"use_adg": false,
"cfg_interval_start": 0.0,
"cfg_interval_end": 1.0,
"shift": 1.0,
"infer_method": "ode",
"sampler_mode": "euler",
"velocity_norm_threshold": 0.0,
"velocity_ema_factor": 0.0,
"dcw_enabled": false,
"dcw_mode": "double",
"dcw_scaler": 0.05,
"dcw_high_scaler": 0.02,
"dcw_wavelet": "haar",
"timesteps": null,
"repainting_start": 0.0,
"repainting_end": -1,
"chunk_mask_mode": "auto",
"repaint_latent_crossfade_frames": 10,
"repaint_wav_crossfade_sec": 0.0,
"repaint_mode": "balanced",
"repaint_strength": 0.5,
"retake_seed": null,
"retake_variance": 0.0,
"flow_edit_morph": false,
"flow_edit_source_caption": "",
"flow_edit_source_lyrics": "",
"flow_edit_n_min": 0.0,
"flow_edit_n_max": 1.0,
"flow_edit_n_avg": 1,
"audio_cover_strength": 0.6,
"cover_noise_strength": 0.0,
"thinking": false,
"lm_temperature": 0.85,
"lm_cfg_scale": 2.0,
"lm_top_k": 0,
"lm_top_p": 0.9,
"lm_negative_prompt": "NO USER INPUT",
"use_cot_metas": false,
"use_cot_caption": false,
"use_cot_lyrics": false,
"use_cot_language": false,
"use_constrained_decoding": true,
"cot_bpm": null,
"cot_keyscale": "",
"cot_timesignature": "",
"cot_duration": null,
"cot_vocal_language": "unknown",
"cot_caption": "",
"cot_lyrics": ""
},
"config": {
"batch_size": 1,
"allow_lm_batch": false,
"use_random_seed": false,
"seeds": [
20261024
],
"lm_batch_chunk_size": 8,
"constrained_decoding_debug": false,
"audio_format": "wav",
"mp3_bitrate": "128k",
"mp3_sample_rate": 48000
},
"sourceIdentity": "Program-synthesized original eight-second motif; not human humming",
"success": true,
"initialization": {
"success": true,
"status": "[OK] Model initialized successfully on cpu\nMain model: /Volumes/Mac Storage/Projects/tokenba-tutorial-next30-20261024/media/ace-checkpoints/acestep-v15-turbo\nVAE: /Volumes/Mac Storage/Projects/tokenba-tutorial-next30-20261024/media/ace-checkpoints/vae\nText encoder: /Volumes/Mac Storage/Projects/tokenba-tutorial-next30-20261024/media/ace-checkpoints/Qwen3-Embedding-0.6B\nDtype: torch.bfloat16\nAttention: sdpa\nCompiled: False\nQuantization: Disabled\nOffload to CPU: True\nOffload DiT to CPU: True\nMLX DiT: Disabled by user\nMLX VAE: Active (native MLX)",
"elapsedSeconds": 14.588632106781006
},
"statusMessage": "Generation completed successfully!",
"audios": [
{
"path": "9ee711de-6432-fa16-3f13-88b833c0f1d4.wav",
"key": "9ee711de-6432-fa16-3f13-88b833c0f1d4",
"sample_rate": 48000,
"params": {
"task_type": "cover",
"instruction": "Fill the audio semantic mask based on the given conditions:",
"reference_audio": null,
"src_audio": "/Volumes/Mac Storage/Projects/tokenba-tutorial-next30-20261024/media/inputs/original-eight-second-melody.wav",
"audio_codes": "",
"caption": "An original warm acoustic instrumental, plucked acoustic guitar carrying the clear C major motif, gentle bass and very light brushed percussion, no vocals, intimate dry studio sound.",
"global_caption": "",
"lyrics": "[Instrumental]",
"instrumental": true,
"vocal_language": "unknown",
"bpm": 60,
"keyscale": "C Major",
"timesignature": "4",
"duration": 8.0,
"enable_normalization": true,
"normalization_db": -1.0,
"fade_in_duration": 0.0,
"fade_out_duration": 0.0,
"latent_shift": 0.0,
"latent_rescale": 1.0,
"inference_steps": 4,
"seed": 20261024,
"guidance_scale": 1.0,
"use_adg": false,
"cfg_interval_start": 0.0,
"cfg_interval_end": 1.0,
"shift": 1.0,
"infer_method": "ode",
"sampler_mode": "euler",
"velocity_norm_threshold": 0.0,
"velocity_ema_factor": 0.0,
"dcw_enabled": false,
"dcw_mode": "double",
"dcw_scaler": 0.05,
"dcw_high_scaler": 0.02,
"dcw_wavelet": "haar",
"timesteps": null,
"repainting_start": 0.0,
"repainting_end": -1,
"chunk_mask_mode": "auto",
"repaint_latent_crossfade_frames": 10,
"repaint_wav_crossfade_sec": 0.0,
"repaint_mode": "balanced",
"repaint_strength": 0.5,
"retake_seed": null,
"retake_variance": 0.0,
"flow_edit_morph": false,
"flow_edit_source_caption": "",
"flow_edit_source_lyrics": "",
"flow_edit_n_min": 0.0,
"flow_edit_n_max": 1.0,
"flow_edit_n_avg": 1,
"audio_cover_strength": 0.6,
"cover_noise_strength": 0.0,
"thinking": false,
"lm_temperature": 0.85,
"lm_cfg_scale": 2.0,
"lm_top_k": 0,
"lm_top_p": 0.9,
"lm_negative_prompt": "NO USER INPUT",
"use_cot_metas": false,
"use_cot_caption": false,
"use_cot_lyrics": false,
"use_cot_language": false,
"use_constrained_decoding": true,
"cot_bpm": null,
"cot_keyscale": "",
"cot_timesignature": "",
"cot_duration": null,
"cot_vocal_language": "unknown",
"cot_caption": "",
"cot_lyrics": "",
"lora_loaded": false,
"use_lora": false,
"lora_scale": 1.0,
"lora_weights_hash": "",
"audio_format": "wav"
}
}
],
"elapsedSeconds": 308.2570161819458,
"maxRssBytes": 4739596288
}