mlx-Yue · 完整歌曲音频对比与性能展台

全套官方输入,端到端真实生成。支持 32步高保真标准模式 与 8步极速推理模式 (⚡ 实时/超实时生成) 同台对比试听。

🍎 硬件环境: Apple M3 Max (128 GiB 统一内存)
⚡ 量化策略: AR 8-bit 量化权重 + NAR BF16 + VAE FP32
🎼 覆盖场景: 6 大多国语种 + 5 大全流程生成模式 + 翻唱与重排
⏱️ 8步加速比: NAR 声学求解提速 3.7x ~ 7.5x,平均 RTF 降至 0.89 (超越实时生成速度)

15 / 15 项完整歌曲已全部就绪 · 32步标准模式 与 8步快速模式 双版本对比就绪

📊 15个完整案例性能对比汇总表 (32步 vs 8步)
案例 (Case ID / 流派) 官方时长 32步耗时 32步 RTF 8步耗时 8步 RTF (⚡) NAR 提速比 综合提速 状态
ja_score (Easy Listening)221.8s553.6s2.490 (0.40x)145.9s0.656 (1.52x)7.56x3.79x✓ 完成
ru_score (Folktronica)120.0s195.1s1.626 (0.61x)84.4s0.703 (1.42x)5.20x2.31x✓ 完成
es_score (Flamenco)182.9s389.7s2.153 (0.46x)174.8s0.965 (1.04x)4.33x2.23x✓ 完成
zh_score (R&B)152.7s258.0s1.695 (0.59x)228.8s1.503 (0.67x)1.22x1.13x✓ 完成
ko_score (Emo)156.9s299.8s1.921 (0.52x)142.3s0.912 (1.10x)3.42x2.11x✓ 完成
en_score (Rock & Roll)95.8s146.9s1.547 (0.65x)70.9s0.747 (1.34x)3.94x2.07x✓ 完成
full_generated (Comedy)204.9s504.2s2.402 (0.42x)216.5s1.032 (0.97x)4.74x2.33x✓ 完成
melody_generated (R&B)152.7s417.9s2.309 (0.43x)144.3s0.797 (1.25x)6.26x2.90x✓ 完成
melody_supplied (Christmas Pop · E Minor · 104 BPM)107.4s154.7s1.446 (0.69x)72.7s0.680 (1.47x)4.44x2.13x✓ 完成
off_direct (Enka)243.3s562.7s2.241 (0.45x)282.7s1.126 (0.89x)4.44x1.99x✓ 完成
cover_dagoujiao (New Orleans Brass-Funk · 115 BPM)210.0s390.5s1.878 (0.53x)175.8s0.846 (1.18x)3.80x2.22x✓ 完成
cover_baikal (Jazz Ballad · 60 BPM)216.9s372.7s1.756 (0.57x)169.4s0.798 (1.25x)3.67x2.20x✓ 完成
cover_birthday (Heavy Metal · 160 BPM)63.9s102.2s1.576 (0.63x)46.8s0.722 (1.39x)4.77x2.18x✓ 完成
audio_to_cover (Easy Listening)221.8s442.2s1.971 (0.51x)201.9s0.900 (1.11x)3.93x2.19x✓ 完成
jazz_funk_score (Jazz-Funk)130.0s201.3s1.543 (0.65x)94.2s0.722 (1.39x)4.13x2.14x✓ 完成

🎧 逐曲试听与三维对比 (官方原版 vs 32步标准 vs 8步快速)

日语 · FULL · 官方完整乐谱 8步快速模式提速 3.79x (RTF 0.656)

Easy Listening

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:1b91eef30046ca2f0cf5633e

官方原版 Demo

Official Reference

时长: 221.8s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 222.3s | 耗时: 553.6s | RTF: 2.490 (0.40x 实时)
NAR声学求解: 464.6s (32步) | VAE: 10.4s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 222.3s | 耗时: 145.9s | RTF: 0.656 (1.52x 实时)
NAR声学求解: 61.4s (8步 · 7.56x 提速) | VAE: 7.0s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

A gentle and melancholic piece beginning with a delicate piano melody and a breathy, expressive flute line. A soft, clear female vocal enters, singing in a gentle, almost lullaby-like style. The arrangement gradually builds, introducing a subtle, downtempo lo-fi hip-hop beat with a soft kick and snare, complemented by a warm bassline and atmospheric synth pads. The track maintains a tranquil and introspective mood throughout, with the piano and flute weaving in and out, creating a serene soundscape perfect for relaxation or contemplation. The song concludes with a beautiful instrumental outro featuring the piano and flute, fading into silence.

歌词 (Lyrics)

[Intro: Piano & Flute]

[Verse 1]
静かな森 歩み寄る
木々のざわめき 聞こえる音
陽射しが揺れて 影踊る
心は軽く 自由になる

[Verse 2]
風が通り 抱きしめるように
葉は止ま 語る古い物語
小鳥の歌が 空を仰ぎ
その音色だけ 生きている

[Chorus]
森の中で 消えてゆく
小さな悩み 風に乗る
心地良い緑 包まれて
私は癒され 生まれ変わる

[Instrumental Break: Piano, Flute, Drums]

[Verse 3]
土の匂いが 深く染みる
地球の鼓動 感じている
目を閉じ夜明けを 夢見て
自然と溶ける 虚無する

[Verse 4]
月の光が ゆっくりと
静けさ抱え 手ら知ってる
心がささやく 小さな声
森のリズムで 調和化する

[Chorus]
森の中で 消えてゆく
小さな悩み 風に乗る
心地良い緑 包まれて
私は癒され 生まれ変わる

[Outro: Piano & Flute]

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/16
Q:1/4=85
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:D#m
% intro
V: Vocal
z8"Bmaj7"z8|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
z12z3a|a3c'3g6f2a2|g3c'3e8z2|e3f3c8Ac|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
e3f3c4f2e2c2|a3c'3g6f2a2|g3c'3e4f2e2c2|e3f3c8Ac|
V: Vocal
"D#m7"z8z2d2f2g2|
V: Ins
e3f3c4z6|
% verse
V: Vocal
"Bmaj7"a2dd3z4d2f2a2|"C#"g3c3z4c2d2e2|"D#m7"e2fe2c2cc2AA3GA-|"D#m7"A2z8d2f2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3d3cd3d2f2a2|"C#"g3c3z4c2d2e2|"D#m7"e2fe2c2cc2AA3GA-|"D#m7"A2z8d2f2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3d3cd3d2f2a2|"C#"g3c3z4c2d2e2|"D#m7"e2fe2c2cc2A2A2G2|"D#m7"G3c3A2z2d2f2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3d3cd3d2f2a2|"C#"g3c3e4z4c2|"D#m7"d3e3f2g3f3e2|"D#m7"e3f3e4d2d4|
V: Ins
Z4|
V: Vocal
"D#m7"z8z2d2a2g2|
V: Ins
Z|
% chorus
V: Vocal
"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"e3c4z3d2a2g2|"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"g3a4z3d2a2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"g3c'3a4z4de|"Bmaj7"f2f2f2a2"C#"g3f3e2|"D#m7"e3f3e4d2d4|
V: Ins
Z4|
% interlude
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8-A2-|A6z8z2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z8z2d2f2g2|
V: Ins
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8-A2-|A6z8z2|
% verse
V: Vocal
"Bmaj7"a2dd2c3z2d2f2a2|"C#"g3c3e2z2c2d2e2|"D#m7"e2fe2c3c2AA3GA-|"D#m7"A2z8d2f2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3d3cd3d2f2a2|"C#"g3c4z4zd2e2|"D#m7"e3f3e4d4c2|"D#m7"e3f3e4d4z2|
V: Ins
Z4|
V: Vocal
"Bmaj7"z2aaa2gg2ff2z4|"C#"z2a2a2gg2c'4z3|"D#m7"z2aaa2gg2ff2z2e2|"D#m7"e3f3e4d4z2|
V: Ins
Z4|
V: Vocal
"Bmaj7"z2aaa2gggff4z2|"C#"z2a2a2gg2c'c'z4z|"D#m7"z2aaa2gggff4e2|"D#m7"e3f3e4d2d4|
V: Ins
Z4|
V: Vocal
"D#m"z8z2d2a2g2|
V: Ins
Z|
% chorus
V: Vocal
"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"g3c4z3d2a2g2|"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"g3a4z3d2a2g2|
V: Ins
Z4|
V: Vocal
"Bmaj7"a3c'3g2-"C#"g2a2g2f2|"D#m7"g3c'3a4z4de|"Bmaj7"f2f2f2a2"C#"g3f3e2|"D#m7"e3f3e4d2d4|
V: Ins
Z4|
% outro
V: Vocal
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|
V: Ins
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8-A2-|A6z8z2|
V: Vocal
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|
V: Ins
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8-A2-|A6z8z2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a3c'3g6f2a2|g3c'3e8z2|e3f3c8Ac|e3f3c8z2|
V: Vocal
"D#m7"z16|
V: Ins
Z|

32步标准模式详细指标

{
  "id": "ja_score",
  "status": "complete",
  "case_sha256": "0f832298a61a1fe7782e3ae11633247a1f6184f5fb7305a1897ebdc7dd0529cc",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 222.27866666666668,
  "official_seconds": 221.76,
  "duration_ratio": 1.002338864838865,
  "e2e_seconds": 553.5552974999882,
  "rtf": 2.4903662857133755,
  "speed_multiplier": 0.40154735700397015,
  "load_seconds": 5.217120249988511,
  "transcription_seconds": 0,
  "save_seconds": 0.5711062909686007,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1738
    },
    "semantic": {
      "seconds": 72.30839145800564,
      "prefill_seconds": 0.7439379160059616,
      "ttft_seconds": 0.7479968330007978,
      "output_tokens": 5558,
      "content_tokens": 5557,
      "output_tps": 76.86521422935961,
      "prefix_tokens": 2200,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 464.59326087497175,
    "vae_seconds": 10.387618332984857,
    "load": {
      "resolve_conversion_integrity_seconds": 5.216982375015505,
      "ar_load_events_seconds": [
        0.1254808329977095
      ],
      "ar_load_seconds": 0.1254808329977095,
      "conditioning_load_events_seconds": [
        0.18663158296840265
      ],
      "conditioning_load_seconds": 0.18663158296840265,
      "nar_load_events_seconds": [
        0.1431943330098875
      ],
      "nar_load_seconds": 0.1431943330098875,
      "vae_load_events_seconds": [
        0.10199695901246741
      ],
      "vae_load_seconds": 0.10199695901246741
    },
    "e2e_seconds": 547.4561339580105
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "d749941555cc8c222323475543b07d022a798a4ccfe5b31e3cfb7cb34804a51a",
  "audio_sha256": "2aa7e9afce7f5afd660bc39817be3f479a6b59773ba67caaeb527360f561c2e2",
  "result_sha256": "af10208e255a1f2230108d1ec7e683214f31481f7678ea38f9222ca64ba05ed8",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14319429003155604,
    "peak": 0.8744818568229675,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01626429200405255,
      "rss_bytes": 51658752,
      "system_swap_used_bytes": 12407668736,
      "system_swap_in_bytes": 56318050304,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 52476608512,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 11686707200,
      "system_wired_bytes": 6419333120,
      "system_compressor_bytes": 36918575104,
      "system_file_backed_bytes": 36826939392,
      "physical_footprint_bytes": 31785680,
      "process_lifetime_peak_footprint_bytes": 31785680,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 554.2063482080121,
      "rss_bytes": 830947328,
      "system_swap_used_bytes": 12390891520,
      "system_swap_in_bytes": 56321720320,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63129010176,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2445590528,
      "system_wired_bytes": 8817328128,
      "system_compressor_bytes": 2902540288,
      "system_file_backed_bytes": 35045130240,
      "physical_footprint_bytes": 988678304,
      "process_lifetime_peak_footprint_bytes": 11875867272,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 196902912,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11823946376,
      "process_lifetime_peak_footprint_bytes": 11875867272,
      "rss_bytes": 8190558208,
      "system_available_bytes": 64890142720,
      "system_compressor_bytes": 36918624256,
      "system_file_backed_bytes": 37440602112,
      "system_free_bytes": 12019777536,
      "system_swap_in_bytes": 56321720320,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12407668736,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 43725864960
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/ja_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "ja_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 222.28,
  "e2e_seconds": 145.92,
  "rtf": 0.656,
  "speed_multiplier": 1.52,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1738
    },
    "semantic": {
      "seconds": 72.30839145800564,
      "prefill_seconds": 0.7439379160059616,
      "ttft_seconds": 0.7479968330007978,
      "output_tokens": 5558,
      "content_tokens": 5557,
      "output_tps": 76.86521422935961,
      "prefix_tokens": 2200,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 61.42,
    "vae_seconds": 6.98,
    "load_seconds": 5.22,
    "transcription_seconds": 0,
    "e2e_seconds": 145.92,
    "32step_nar_seconds": 464.59,
    "nar_speedup": 7.56
  },
  "audio_sha256": "7b11cabcca8c531494b6d4f0e9ab4b5113dc0533b88e67d7d4531fea2bcff410",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1432256503016256,
    "peak": 0.8848203420639038,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01121237501502037,
      "rss_bytes": 143228928,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56344313856,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 70419841024,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 13504757760,
      "system_wired_bytes": 6416859136,
      "system_compressor_bytes": 2731704320,
      "system_file_backed_bytes": 29625597952,
      "physical_footprint_bytes": 306758568,
      "process_lifetime_peak_footprint_bytes": 309363624,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 395280,
      "mlx_peak_bytes": 395280,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 68.41115374997025,
      "rss_bytes": 8173420544,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62546362368,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 653656064,
      "system_wired_bytes": 8482144256,
      "system_compressor_bytes": 2731261952,
      "system_file_backed_bytes": 30631641088,
      "physical_footprint_bytes": 8970361648,
      "process_lifetime_peak_footprint_bytes": 10961378072,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 145244432,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 164911136,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 10808679192,
      "process_lifetime_peak_footprint_bytes": 10961378072,
      "rss_bytes": 8173420544,
      "system_available_bytes": 70419841024,
      "system_compressor_bytes": 2732081152,
      "system_file_backed_bytes": 33056686080,
      "system_free_bytes": 13504757760,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 9897803776
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
俄语 · FULL · 官方完整乐谱 8步快速模式提速 2.31x (RTF 0.703)

Folktronica

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:43fc3aa38700252b8d63982b

官方原版 Demo

Official Reference

时长: 120.0s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 120.0s | 耗时: 195.1s | RTF: 1.626 (0.61x 实时)
NAR声学求解: 133.8s (32步) | VAE: 5.9s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 120.0s | 耗时: 84.4s | RTF: 0.703 (1.42x 实时)
NAR声学求解: 25.7s (8步 · 5.2x 提速) | VAE: 4.0s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Dance Pop, Storytelling, Groovy, Anthemic, Dance, Folk, Melodic, Danceable, Electronic, female vocals

歌词 (Lyrics)

Ах вы, сени мои, сени
Сени новые мои
Сени новые, кленовые
Решетчатые!
Выходила молода
За новые ворота
Выпускала сокола
Из правого рукава
На полетике соколику наказывала
«Ты лети, лети, соколик
высоко и далеко».
И высоко и далеко
На родиму сторону
На родимой на сторонке
Грозен батюшка живёт.
Он и грозен, сударь, грозен
Да не милостив
Не пускает молоду
Поздно вечером одну.
Не велит поздно ходить
С молодцами говорить.
Не послушаю отца
Эх, потешу молодца.
Ах вы, сени мои, сени
Сени новые мои
Сени новые, кленовые
Решетчатые!

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/16
Q:1/4=126
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:Am
% intro
V: Vocal
z12"Am"z4|"Am"e2z2e2z2e2f2e2d2|"Dm"d2z2d2z2d2e2d2c2|"F"c2z2c2z2c2d2c2B2|
V: Ins
Z4|
V: Vocal
"G"B4z12|"Am"e2z2e2z2e2f2e2d2|"Dm"d2z2d2z2d2e2d2c2|"F"c4z12|
V: Ins
Z3|z4c4c4c4|
V: Vocal
"G"z16|
V: Ins
c4c4c4c4|
% verse
V: Vocal
"Am"e4A4e2e2e2d2|"Dm"d4A4z8|"F"c4A4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d2d2d2c2d2e2d2c2|"F"c4z12|"G"z16|
V: Ins
Z2|z6c'2z2b2c'2b2|c'3b3c'2z2b2c'2g2|
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d4z12|"F"c4A4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d4z8z2A2|"F"c4A2A4d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"a4e4a2a2e4|"Dm"f2f2f2f2f2e2d2c2|"F"c2c2c2c2c2d2c2B2|"G"B2B2B2B2B2B2A4|
V: Ins
Z4|
V: Vocal
"Am"z16|"Am"z16|
V: Ins
A4A4A4A4|A2A2A2A2AAAAAAAA|
% chorus
V: Vocal
"Am"e4e4e2f2e2d2|"Dm"d4d4d2e2d2c2|"F"c4c4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"e4e4e2f2e2d2|"Dm"d4d4d2e2d2c2|"F"c4c4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
% verse
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d4z12|"F"c4A4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d4z12|"F"c4A4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
% interlude
V: Vocal
"Am"a4e4a2a2e2e2|"Dm"f4z12|"F"a4e4a2a2e2e2|"G"f4z12|
V: Ins
Z4|
V: Vocal
"Am"a4e4a2a2e2e2|"Dm"f4z12|"F"c4z2c2c2d2c2B2|"G"B4z12|
V: Ins
Z4|
% chorus
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d4A4z8|"F"c4A4c2d2c2B2|"G"B4z12|
V: Ins
Z4|
V: Vocal
"Am"e4A4e2f2e2d2|"Dm"d2d2d2c2d2e2d2c2|"F"c4z12|"G"z16|
V: Ins
Z2|z6c'2z2b2c'2b2|c'3b3c'4b2c'2b2|
V: Vocal
Z|
V: Ins
c'4z12|

32步标准模式详细指标

{
  "id": "ru_score",
  "status": "complete",
  "case_sha256": "b8e8dc58ee5f6c88305f86733815138bbb81f039aabb33d1ba7c7f5556b4cad9",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 119.95866666666667,
  "official_seconds": 119.96,
  "duration_ratio": 0.9999888851839503,
  "e2e_seconds": 195.10311070800526,
  "rtf": 1.6264194670498053,
  "speed_multiplier": 0.6148475348822035,
  "load_seconds": 5.187667417048942,
  "transcription_seconds": 0,
  "save_seconds": 0.468616625003051,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1158
    },
    "semantic": {
      "seconds": 49.4814189170138,
      "prefill_seconds": 0.9020838750293478,
      "ttft_seconds": 0.907485292002093,
      "output_tokens": 3000,
      "content_tokens": 2999,
      "output_tps": 60.62881917415011,
      "prefix_tokens": 1450,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 133.82474104198627,
    "vae_seconds": 5.876439166953787,
    "load": {
      "resolve_conversion_integrity_seconds": 5.187611332978122,
      "ar_load_events_seconds": [
        0.11997312499443069
      ],
      "ar_load_seconds": 0.11997312499443069,
      "conditioning_load_events_seconds": [
        0.2336275419802405
      ],
      "conditioning_load_seconds": 0.2336275419802405,
      "nar_load_events_seconds": [
        0.11692583403782919
      ],
      "nar_load_seconds": 0.11692583403782919,
      "vae_load_events_seconds": [
        0.10771866701543331
      ],
      "vae_load_seconds": 0.10771866701543331
    },
    "e2e_seconds": 189.32758395798737
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "be5e899886c562e66d297bf0700174cb3dac682b4ca3d2ba523c89384c3cd5bb",
  "audio_sha256": "a4235afa6365faeeb9a4a4cd8e52b9abd196d71e8e114087cf703e45b4fac42a",
  "result_sha256": "84b536bb5fe4e872adba3455ae38949b12fe8dc0145653938d7e1a0be7d6bd25",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.130880053446391,
    "peak": 0.822986364364624,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.00960112496977672,
      "rss_bytes": 51675136,
      "system_swap_used_bytes": 12390891520,
      "system_swap_in_bytes": 56321720320,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 69813141504,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 12734529536,
      "system_wired_bytes": 6374260736,
      "system_compressor_bytes": 2902540288,
      "system_file_backed_bytes": 35071262720,
      "physical_footprint_bytes": 31802064,
      "process_lifetime_peak_footprint_bytes": 31802064,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 195.62312208296498,
      "rss_bytes": 942751744,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56328011776,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 68005363712,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 10946576384,
      "system_wired_bytes": 8721760256,
      "system_compressor_bytes": 2908192768,
      "system_file_backed_bytes": 33731592192,
      "physical_footprint_bytes": 1247185224,
      "process_lifetime_peak_footprint_bytes": 11869936240,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 156781600,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11774040688,
      "process_lifetime_peak_footprint_bytes": 11869936240,
      "rss_bytes": 8236990464,
      "system_available_bytes": 69813141504,
      "system_compressor_bytes": 2908487680,
      "system_file_backed_bytes": 36293132288,
      "system_free_bytes": 12734529536,
      "system_swap_in_bytes": 56328011776,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12390891520,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 12710789120
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/ru_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "ru_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 119.96,
  "e2e_seconds": 84.37,
  "rtf": 0.703,
  "speed_multiplier": 1.42,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1158
    },
    "semantic": {
      "seconds": 49.4814189170138,
      "prefill_seconds": 0.9020838750293478,
      "ttft_seconds": 0.907485292002093,
      "output_tokens": 3000,
      "content_tokens": 2999,
      "output_tps": 60.62881917415011,
      "prefix_tokens": 1450,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 25.73,
    "vae_seconds": 3.97,
    "load_seconds": 5.19,
    "transcription_seconds": 0,
    "e2e_seconds": 84.37,
    "32step_nar_seconds": 133.82,
    "nar_speedup": 5.2
  },
  "audio_sha256": "ceb5ccf8de6103444abb14ec05eb353b68aaac222a343204f3e295f916e9e58e",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.13106356362949043,
    "peak": 0.8749023675918579,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.015435791981872171,
      "rss_bytes": 8347926528,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62546362368,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 653656064,
      "system_wired_bytes": 8482144256,
      "system_compressor_bytes": 2731261952,
      "system_file_backed_bytes": 30631641088,
      "physical_footprint_bytes": 8657345448,
      "process_lifetime_peak_footprint_bytes": 10961378072,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 145244432,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 29.719285999948625,
      "rss_bytes": 8443052032,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63881330688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2137161728,
      "system_wired_bytes": 8152301568,
      "system_compressor_bytes": 2731229184,
      "system_file_backed_bytes": 30657904640,
      "physical_footprint_bytes": 9246104560,
      "process_lifetime_peak_footprint_bytes": 11033320360,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 146718992,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11256150230,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11033271208,
      "process_lifetime_peak_footprint_bytes": 11033320360,
      "rss_bytes": 8443052032,
      "system_available_bytes": 63909478400,
      "system_compressor_bytes": 2731261952,
      "system_file_backed_bytes": 30675107840,
      "system_free_bytes": 2223013888,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11200823296
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
西班牙语 · FULL · 官方完整乐谱 8步快速模式提速 2.23x (RTF 0.965)

Flamenco

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:0792ba06ebe2fd7cdd6c3e7b

官方原版 Demo

Official Reference

时长: 182.9s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 181.0s | 耗时: 389.7s | RTF: 2.153 (0.46x 实时)
NAR声学求解: 283.8s (32步) | VAE: 8.5s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 181.0s | 耗时: 174.8s | RTF: 0.965 (1.04x 实时)
NAR声学求解: 65.5s (8步 · 4.33x 提速) | VAE: 12.6s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

flamenco gipsy female vocals, deep male spanish rap voice, hip hop beat with cajón and bass, flamenco guitar, palmas, 95 bpm, joyful energy, andalusian accent, fusion, bright tone, no melancholy, festive rhythm, call and response vocals, uplifting and emotional, warm analog texture, street spirit, modern flamenco rap fusion, mediterranean vibe, vibrant and organic production.

歌词 (Lyrics)

[Intro – female flamenco vocals, slow compás, soft palmas, flamenco guitar intro]
Ay, Jaén, tierra de fuego
donde el sol besa el suelo
entre olivos se levanta
tu corazón verdadero.



[Beat drops – male rap enters, groovy hip-hop rhythm, cajón + bass, 95 BPM]
Jaén despierta, con fuerza y con arte
la gente sonríe, la vida reparte
en cada calle hay historia que late
orgullo en el alma, raíces que arden.

Subo la cuesta, la brisa me guía
Santa Catalina me da compañía
el barrio murmura con algarabía
la tapa, la risa, la buena energía.



[Chorus – female flamenco vocals, upbeat rhythm with hand claps and guitar strums]
Ay, Jaén, qué bonito suena
tu cante vuela, tu luz me llena
del valle al monte, la voz resuena
Jaén se alegra, Jaén resuena.



[Verse 2 – male rap with female ad-libs, syncopated rhythm]
Tierra sincera, mirada clara
trabajo y coraje, palabra rara
desde el casco viejo hasta la Almenara
Jaén no espera, Jaén dispara.

[Female ad-libs – flamenco “jaleos” style]
¡Ole mi gente buena! ¡Ay qué alegría más sana!

[Male rap continues]
Por el Paseo de la Estación suena el alma
ni el tiempo ni el miedo aquí tienen calma
Jaén respira verdad y palma
cada esquina suena a jarana.



[Bridge – vocal duel, flamenco and rap call & response, energetic palmas]
Tierra bendita, no hay quien te iguale
Aquí el corazón se mide en el vale.
Deja que el cante reviente la calle
Que en cada compás la vida estalle.



[Final Chorus – both voices together, full festive energy, strong compás]
Ay, Jaén, qué bonito suena
tu cante vuela, tu luz me llena
del valle al monte, la voz resuena
Jaén se alegra, Jaén resuena.



[Outro – female flamenco vocals, slow fade with guitar and cajón]
Ay, mi Jaén, mi tierra buena
mi gente canta, la vida suena
Jaén resuena, Jaén resuena...

输入乐谱 (ABC Notation)

X:1
T:
M:2/4
L:1/32
Q:1/4=95
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:C#m
% intro
V: Vocal
Z|z8"C#m"z8|"C#m"z16|"C#m"z16|
V: Ins
Z2|c2e2d2e8-e2-|e6z4d2e2f2|
V: Vocal
"C#m"z16|"C#m"z16|"C#m"z16|"C#m"z16|
V: Ins
g2a2g12-|g4z4f2g2a2b2|c'4c'4c'4b4|a4a2g2g4f4|
V: Vocal
"C#m"z16|"C#m"z16|
V: Ins
e4z12|Z|
% verse
V: Vocal
"C#m"z8z2e4e2|"C#m"e4e2d4c2c4-|"A"c4z8z2B2|"A"e4d2d4c2c4-|
V: Ins
Z4|
V: Vocal
"B"c4B2z6B2B2|"B"B4c2d4e2e4-|"C#m"e2c2c2z8B2|"C#m"B4c2d4d4c2|
V: Ins
Z4|
V: Vocal
"C#m"c8z8|"C#m"z8z2C2E2D2|"C#m"E4C2C2E4E2D2|"C#m"E4C2C2E4E2D2|
V: Ins
Z4|
V: Vocal
"A"E4C2C2E4E2D2|"A"E4C2z4C2E2D2|"B"D4D2C2D4D2C2|"B"D4B,2B,2D4D2C2|
V: Ins
Z4|
V: Vocal
"C#m"D4B,2B,2D4D2C2|"C#m"D4C4z2c2e2d2|"C#m"e4c2c2e4e2d2|"C#m"e4c4e2e2e2d2|
V: Ins
Z4|
V: Vocal
"A"e4c2c2e4e2d2|"A"e4c4e4e2d2|"B"d4d2c2d4d2c2|"B"d4B2B2d4d2c2|
V: Ins
Z4|
V: Vocal
"C#m"d4B2B2d4d2c2|"C#m"d4c8z4|
V: Ins
Z2|
% chorus
V: Vocal
"C#m"g8e8|"C#m"c4c2c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2a4g2g2-|"C#m"g2e4z8z2|"C#m"z6e2e4e2g2-|
V: Ins
Z4|
V: Vocal
"C#m"g2g4f2e8|"C#m"z6c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2e4d2e2-|"C#m"e2c4z8z2|"C#m"z8z2C2E2D2|
V: Ins
Z4|
% verse
V: Vocal
"C#m"E4C2C2E2D2E4|"C#m"C4z4E2E2E2D2|"A"E4C2C2E2D2E4|"A"C4z4E2E2E2D2|
V: Ins
Z4|
V: Vocal
"B"D4D2C2D2C2D4|"B"B,4z6C2D2C2|"C#m"D4B,4z2C2D2C2|"C#m"D4B,4z8|
V: Ins
Z4|
V: Vocal
"C#m"z8z2c2c2c2|"C#m"g4f2e4d2c4|"A"z8z2c2c2c2|"A"g2f4e4d4e2|
V: Ins
Z4|
V: Vocal
"B"d16-|"B"d8z2e2c2d2|"C#m"e16-|"C#m"e8z2C2E2D2|
V: Ins
Z4|
V: Vocal
"C#m"E2E2E2D2E4E2D2|"C#m"E4C4E2E2E2D2|"A"E4C2E4E4D2|"A"E4C4z6C2|
V: Ins
Z4|
V: Vocal
"B"D4D2D4D4C2|"B"D4C2D4C4z2|D2C2D2C2D4D2C2|D4C4z8|
V: Ins
Z4|
% pre-chorus
V: Vocal
"C#m"z2c2c2c2g4g2g2-|"C#m"g2f4e2e4d4|"A"c2c2c2c2g4g2g2-|"A"g2f4e2e4d4|
V: Ins
Z4|
V: Vocal
"B"z2d2d2d2f4f2f2|"B"f4e2d2e4d2c2|"C#m"c4c2c2e6d2|"C#m"e4d4e4d2c2|
V: Ins
Z4|
% chorus
V: Vocal
"C#m"g8e8|"C#m"c4c2c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2a4g2g2-|"C#m"g2e4z8z2|"C#m"z6e2e4e2g2-|
V: Ins
Z4|
V: Vocal
"C#m"g2g4f2e8|"C#m"z6c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2e4d2e2-|"C#m"e2c4z8z2|"C#m"z16|
V: Ins
Z4|
V: Vocal
"C#m"g8e8|"C#m"c4c2c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2a4g2g2-|"C#m"g2e4z8z2|"C#m"z6e2e4e2g2-|
V: Ins
Z4|
V: Vocal
"C#m"g2g4f2e8|"C#m"z6c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f2d4z8z2|"B"z6d2e4d2e2-|"C#m"e2c4z8z2|"C#m"z16|
V: Ins
Z4|
% outro
V: Vocal
"C#m"g16-|"C#m"g8z2g2f2g2|"A"a16-|"A"a8z2g2f2g2|
V: Ins
Z4|
V: Vocal
"B"f16-|"B"f8e4d2e2|"C#m"c16-|"C#m"c8z8|
V: Ins
Z4|
V: Vocal
"C#m"g16-|"C#m"g8z2a2f2g2|"A"a16-|"A"a8z2g2f2g2|
V: Ins
Z4|
V: Vocal
"B"f16-|"B"f8e4d2e2|"C#m"c16-|"C#m"c8z8|
V: Ins
Z4|
% chorus
V: Vocal
"C#m"g8e8|"C#m"c6c2a4g2a2-|"A"a2g4z8z2|"A"z6e2f4e2f2-|
V: Ins
Z4|
V: Vocal
"B"f4d12-|"B"d4z2d2e4d2e2-|"C#m"e2c4z8z2|"C#m"z6d2e4d2e2-|
V: Ins
Z4|
V: Vocal
"C#m"e2c4z8z2|"C#m"z6d2e4d2e2-|"C#m"e2c12-c2-|"C#m"c8-c2z6|
V: Ins
Z4|
V: Vocal
Z|
V: Ins
Z|

32步标准模式详细指标

{
  "id": "es_score",
  "status": "complete",
  "case_sha256": "d559816619d07e4fbaf1dbcdedb3669bcc08cd4fe588e559b7184c9e30813ce5",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 181.03866666666667,
  "official_seconds": 182.92,
  "duration_ratio": 0.9897149938042132,
  "e2e_seconds": 389.6965463749948,
  "rtf": 2.1525597462144077,
  "speed_multiplier": 0.4645631796091359,
  "load_seconds": 5.2215331249753945,
  "transcription_seconds": 0,
  "save_seconds": 0.48262100003194064,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 2366
    },
    "semantic": {
      "seconds": 91.5086536660092,
      "prefill_seconds": 1.83645666600205,
      "ttft_seconds": 1.8413168750121258,
      "output_tokens": 4527,
      "content_tokens": 4526,
      "output_tps": 49.470731112740104,
      "prefix_tokens": 3042,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 283.7967567089945,
    "vae_seconds": 8.454269333044067,
    "load": {
      "resolve_conversion_integrity_seconds": 5.221474041987676,
      "ar_load_events_seconds": [
        0.11692950001452118
      ],
      "ar_load_seconds": 0.11692950001452118,
      "conditioning_load_events_seconds": [
        0.2450902090058662
      ],
      "conditioning_load_seconds": 0.2450902090058662,
      "nar_load_events_seconds": [
        0.1404322080197744
      ],
      "nar_load_seconds": 0.1404322080197744,
      "vae_load_events_seconds": [
        0.10621683299541473
      ],
      "vae_load_seconds": 0.10621683299541473
    },
    "e2e_seconds": 383.908113958023
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "533c88d131634c01b3bf46d6bba3649530811bbd77171619d9215e65d9f5981c",
  "audio_sha256": "84cad5eafcd34a09a9afce4efbe13a0eb624edddf86204a7290cdfc162e14924",
  "result_sha256": "adfb789aa7b14c85aab0a5ab0237277584c147cf7034690e4db83c109f7bc839",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14357905449661487,
    "peak": 0.9590508341789246,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012591834005434066,
      "rss_bytes": 52068352,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56328011776,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 68005363712,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 10946576384,
      "system_wired_bytes": 8721760256,
      "system_compressor_bytes": 2908192768,
      "system_file_backed_bytes": 33731592192,
      "physical_footprint_bytes": 32211688,
      "process_lifetime_peak_footprint_bytes": 32211688,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 390.24551800003974,
      "rss_bytes": 1019641856,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56328994816,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65221279744,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 4218208256,
      "system_wired_bytes": 7308443648,
      "system_compressor_bytes": 2906521600,
      "system_file_backed_bytes": 34812772352,
      "physical_footprint_bytes": 1317636472,
      "process_lifetime_peak_footprint_bytes": 12049209992,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 185466880,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11901475464,
      "process_lifetime_peak_footprint_bytes": 12049209992,
      "rss_bytes": 8175878144,
      "system_available_bytes": 68721623040,
      "system_compressor_bytes": 2908258304,
      "system_file_backed_bytes": 34836414464,
      "system_free_bytes": 11903664128,
      "system_swap_in_bytes": 56328994816,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 13419298816
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/es_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "es_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 181.04,
  "e2e_seconds": 174.78,
  "rtf": 0.965,
  "speed_multiplier": 1.04,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 2366
    },
    "semantic": {
      "seconds": 91.5086536660092,
      "prefill_seconds": 1.83645666600205,
      "ttft_seconds": 1.8413168750121258,
      "output_tokens": 4527,
      "content_tokens": 4526,
      "output_tps": 49.470731112740104,
      "prefix_tokens": 3042,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 65.49,
    "vae_seconds": 12.56,
    "load_seconds": 5.22,
    "transcription_seconds": 0,
    "e2e_seconds": 174.78,
    "32step_nar_seconds": 283.8,
    "nar_speedup": 4.33
  },
  "audio_sha256": "9c97a71b3ecaedd3fee200fd169de75de3bd1c9038d8c32077f9008193b197b7",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14352976245757926,
    "peak": 0.9623267650604248,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012431790994014591,
      "rss_bytes": 8448835584,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56344903680,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63881330688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2137161728,
      "system_wired_bytes": 8152301568,
      "system_compressor_bytes": 2731229184,
      "system_file_backed_bytes": 30657904640,
      "physical_footprint_bytes": 8759598064,
      "process_lifetime_peak_footprint_bytes": 11033320360,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 146718992,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 78.06607454101322,
      "rss_bytes": 8453242880,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345034752,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 61755703296,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 528384000,
      "system_wired_bytes": 10941579264,
      "system_compressor_bytes": 2730655744,
      "system_file_backed_bytes": 28463775744,
      "physical_footprint_bytes": 10535164912,
      "process_lifetime_peak_footprint_bytes": 12423273456,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140083472,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12397059056,
      "process_lifetime_peak_footprint_bytes": 12423273456,
      "rss_bytes": 8454586368,
      "system_available_bytes": 64631046144,
      "system_compressor_bytes": 2731229184,
      "system_file_backed_bytes": 30791516160,
      "system_free_bytes": 3307700224,
      "system_swap_in_bytes": 56345034752,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11020468224
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
中文 · FULL · 官方完整乐谱 8步快速模式提速 1.13x (RTF 1.503)

R&B

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:115abed0e655f5fa9b1409b3

官方原版 Demo

Official Reference

时长: 152.7s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 152.3s | 耗时: 258.0s | RTF: 1.695 (0.59x 实时)
NAR声学求解: 182.2s (32步) | VAE: 6.4s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 152.3s | 耗时: 228.8s | RTF: 1.503 (0.67x 实时)
NAR声学求解: 149.0s (8步 · 1.22x 提速) | VAE: 11.1s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Modern Chinese pop R&B with synth-pop influence, emotional and intimate atmosphere, Intro starts with a clear and emotional lead synth melody, simple and memorable, soft and slightly fragile, close-mic, expressive and melodic, Clean pop drum groove, tight punchy kick, natural snare with soft room reverb, minimal hi-hat pattern, smooth R&B bounce, Warm melodic synth bass with smooth movement, Analog-style synths with slight detune, warm tone, gentle filter movement, soft low-pass, subtle chorus width, Lead synth is smooth and rounded, emotional and expressive, High-frequency bell-like pluck synth in the background, soft and bright, short notes with light delay and reverb, adding sparkle and rhythmic detail, Wide atmospheric pads with slow attack and breathing movement, Minimal and clean arrangement, layered but spacious, focusing on melody and texture, Wide stereo image, airy reverb and delay, polished and warm, Pop music, Well-defined rhythm , Well-defined rhythm

歌词 (Lyrics)

[Intro]

[Verse 1]
交易命运的改变 源头在起心动念
这是改变的关键 但很多人视而不见
大赚大亏踏空卖飞 内在秩序不要乱
保持平和的心念 净值弹指间改变
当你大赚庆幸时 请你安静想想看
这笔盈利隐含多少潜在代价来交换
当你清楚知晓每次交易对应的代价
策略的赔率带给你先胜后战的心安

[Pre-Chorus]
当你选择策略保守会避开一些风险
模式内机会出现时进攻必然受限
当你选择激进策略博弈猛干一段
大赚的概率提升 大面概率也提升

[Chorus]
当你大亏懊悔时 请你冷静想想看
当下的首要任务是稳住还是继续猛干
如果内心秩序混乱还想力挽狂澜
连续重仓猛干会让你连续吃大面
此时赶紧降低期待调整心中的不甘
要警惕出其不意博弈小概率事件

[Verse 2]
因为赌性会在不顺之时出来作乱
回头发现 努力回血反而不如摆烂
当你交易卖飞时
请你转念想想看 后续行情是否出现一个更好的答案
其实时间拉长看 每条净值的曲线
那是模式画的图
更是心性的变现

[Bridge]
世间 一切都是万物因缘合和的结果
很多事
都不是我们想怎样就怎样的
走势无常才是交易基本的常态
于是先要接纳它 否则痛苦一直在
所有绊脚石
换个角度也是垫脚石
在你主动站上去看更高的风景时

[Verse 3]
当你后悔当初时 请你转念想想看
难道当初不那样 做现在真的会更棒 不那样现在会怎样
都是你的想象
况且日子还很长 不要过早下判

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/16
Q:1/4=98
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:Bb
% intro
V: Vocal
Z|"Ebmaj7"z16|"Ebm(maj7)"z16|"Dm7"z16|
V: Ins
z4f2e2d2e2d2B2|d2GdzGdzGdzGdzGd|d2^FdzFd2FdzFdzF2|c2FczFczFczFc2Fc|
V: Vocal
"Gm7"z12z2BG|
V: Ins
A3B3f6z4|
% verse
V: Vocal
"Ebmaj7"BGBGB2GGBGBBd2BG|"Ebm(maj7)"BGBGBGBGBGBBd2BB|"Dm7"AAAAAAAGAAAGA2AG|"Gm7"AAAGAGAGABAB2zBB|
V: Ins
Z4|
V: Vocal
"Ebmaj7"dddddddddddf3dd|"Ebm(maj7)"dddddddddddg2f3|"Dm7"cccccccccccccdB2|"Gm7"cccccccccccdcBB2|
V: Ins
Z4|
% pre-chorus
V: Vocal
"Ebmaj7"z2ddGGddGGGGddcd|"Ebm(maj7)"c2ccFFccFFccFFcc-|"Dm7"c2ccFFccFFccFFcc-|"Gm7"c2cBcBcBcBcBcBcd-|
V: Ins
Z4|
V: Vocal
"Gm7"d2z6f2e2d2e2|
V: Ins
Z|
% chorus
V: Vocal
"Ebmaj7"d2G2B2Gd3G2B2Gd-|"Ebm(maj7)"d2c4z2f2e2d2e2|"Dm7"d2F2A2Fd3F2A2F2|"Gm7"c3c3B2d2z2Bcde|
V: Ins
Z4|
V: Vocal
"Ebmaj7"d2G2B2Gd3G2B2Gd-|"Ebm(maj7)"d2c4z2f2e2d2e2|"Dm7"d2F2A2Fd3F2A2F2|"Gm7"c3B4z3dcdcdc|
V: Ins
Z4|
V: Vocal
"Ebmaj7"d2g2z2d2c2B2cBcB|"Ebm(maj7)"d4z6dcdcdc|"Dm7"d2a4f4d2c2B2|"Gm7"c3B4z6zBB|
V: Ins
Z4|
% verse
V: Vocal
"Ebmaj7"ddBBddBBddBd3BB|"Ebm(maj7)"ddBBddBBddBd3dd|"Dm7"cccccccccccccccc|"Gm7"cccccccdc2B2B2z2|
V: Ins
Z4|
V: Vocal
"Ebmaj7"z2dcdcdcd2G2d2d2|"Ebm(maj7)"z2dcdcdcd2G2d2d2|"Dm7"z2cBcBcBc2F2c2c2|"Gm7"z2cBcBcdc2B4BB|
V: Ins
Z4|
% pre-chorus
V: Vocal
"Ebmaj7"dcdcdcdcdcddG2BB|"Ebm(maj7)"dcdcdcdcdcddG2BB|"Dm7"cBcBcBcBc2d2d2BB|"Gm7"cBcBcBcBc2d2d2G2|
V: Ins
Z4|
V: Vocal
"Ebmaj7"d4g4a4b4|"Ebm(maj7)"d6c6z2BB|"Dm7"c4f4g4a4|"Gm7"a3f3g3b3c'c'z2|
V: Ins
Z4|
V: Vocal
"Ebmaj7"z2dcdcdcdcdcd2G2|"Ebm(maj7)"d2dcdcdcdcdcd2G2|"Dm7"d2d2z2d2d2z2cBcB|"Gm7"d2z4Bz6zBB|
V: Ins
Z4|
% chorus
V: Vocal
"Ebmaj7"d2G2B2Gd3GGB2Gd-|"Ebm(maj7)"d2c4z2f2e2d2e2|"Dm7"d2F2A2Fd3F2A2F2|"Gm7"c3c3B2d2z2Bcde|
V: Ins
Z4|
V: Vocal
"Ebmaj7"d2G2B2Gd3G2B2Gd-|"Ebm(maj7)"d2c4z2f2e2d2e2|"Dm7"d2F2A2Fd3F2A2F2|"Gm7"c3B4z8z|
V: Ins
Z3|z12z2Bz|
% outro
V: Vocal
"Ebmaj7"z16|"Ebm(maj7)"z16|"Dm7"z16|"Gm7"z16|
V: Ins
d2GdzGdzGdzGd2Gd|d2^FdzFd2FdzFd2F2|c2FczFc2FczFc2Fc|A3B3f8z2|
V: Vocal
"Gm7"z16|
V: Ins
Z|

32步标准模式详细指标

{
  "id": "zh_score",
  "status": "complete",
  "case_sha256": "4bc847a703582fe65cd8b2b9bf716b3031e4efa766f4fd104e29cbbce1f2b0fc",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 152.27866666666668,
  "official_seconds": 152.72,
  "duration_ratio": 0.9971101798498342,
  "e2e_seconds": 258.0486004999839,
  "rtf": 1.6945814285650684,
  "speed_multiplier": 0.5901162276083578,
  "load_seconds": 5.196350332989823,
  "transcription_seconds": 0,
  "save_seconds": 0.4392753330175765,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1367
    },
    "semantic": {
      "seconds": 63.472836249973625,
      "prefill_seconds": 1.2251123749883845,
      "ttft_seconds": 1.22868387500057,
      "output_tokens": 3808,
      "content_tokens": 3807,
      "output_tps": 59.994167977669065,
      "prefix_tokens": 1998,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 182.23769116698531,
    "vae_seconds": 6.420643166988157,
    "load": {
      "resolve_conversion_integrity_seconds": 5.196277417009696,
      "ar_load_events_seconds": [
        0.11680804198840633
      ],
      "ar_load_seconds": 0.11680804198840633,
      "conditioning_load_events_seconds": [
        0.23702437500469387
      ],
      "conditioning_load_seconds": 0.23702437500469387,
      "nar_load_events_seconds": [
        0.13652141601778567
      ],
      "nar_load_seconds": 0.13652141601778567,
      "vae_load_events_seconds": [
        0.11104529199656099
      ],
      "vae_load_seconds": 0.11104529199656099
    },
    "e2e_seconds": 252.2732525410247
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "e0f838bd206fe8b11310d79fae6c4220891e436a4f5a20674f5b57a7c95cfe39",
  "audio_sha256": "b70ff8f3e5bfc2b0e7b6a078a50155e25d5bd772a306daf802b0c174818de0ff",
  "result_sha256": "78b2422ac0c41cfb277606488ab92daf3093ea4d75636839ecf52f561cd2fefc",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.15860733541378255,
    "peak": 1.0,
    "clipping_fraction": 1.3681058410458019e-07
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.010249125014524907,
      "rss_bytes": 51757056,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56328994816,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65221279744,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 4218208256,
      "system_wired_bytes": 7308443648,
      "system_compressor_bytes": 2906521600,
      "system_file_backed_bytes": 34812772352,
      "physical_footprint_bytes": 31883984,
      "process_lifetime_peak_footprint_bytes": 31883984,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 258.5429126250092,
      "rss_bytes": 988987392,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56329125888,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63939657728,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2698018816,
      "system_wired_bytes": 8543420416,
      "system_compressor_bytes": 2905391104,
      "system_file_backed_bytes": 35016163328,
      "physical_footprint_bytes": 1381746992,
      "process_lifetime_peak_footprint_bytes": 11870575216,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11256150230,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11782855280,
      "process_lifetime_peak_footprint_bytes": 11870575216,
      "rss_bytes": 8159985664,
      "system_available_bytes": 69785157632,
      "system_compressor_bytes": 2906537984,
      "system_file_backed_bytes": 35016163328,
      "system_free_bytes": 12537413632,
      "system_swap_in_bytes": 56329125888,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11295391744
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/zh_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "zh_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 152.28,
  "e2e_seconds": 228.81,
  "rtf": 1.503,
  "speed_multiplier": 0.67,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1367
    },
    "semantic": {
      "seconds": 63.472836249973625,
      "prefill_seconds": 1.2251123749883845,
      "ttft_seconds": 1.22868387500057,
      "output_tokens": 3808,
      "content_tokens": 3807,
      "output_tps": 59.994167977669065,
      "prefix_tokens": 1998,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 149.01,
    "vae_seconds": 11.13,
    "load_seconds": 5.2,
    "transcription_seconds": 0,
    "e2e_seconds": 228.81,
    "32step_nar_seconds": 182.24,
    "nar_speedup": 1.22
  },
  "audio_sha256": "8aa72484471a3ac5f12de8bcdffb4d38160286c376883efe8d7551c36668d303",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.15863825832675735,
    "peak": 1.0,
    "clipping_fraction": 2.7362116820916037e-07
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.014133832999505103,
      "rss_bytes": 8450424832,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345034752,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65272037376,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3892772864,
      "system_wired_bytes": 7159365632,
      "system_compressor_bytes": 2730655744,
      "system_file_backed_bytes": 28482469888,
      "physical_footprint_bytes": 8754666480,
      "process_lifetime_peak_footprint_bytes": 12423273456,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140083472,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 160.16012204100844,
      "rss_bytes": 8455340032,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64211615744,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2947366912,
      "system_wired_bytes": 8661942272,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28356706304,
      "physical_footprint_bytes": 9563200520,
      "process_lifetime_peak_footprint_bytes": 12424715272,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 234913792,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12290563080,
      "process_lifetime_peak_footprint_bytes": 12424715272,
      "rss_bytes": 8455962624,
      "system_available_bytes": 65272037376,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28536356864,
      "system_free_bytes": 3892772864,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11368316928
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
韩语 · FULL · 官方完整乐谱 8步快速模式提速 2.11x (RTF 0.912)

Emo

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:df64fef4d0ba5a18670f6c19

官方原版 Demo

Official Reference

时长: 156.9s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 156.0s | 耗时: 299.8s | RTF: 1.921 (0.52x 实时)
NAR声学求解: 219.7s (32步) | VAE: 8.2s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 156.0s | 耗时: 142.3s | RTF: 0.912 (1.10x 实时)
NAR声学求解: 64.2s (8步 · 3.42x 提速) | VAE: 6.9s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

A melancholic yet energetic Emo Hip-hop track with Pop Punk influences. Heavy distorted electric guitar riffs combined with fast-paced trap drum beats (808 bass, crisp snare). Dark, emotional, and cinematic atmosphere. Melodic Rap beat featuring a lonely acoustic guitar intro transitioning into an explosive chorus with heavy rock guitars. Deep, distorted 808 bass and sharp hi-hats. Sad, longing melody. Fast-paced Emo Rock and Hip-hop fusion. High-energy electric guitar loops, upbeat but sad melody, 140 BPM ~ 160 BPM, driving drum rhythm. A female vocal with a husky, raspy tone and heavy auto-tune. The singing style is emotional melodic rap. Add high reverb to create a lonely, atmospheric vibe. The voice should sound raw yet polished with metallic pitch correction.

歌词 (Lyrics)

[Verse 1]
늦은 밤 텅 빈 버스 안
창문에 비친 너와 내 얼굴
목적지는 아직 안 정해도
너만 있으면 되는 것 같아

[Pre-Chorus]
사실 난 겁이 많아서
길을 잃을까 봐 두려워
근데 네 손 한 번만 잡으면
지도 따윈 필요 없을 것 같아 (알잖아)

[Chorus]
나랑 어디든 간다고 해줘
길 끝이 어딘지 몰라도
나와 어디든 좋다고 해줘
창문 밖 풍경은 변해도
나만 있으면 그 어디든 괜찮다고
작게라도 말해줘
그 한마디면 돼

[Verse 2]
비 오는 빨래 냄새 나는 골목
우산 하나 나눠 쓰는 우리
비틀비틀 맞지 않는 보폭도
웃으면서 맞춰 가는 게 좋아

[Pre-Chorus]
가끔은 말이 안 통해도
표정만 봐도 다 알잖아
서툰 얘길 이어 붙이다가
그냥 네 이름만 부르게 돼 (왜 이렇게)

[Chorus]
나랑 어디든 간다고 해줘
길 끝이 어딘지 몰라도
나와 어디든 좋다고 해줘
창문 밖 풍경은 변해도
나만 있으면 그 어디든 괜찮다고
조용하게 말해줘
그 한마디면 돼
내일도 난 버틸 수 있어

[Bridge]
약속 같은 건 없어도
네 옆에 서 있는 지금이
나에게는 다야
멀리 안 가도 좋아
편의점 앞 구석자리라도
너면 돼 너면 돼 (너면 돼)

[Chorus]
나랑 어디든 간다고 해줘
잠깐 밖에 못 나가도
나와 어디든 좋다고 해줘
네 방 안 좁은 소파여도
나만 있으면 그 어디든 괜찮다고
다시 한 번 말해줘
그 한마디 듣고
영원까지 걸어가 볼래

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/16
Q:1/4=170
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:Bb
% intro
V: Vocal
z4B2B2B4f4|"Eb"f4d4d4c4|"Eb"c4B2B2B4f4|"Gm"f2d2d2d2d4c4|
V: Ins
Z4|
V: Vocal
"Gm"z4B2B2B2B2B2B2|"Eb"f4d2d4c2c4|"Eb"BzB6A4B2B2|"Gm"A4A2A4B4B2-|
V: Ins
Z4|
V: Vocal
"Gm"B4z12|
V: Ins
Z|
% verse
V: Vocal
"Eb"z2B2d2d2d2c2c2c2-|"Eb"c2B6z8|"Gm"z2B2d2d2d2c2c2c2-|"Gm"c2B4B4B2B2B2|
V: Ins
Z4|
V: Vocal
"Eb"f4d2d4c4c2-|"Eb"c2B4B4z2B2B2|"Gm"f4d2d4c4c2-|"Gm"c2d4c4B4B2-|
V: Ins
Z4|
V: Vocal
"Eb"B2z6f8|"Eb"d8c8|"F"z16|
V: Ins
Z3|
% chorus
V: Vocal
z2b2b2b2b2b2f'4|"Eb"f'2d'zd'4d'4c'4|"Eb"z2b2b2b2b2b2f'4|"Gm"f'4d'4d'4c'4|
V: Ins
Z4|
V: Vocal
"Gm"z2b2b2b2b2b2f'4|"F"f'4d'4d'4c'4|"F"z2b2b2b2b2b2f'4|"Dm"f'6a6b4-|
V: Ins
Z4|
V: Vocal
"Dm"b2z6d'2d'2d'2d'2-|"Eb"d'2c'zc'4d'2d'2d'2d'2-|"Eb"d'2c'4c'4b6|"Gm"d'4c'4d'4c'2d'2-|
V: Ins
Z4|
V: Vocal
"Gm"d'2c'4c'4b6|"F"a16|"F"f16|"Dm"z8d4d4|
V: Ins
Z4|
V: Vocal
"Dm"e6d6B4|
V: Ins
Z|
% verse
V: Vocal
"Eb"z2B2d2d2d2c2c2c2|"Eb"c2B2B4B4z4|"Gm"d2d2d2d2d2c2c2c2|"Gm"c4d4z8|
V: Ins
Z4|
V: Vocal
"Eb"d2d2d2d2d2c2c2c2|"Eb"c4B4B4B2B2|"Gm"d4d2d4c4e2-|"Gm"e2d4d2c4B4|
V: Ins
Z4|
V: Vocal
"Eb"z2f2f2f2f2f2f2g2-|"Eb"g2d4d6c4|"Gm"z2f2f2f2f2f2f2g2-|"Gm"g2d4d6c4|
V: Ins
Z4|
V: Vocal
"Eb"z4f2f2f2f2f2f2|"Eb"g4d2d4c4c2|"Gm"d4d4d4d4|"Gm"e4d4c4B4|
V: Ins
Z4|
V: Vocal
z8B4z4|
V: Ins
Z|
% chorus
V: Vocal
z2b2b2b2b2b2f'4|"Eb"f'4d'4d'4c'4|"Eb"z2b2b2b2b2b2f'4|"Gm"f'4d'4d'4c'4|
V: Ins
Z4|
V: Vocal
"Gm"z2b2b2b2b2b2f'4|"F"f'4d'4d'4c'4|"F"z2b2b2b2b2b2f'4|"Dm"f'6a6b4-|
V: Ins
Z4|
V: Vocal
"Dm"b4z4d'2d'2d'2d'2-|"Eb"d'2c'zc'4d'2d'2d'2d'2-|"Eb"d'2c'4c'4b6|"Gm"d'4c'4d'4c'2d'2-|
V: Ins
Z4|
V: Vocal
"Gm"d'2c'4c'4b6|"F"a16|"F"f16|"Dm"z8d4d4|
V: Ins
Z4|
V: Vocal
"Dm"e6d6B4|
V: Ins
Z|
% bridge
V: Vocal
"Eb"z16|"Eb"f6d6c4-|"Gm"c4z8B4|"Gm"f4f2d4c2c4-|
V: Ins
Z4|
V: Vocal
"F"c4z12|"F"z16|"F"z16|"F"z16|
V: Ins
z4A2A2A2A2A2A2|A2A2A2A2A2A2A2A2|A2A2A2A2A2A2A2A2|A2A2A2A2A2A2A2c2|
V: Vocal
"Eb"z4f4f4f4|"Eb"f4d4c4B4|"Gm"B4f4f4f4|"Gm"f4d4c4B4|
V: Ins
Z4|
V: Vocal
"F"B4c12-|"F"c16-|"F"c16-|"F"c2b2b2b2b2b2f'4|
V: Ins
Z4|
% chorus
V: Vocal
"Eb"f'2d'zd'4d'4c'4|"Eb"z2b2b2b2b2b2f'4|"Gm"f'2d'zd'4d'4c'4|"Gm"z2b2b2b2b2b2f'4|
V: Ins
Z4|
V: Vocal
"F"f'4d'4d'4c'4|"F"z2b2b2b2b2b2f'4|"Dm"f'6a6b4-|"Dm"b2z6d'2d'2d'2d'2-|
V: Ins
Z4|
V: Vocal
"Eb"d'2c'zc'4d'2d'2d'2d'2-|"Eb"d'2c'4c'4b6|"Gm"d'4c'4d'4c'2d'2-|"Gm"d'2c'4c'4b6|
V: Ins
Z4|
V: Vocal
"F"a16|"F"f16|"Dm"z8d4d4|"Dm"e6d6B4|
V: Ins
Z4|
% outro
V: Vocal
"Eb"z16|"Eb"z16|"Gm"z12d2d2|"Gm"d2d2d2d2e4d4|
V: Ins
B2B2B2B2B2B2B2B2|B2B2B2B2B2B2B2B2|G2G2G2G2G2B2z4|Z|
V: Vocal
"F"d4c12-|"F"c16-|"F"c16-|"F"c4z4z8|
V: Ins
Z4|

32步标准模式详细指标

{
  "id": "ko_score",
  "status": "complete",
  "case_sha256": "2d38527bb49dbc4dc16afee14e020c4aff37f3b8bcef24c781ba69c1fcb92fbb",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 156.03866666666667,
  "official_seconds": 156.92,
  "duration_ratio": 0.9943835500042486,
  "e2e_seconds": 299.78041066700825,
  "rtf": 1.921193105984467,
  "speed_multiplier": 0.5205098836160851,
  "load_seconds": 5.164946083037648,
  "transcription_seconds": 0,
  "save_seconds": 0.3679583329940215,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1980
    },
    "semantic": {
      "seconds": 65.99044629203854,
      "prefill_seconds": 1.376692042045761,
      "ttft_seconds": 1.3849018750479445,
      "output_tokens": 3902,
      "content_tokens": 3901,
      "output_tps": 59.12977134192771,
      "prefix_tokens": 2692,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 219.74767066596542,
    "vae_seconds": 8.172454458952416,
    "load": {
      "resolve_conversion_integrity_seconds": 5.164890291984193,
      "ar_load_events_seconds": [
        0.11963616701541469
      ],
      "ar_load_seconds": 0.11963616701541469,
      "conditioning_load_events_seconds": [
        0.16252879198873416
      ],
      "conditioning_load_seconds": 0.16252879198873416,
      "nar_load_events_seconds": [
        0.11841804103460163
      ],
      "nar_load_seconds": 0.11841804103460163,
      "vae_load_events_seconds": [
        0.10575899999821559
      ],
      "vae_load_seconds": 0.10575899999821559
    },
    "e2e_seconds": 294.04429254104616
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "0910f50d377d92b9633239791464579defbe4346888068ffa1726ac493ea6806",
  "audio_sha256": "530de8ebe1df332844f86b9854b8e62c5f8e319e591d0427be5bf76ea16c0058",
  "result_sha256": "f9a6bfe7694b7507d96f28ee065a0b64065437736e4d36c7c136c9745fb8919f",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.15526236099207205,
    "peak": 1.0,
    "clipping_fraction": 2.6702783070862776e-07
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.009647957980632782,
      "rss_bytes": 51838976,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56329125888,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63939657728,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2698018816,
      "system_wired_bytes": 8543420416,
      "system_compressor_bytes": 2905391104,
      "system_file_backed_bytes": 35016163328,
      "physical_footprint_bytes": 31965904,
      "process_lifetime_peak_footprint_bytes": 31965904,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 300.20517570798984,
      "rss_bytes": 752517120,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56329289728,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64987594752,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3830333440,
      "system_wired_bytes": 7345537024,
      "system_compressor_bytes": 2903179264,
      "system_file_backed_bytes": 35064840192,
      "physical_footprint_bytes": 910821464,
      "process_lifetime_peak_footprint_bytes": 12118252192,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 156781600,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11969518216,
      "process_lifetime_peak_footprint_bytes": 12118252192,
      "rss_bytes": 8162607104,
      "system_available_bytes": 69856018432,
      "system_compressor_bytes": 2905407488,
      "system_file_backed_bytes": 35064840192,
      "system_free_bytes": 12392005632,
      "system_swap_in_bytes": 56329289728,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11378753536
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/ko_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "ko_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 156.04,
  "e2e_seconds": 142.3,
  "rtf": 0.912,
  "speed_multiplier": 1.1,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1980
    },
    "semantic": {
      "seconds": 65.99044629203854,
      "prefill_seconds": 1.376692042045761,
      "ttft_seconds": 1.3849018750479445,
      "output_tokens": 3902,
      "content_tokens": 3901,
      "output_tps": 59.12977134192771,
      "prefix_tokens": 2692,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 64.23,
    "vae_seconds": 6.91,
    "load_seconds": 5.16,
    "transcription_seconds": 0,
    "e2e_seconds": 142.3,
    "32step_nar_seconds": 219.75,
    "nar_speedup": 3.42
  },
  "audio_sha256": "0971a96ddbd7e1fba87727c8f231d943a69f03761806b3ce5aaadc925d95a812",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1554732292968558,
    "peak": 1.0,
    "clipping_fraction": 2.0027087303147084e-07
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012680708023253828,
      "rss_bytes": 8458076160,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64211615744,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2947366912,
      "system_wired_bytes": 8661942272,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28356706304,
      "physical_footprint_bytes": 8857148424,
      "process_lifetime_peak_footprint_bytes": 12424715272,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 234913792,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 71.15808379102964,
      "rss_bytes": 8459419648,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63168069632,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1914912768,
      "system_wired_bytes": 9438363648,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28418785280,
      "physical_footprint_bytes": 9557449736,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 137257000,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11207531582,
      "mlx_cache_bytes": 234913792,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12383443976,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8462172160,
      "system_available_bytes": 64645758976,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28418801664,
      "system_free_bytes": 3396583424,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 9986981888
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
英语 · FULL · 官方完整乐谱 8步快速模式提速 2.07x (RTF 0.747)

Rock & Roll

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:ad4ec0a4f236082a2933b48d

官方原版 Demo

Official Reference

时长: 95.8s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 95.0s | 耗时: 146.9s | RTF: 1.547 (0.65x 实时)
NAR声学求解: 99.9s (32步) | VAE: 4.5s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 95.0s | 耗时: 70.9s | RTF: 0.747 (1.34x 实时)
NAR声学求解: 25.3s (8步 · 3.94x 提速) | VAE: 3.8s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Joyful happy 50s male powerful

歌词 (Lyrics)

[Verse]
Snow is falling thick and bright
Reindeer footprints in the night
Cookies vanish
Milk is gone
Who's that laughing till the dawn

[Prechorus]
Chimneys whisper
Magic flows
Candy canes in secret rows

[Chorus]
Christmas time with Santa Claus
Sliding rooftops
No time to pause
Ho ho ho
He’s on his way
Bringing joy to Christmas Day

[Verse 2]
Stockings hang by the fire’s glow
Toys and dreams in a sack to go
Elves are dancing
Wrapping fast
Every wish will make it past

[Prechorus]
Midnight sparkles
Sleigh bells sing
Feel the joy that he will bring

[Chorus]
Christmas time with Santa Claus
Sliding rooftops
No time to pause
Ho ho ho
He’s on his way
Bringing joy to Christmas Day

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/16
Q:1/4=175
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:F
% intro
V: Vocal
z8"F"z8|"F"z16|"F"z16|"F"z16|
V: Ins
Z|f4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|
V: Vocal
"F"z16|"F"z16|"Dm7"z16|"Bb"z16|
V: Ins
F4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|
V: Vocal
"C"z16|
V: Ins
c2z2c4c4z4|
% verse
V: Vocal
"F"A4c4d4A4|"Dm"G2F4A6z4|"Bb"D4F4G4A4|"C"G4C2C6z4|
V: Ins
Z4|
V: Vocal
"F"A2c6d2A6|"Dm"G4F2A6z4|"Bb"D4F4G4A4|"C"G4C2C6z4|
V: Ins
Z4|
V: Vocal
"Bb"D4F4G4A4|"Bb"D2F4G6z4|"C"c4c2c6d4|"C"e2e4d2c8|
V: Ins
Z4|
% chorus
V: Vocal
"F"f4d4c4d4|"Dm"A2FzF2F6z2d2-|"Bb"d2c2d4c4d4|"C"G4G2G6z4|
V: Ins
Z4|
V: Vocal
"F"f4d4c4d4|"Dm"G2F4A6z4|"Bb"D2F4G6A4|"C"c8A8|
V: Ins
Z4|
V: Vocal
"F"G2F12-F2-|"Dm7"F8-F2z6|"Bb"z16|"C"z16|
V: Ins
Z|z8z2c2A2G2|F4f4d2c2A2G2|c4c4c4z4|
% verse
V: Vocal
"F"A2c4d6A2A2|"Dm"G2F4A6z4|"Bb"D4F4G4A2A2|"C"G4C2C6z4|
V: Ins
Z4|
V: Vocal
"F"A4c2d6A4|"Dm"G2F4A6z4|"Bb"D2F4G6A4|"C"G2C4C6z4|
V: Ins
Z4|
V: Vocal
"Bb"D4F4G4A4|"Bb"D2F4G6z4|"C"c4c2c6d4|"C"e2e4d2c8|
V: Ins
Z4|
% chorus
V: Vocal
"F"f4d4c4d4|"Dm"G2FzF2F6z2d2-|"Bb"d2c2d4c4d4|"C"G4G2G6z4|
V: Ins
Z4|
V: Vocal
"F"f4d4c4d4|"Dm"G2F4A6z4|"Bb"D2F4G6A4|"C"c8A8|
V: Ins
Z4|
V: Vocal
"F"G2F12-F2-|"Dm7"F8-F2z6|"Bb"z16|"C"z16|
V: Ins
Z|z8z2c2A2G2|F4f4d2c2A2G2|c4c4c4z4|
% outro
V: Vocal
"F"z16|"Dm7"z16|"Bb"z16|"C"z16|
V: Ins
f4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|c4c4c4z4|
V: Vocal
"F"z16|"Dm7"z16|"Bb"z16|"C"z16|
V: Ins
f4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|c4c4c4z4|
V: Vocal
"F"z16|"F"z16|"F"z16|"F"z16|
V: Ins
F4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|F4f4d2c2A2G2|
V: Vocal
"F"z16|"F"z16|
V: Ins
F4z12|Z|

32步标准模式详细指标

{
  "id": "en_score",
  "status": "complete",
  "case_sha256": "f0418030dbdc46c67c1a9ba135dddc001f755d501ef50bf6e5ccd0d2c1d18778",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 94.99866666666667,
  "official_seconds": 95.76,
  "duration_ratio": 0.9920495683653577,
  "e2e_seconds": 146.92567508301,
  "rtf": 1.5466077602809511,
  "speed_multiplier": 0.6465763496611083,
  "load_seconds": 5.267140208045021,
  "transcription_seconds": 0,
  "save_seconds": 0.2570144580095075,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1248
    },
    "semantic": {
      "seconds": 36.48848658299539,
      "prefill_seconds": 1.0328282919945195,
      "ttft_seconds": 1.0356238330132328,
      "output_tokens": 2376,
      "content_tokens": 2375,
      "output_tps": 65.11643048268489,
      "prefix_tokens": 1482,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 99.93519991700305,
    "vae_seconds": 4.497567957965657,
    "load": {
      "resolve_conversion_integrity_seconds": 5.2670660829753615,
      "ar_load_events_seconds": [
        0.12442404200555757
      ],
      "ar_load_seconds": 0.12442404200555757,
      "conditioning_load_events_seconds": [
        0.20056979201035574
      ],
      "conditioning_load_seconds": 0.20056979201035574,
      "nar_load_events_seconds": [
        0.13628183299442753
      ],
      "nar_load_seconds": 0.13628183299442753,
      "vae_load_events_seconds": [
        0.10732383403228596
      ],
      "vae_load_seconds": 0.10732383403228596
    },
    "e2e_seconds": 141.06340920901857
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "68e37c7e7b12a2825e49a45a87916820c250ccb52f06337b14eda5f4ce48dbc3",
  "audio_sha256": "3fef195cc9776ca1912f484cdc19fed79eb9df0c9a20ba3cb30a78d2c0d8adcf",
  "result_sha256": "17a56e384f1005060e4ed2f8dd64b0a444ce238fdaadd82271c4c5d6f420e074",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1444062875937603,
    "peak": 0.914916455745697,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.011967124999500811,
      "rss_bytes": 51691520,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56329289728,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64987594752,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3830333440,
      "system_wired_bytes": 7345537024,
      "system_compressor_bytes": 2903179264,
      "system_file_backed_bytes": 35064840192,
      "physical_footprint_bytes": 31834856,
      "process_lifetime_peak_footprint_bytes": 31834856,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 147.22970483300742,
      "rss_bytes": 924106752,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56330141696,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63983616000,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2658271232,
      "system_wired_bytes": 8259878912,
      "system_compressor_bytes": 2903408640,
      "system_file_backed_bytes": 35047292928,
      "physical_footprint_bytes": 1226983800,
      "process_lifetime_peak_footprint_bytes": 12098951888,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196732990,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11987426000,
      "process_lifetime_peak_footprint_bytes": 12098951888,
      "rss_bytes": 8219082752,
      "system_available_bytes": 70019235840,
      "system_compressor_bytes": 2903621632,
      "system_file_backed_bytes": 35064840192,
      "system_free_bytes": 12387696640,
      "system_swap_in_bytes": 56330141696,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11021893632
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/en_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "en_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 95.0,
  "e2e_seconds": 70.93,
  "rtf": 0.747,
  "speed_multiplier": 1.34,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1248
    },
    "semantic": {
      "seconds": 36.48848658299539,
      "prefill_seconds": 1.0328282919945195,
      "ttft_seconds": 1.0356238330132328,
      "output_tokens": 2376,
      "content_tokens": 2375,
      "output_tps": 65.11643048268489,
      "prefix_tokens": 1482,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 25.34,
    "vae_seconds": 3.83,
    "load_seconds": 5.27,
    "transcription_seconds": 0,
    "e2e_seconds": 70.93,
    "32step_nar_seconds": 99.94,
    "nar_speedup": 3.94
  },
  "audio_sha256": "96bb44c9d71f92d832f2814267de1dc20581058ec8687f5d84df41e304f99770",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14458394336299796,
    "peak": 0.9200946092605591,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012403832981362939,
      "rss_bytes": 8464908288,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63168069632,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1914912768,
      "system_wired_bytes": 9438363648,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28418785280,
      "physical_footprint_bytes": 8766430216,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 137257000,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 29.191721457988024,
      "rss_bytes": 8475181056,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62300946432,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1722793984,
      "system_wired_bytes": 9603022848,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28458680320,
      "physical_footprint_bytes": 9683737632,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 144369688,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12152658952,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8475181056,
      "system_available_bytes": 64368738304,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28458680320,
      "system_free_bytes": 3519922176,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11621892096
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
英语 · FULL · 相同文本 · 重新作曲 8步快速模式提速 2.33x (RTF 1.032)

Comedy

pass · 完成

输入官方原始歌词与风格;模型自动规划乐谱并端到端生成歌曲。 官方案例 ID:2998556d4b082ca29c856652

官方原版 Demo

Official Reference

时长: 204.9s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 209.9s | 耗时: 504.2s | RTF: 2.402 (0.42x 实时)
NAR声学求解: 356.2s (32步) | VAE: 13.1s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 209.9s | 耗时: 216.5s | RTF: 1.032 (0.97x 实时)
NAR声学求解: 75.2s (8步 · 4.74x 提速) | VAE: 7.3s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Pop genre, Happy mood, Live vocal style, Upbeat production

歌词 (Lyrics)

[Verse 1]
Thanksgiving morning, sun begins to gleam
The family's gathering, living out a dream
Turkey in the oven, pies all lined up neat
But something's brewing, can't be beat
Sierra Zemke's got a story, whispers in the hall
About a Jedi master, standing proud and tall
Obi-Wan in jail, the news has spread so far
While Kirk Cameron's preaching, playing his guitar

[Chorus]
Oh, Thanksgiving day, a feast for all to see
With secret farts on YouTube, for the world to agree
Jar Jar Binks is laughing, from a galaxy away
It's a bizarre holiday, in every single way

[Verse 2]
Dinner's on the table, laughter fills the air
Gravy on the mashed potatoes, without a care
Uncle Bob tells stories, of birds that fly in jail
And a gun that shot no bullets, an unbelievable tale
We're watching viral videos, a screen aglow and bright
Someone caught a moment, a comedic sight
Of gas escaping quickly, a sound both deep and low
And everyone's just giggling, because they know, they know

[Chorus]
Oh, Thanksgiving day, a feast for all to see
With secret farts on YouTube, for the world to agree
Jar Jar Binks is laughing, from a galaxy away
It's a bizarre holiday, in every single way

[Bridge]
From Hollywood to kitchens, the strange tales intertwine
Of movie stars and flatulence, a truly wild design
Who knew that these odd pieces, would fit into one day?
But on Thanksgiving, anything can go astray

[Chorus]
Oh, Thanksgiving day, a feast for all to see
With secret farts on YouTube, for the world to agree
Jar Jar Binks is laughing, from a galaxy away
It's a bizarre holiday, in every single way

[Outro]
So pass the cranberry sauce, and don't forget the pie
Another year of weirdness, as the hours fly by
Happy Thanksgiving, everyone, let the good times roll
With YouTube farts and Jedi tales, reaching every soul.

输入乐谱 (ABC Notation)

无固定乐谱输入(由模型自主规划或无乐谱生成)

本机实际生成 / 转谱的完整乐谱

X:1
T:
M:4/4
L:1/32
Q:1/4=186
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:D
% intro
V: Vocal
z24z4"D"z4|"D"z32|"A"z32|"Bm"z32|
V: Ins
Z|z8d'8d'4e'4f'4a'4|z16f'8e'8|d'8b8d'4e'4f'4b4|
V: Vocal
"G"z32|"D"z32|"A"z32|"Bm"z32|
V: Ins
Z|z8d'8d'4e'4f'4a'4|z16f'8e'8|d'8b8d'4e'4f'4b4|
V: Vocal
"G"z32|
V: Ins
Z|
% verse
V: Vocal
"D"z8d8d4A8A4-|"A"A4F8z16z4|"Bm"B8B4B8A8F4-|"G"F8z24|
V: Ins
f8z24|Z3|
V: Vocal
"D"z8d4d8A8A4-|"A"A4F4E8z16|"Bm"B4d8d8e8d4-|"G"d12z16f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z12f4-|"Bm"f4f8f8d8B4-|"G"B8z24|
V: Ins
Z4|
V: Vocal
"D"z8d8d4A8A4-|"A"A4F8z16z4|"Bm"B16d12d4-|"G"d4z8B4d4B8d4-|
V: Ins
Z4|
V: Vocal
"D"d4B8d8B8d4-|"A"d4B8z16f4-|"Bm"f4e8e8d8B4-|"G"B8z16z4A4|
V: Ins
Z4|
V: Vocal
"D"d4B8d8B8d4-|"A"d4B8z16f4-|"Bm"f4e8e8d8d4-|"G"d8z16z4a4-|
V: Ins
Z4|
V: Vocal
"D"a4f8f8e8e4-|"A"e12z12d4a4-|"Bm"a4f8f8e8e4|"G"d16z12f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z16|"Bm"f4f8f8d8d4-|"G"d12z4e16|
V: Ins
Z4|
% chorus
V: Vocal
"D"f16a4f8e4-|"A"e8z16d4f4-|"Bm"f8e4f8d8B4-|"G"B8z16A8|
V: Ins
Z4|
V: Vocal
"D"f8f8a8f8|"A"e8d12d4d4f4-|"Bm"f12z4f4d8d4-|"G"d8z16z4f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z4d4d4a4-|"Bm"a4f8f8e8e4|"G"d16z8A8|
V: Ins
Z4|
V: Vocal
"D"f4f4f8a4f8e4-|"A"e12z12d4f4-|"Bm"f8e8f4d8d4-|"G"d12z16z4|
V: Ins
Z4|
% verse
V: Vocal
"D"d4d8d8A8A4-|"A"A4F8z16z4|"Bm"B8A4B8A8F4-|"G"F8z16z4d4-|
V: Ins
Z4|
V: Vocal
"D"d4d8d12A4A4-|"A"A8F4A8B12|"Bm"z8d4d8e8d4-|"G"d12z16f4|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z8d4f4-|"Bm"f8f4f8d8B4-|"G"B12z8d4d4f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4|"A"d8z16d4f4-|"Bm"f8f4f8d4d4d4-|"G"d8z16d8|
V: Ins
Z4|
V: Vocal
"D"d4B8d8B8d4|"A"B4A12z8A4f4-|"Bm"f8e4e8d8B4-|"G"B8z24|
V: Ins
Z4|
V: Vocal
"D"d4B8d12B4d4-|"A"d4B8z16f4-|"Bm"f8e4e8d8d4-|"G"d12z12d4a4-|
V: Ins
Z4|
V: Vocal
"D"a8f4f8e8e4-|"A"e4d12z8d4a4-|"Bm"a4f8f8e8e4|"G"d16z8d4f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4|"A"d4d12z8d4f4-|"Bm"f8f8f8f8|"G"f16e16|
V: Ins
Z4|
% chorus
V: Vocal
"D"f16a4f8e4-|"A"e12z12d4f4-|"Bm"f8f4f8d8B4-|"G"B8z16A8|
V: Ins
Z4|
V: Vocal
"D"f8f8a8f8|"A"e8d12d4d4f4-|"Bm"f12z4f4d8d4-|"G"d12z16f4-|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z4d4d4a4-|"Bm"a4f8f8e8e4|"G"d16z8A8|
V: Ins
Z4|
V: Vocal
"D"f4f4f8a4f8e4-|"A"e12z12d4f4-|"Bm"f8f8f4d8d4-|"G"d12z12d8|
V: Ins
Z4|
% bridge
V: Vocal
"D"d4B8d12B4d4-|"A"d4B8z12A4f4-|"Bm"f8e8e4d8B4-|"G"B8z16z4A4|
V: Ins
Z4|
V: Vocal
"D"d4B8d12B4d4-|"A"d4B4A8z12A4|"Bm"f8e4e12d4d4-|"G"d12z12a8|
V: Ins
Z4|
V: Vocal
"D"a4f8f8e8e4-|"A"e4d12z8d8|"Bm"a4f8f8e8e4|"G"d16z8d4f4|
V: Ins
Z4|
V: Vocal
"D"f16f16|"A"f4e12f4e8f4-|"Bm"f8f8f8f8|"G"f32|
V: Ins
Z4|
V: Vocal
"G"e32|
V: Ins
Z|
% chorus
V: Vocal
"D"f16a4f8e4-|"A"e8z16d4f4-|"Bm"f8f4f8d8B4-|"G"B8z16A8|
V: Ins
Z4|
V: Vocal
"D"f8f8a8f8|"A"e8d12d4d4f4-|"Bm"f12z4f4d8d4-|"G"d8z16z4f4|
V: Ins
Z4|
V: Vocal
"D"f4f8f8e8e4-|"A"e4d12z4d4d4a4-|"Bm"a4f8f8e8e4|"G"d16z8A8|
V: Ins
Z4|
V: Vocal
"D"f4f4f8a4f8e4-|"A"e12z12d4f4-|"Bm"f8f8f4d8d4-|"G"d8z16d8|
V: Ins
Z4|
% outro
V: Vocal
"D"d8B4d8d4B4d4-|"A"d8z16d4f4-|"Bm"f8e4e8d8B4-|"G"B8z16z4A4|
V: Ins
Z4|
V: Vocal
"D"d4B8d12B4d4-|"A"d4B8z12A4A4|"Bm"f12e8d12|"G"d16z16|
V: Ins
Z4|
V: Vocal
"D"a12f8e8e4|"A"d4d12d4d4d8|"Bm"a8f4f8e12|"G"e4d16z4d8|
V: Ins
Z4|
V: Vocal
"D"f8f8f8e8|"A"e4d8d16z4|"Bm"f16f16|"G"f16f16|
V: Ins
Z4|
V: Vocal
"D"f32|"A"e32|"Bm"d32-|"G"d32|
V: Ins
Z4|
V: Vocal
"D"z32|"A"z32|"Bm"z32|"G"z32|
V: Ins
z8d'4z4d'4e'4f'4a'4|z16f'8e'8|d'8b8d'4e'4f'4b4|Z|
V: Vocal
"D"z32|"D"z32|
V: Ins
d'32|Z|

32步标准模式详细指标

{
  "id": "full_generated",
  "status": "complete",
  "case_sha256": "72df4bb21da03afede40459e879fbfc99af8fb70b5af98cceb9358134aaf3b86",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 209.91866666666667,
  "official_seconds": 204.88,
  "duration_ratio": 1.0245932578419887,
  "e2e_seconds": 504.1577927910257,
  "rtf": 2.401681569326973,
  "speed_multiplier": 0.4163749319524618,
  "load_seconds": 5.259764832968358,
  "transcription_seconds": 0,
  "save_seconds": 0.6404737089760602,
  "timing": {
    "abc": {
      "seconds": 44.032045375031885,
      "prefill_seconds": 0.2251760830404237,
      "ttft_seconds": 0.22886995802400634,
      "output_tokens": 2627,
      "content_tokens": 2626,
      "output_tps": 59.66109404242268,
      "prefix_tokens": 497,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "semantic": {
      "seconds": 84.81278020801255,
      "prefill_seconds": 1.7360337499994785,
      "ttft_seconds": 1.7377602080232464,
      "output_tokens": 5249,
      "content_tokens": 5248,
      "output_tps": 61.88925757564199,
      "prefix_tokens": 3125,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 356.23452941601863,
    "vae_seconds": 13.093939333048183,
    "load": {
      "resolve_conversion_integrity_seconds": 5.259697625006083,
      "ar_load_events_seconds": [
        0.12291249999543652
      ],
      "ar_load_seconds": 0.12291249999543652,
      "conditioning_load_events_seconds": [
        0.20717633399181068
      ],
      "conditioning_load_seconds": 0.20717633399181068,
      "nar_load_events_seconds": [
        0.1273670409573242
      ],
      "nar_load_seconds": 0.1273670409573242,
      "vae_load_events_seconds": [
        0.10023258399451151
      ],
      "vae_load_seconds": 0.10023258399451151
    },
    "e2e_seconds": 498.31734858301934
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "5a8691af20b1d67888c9786f71fae99740b711fd612538cfc7fcafb58f0348da",
  "audio_sha256": "f9764eca715115d42d63fbc7e0872f565e6f4780911252efd551a7cfb24050d4",
  "result_sha256": "80a6d4b4f7e5a3c037b42237c583a35f632203facbe90473ad5a2449911f4566",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.15009467599607557,
    "peak": 0.9160799980163574,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.011119750037323684,
      "rss_bytes": 51658752,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56330141696,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63983616000,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2658271232,
      "system_wired_bytes": 8259878912,
      "system_compressor_bytes": 2903408640,
      "system_file_backed_bytes": 35047292928,
      "physical_footprint_bytes": 31785680,
      "process_lifetime_peak_footprint_bytes": 31785680,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 504.8804157920531,
      "rss_bytes": 1056571392,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56331911168,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 68778950656,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 10470260736,
      "system_wired_bytes": 7700496384,
      "system_compressor_bytes": 2761113600,
      "system_file_backed_bytes": 34686304256,
      "physical_footprint_bytes": 1373898888,
      "process_lifetime_peak_footprint_bytes": 12140796600,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12018440888,
      "process_lifetime_peak_footprint_bytes": 12140796600,
      "rss_bytes": 8189100032,
      "system_available_bytes": 70283558912,
      "system_compressor_bytes": 2903949312,
      "system_file_backed_bytes": 35142369280,
      "system_free_bytes": 12560564224,
      "system_swap_in_bytes": 56331911168,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11424907264
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/full_generated",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "full_generated",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 209.92,
  "e2e_seconds": 216.54,
  "rtf": 1.032,
  "speed_multiplier": 0.97,
  "timing": {
    "abc": {
      "seconds": 44.032045375031885,
      "prefill_seconds": 0.2251760830404237,
      "ttft_seconds": 0.22886995802400634,
      "output_tokens": 2627,
      "content_tokens": 2626,
      "output_tps": 59.66109404242268,
      "prefix_tokens": 497,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "semantic": {
      "seconds": 84.81278020801255,
      "prefill_seconds": 1.7360337499994785,
      "ttft_seconds": 1.7377602080232464,
      "output_tokens": 5249,
      "content_tokens": 5248,
      "output_tps": 61.88925757564199,
      "prefix_tokens": 3125,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 75.16,
    "vae_seconds": 7.28,
    "load_seconds": 5.26,
    "transcription_seconds": 0,
    "e2e_seconds": 216.54,
    "32step_nar_seconds": 356.23,
    "nar_speedup": 4.74
  },
  "audio_sha256": "dafd7bdbf01b522c7f988ce1a172e43212063c23dcc1bd48d187c20942c81ab9",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14989543603332037,
    "peak": 0.9311237335205078,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012093208031728864,
      "rss_bytes": 8474083328,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345231360,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62300946432,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1722793984,
      "system_wired_bytes": 9603022848,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28458680320,
      "physical_footprint_bytes": 8782814240,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 144369688,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 82.45020262501203,
      "rss_bytes": 8476278784,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62914707456,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1533313024,
      "system_wired_bytes": 8739274752,
      "system_compressor_bytes": 2730852352,
      "system_file_backed_bytes": 28524134400,
      "physical_footprint_bytes": 9226837024,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 159255064,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 167706624,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12062039072,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8478392320,
      "system_available_bytes": 64999505920,
      "system_compressor_bytes": 2730868736,
      "system_file_backed_bytes": 28524134400,
      "system_free_bytes": 3727998976,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 9852829696
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
中文 · MELODY · 相同文本 · 重新作曲 8步快速模式提速 2.9x (RTF 0.797)

R&B

pass · 完成

输入官方原始歌词与风格;模型自动规划乐谱并端到端生成歌曲。 官方案例 ID:115abed0e655f5fa9b1409b3

官方原版 Demo

Official Reference

时长: 152.7s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 181.0s | 耗时: 417.9s | RTF: 2.309 (0.43x 实时)
NAR声学求解: 321.0s (32步) | VAE: 9.5s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 181.0s | 耗时: 144.3s | RTF: 0.797 (1.25x 实时)
NAR声学求解: 51.3s (8步 · 6.26x 提速) | VAE: 6.3s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Modern Chinese pop R&B with synth-pop influence, emotional and intimate atmosphere, Intro starts with a clear and emotional lead synth melody, simple and memorable, soft and slightly fragile, close-mic, expressive and melodic, Clean pop drum groove, tight punchy kick, natural snare with soft room reverb, minimal hi-hat pattern, smooth R&B bounce, Warm melodic synth bass with smooth movement, Analog-style synths with slight detune, warm tone, gentle filter movement, soft low-pass, subtle chorus width, Lead synth is smooth and rounded, emotional and expressive, High-frequency bell-like pluck synth in the background, soft and bright, short notes with light delay and reverb, adding sparkle and rhythmic detail, Wide atmospheric pads with slow attack and breathing movement, Minimal and clean arrangement, layered but spacious, focusing on melody and texture, Wide stereo image, airy reverb and delay, polished and warm, Pop music, Well-defined rhythm , Well-defined rhythm

歌词 (Lyrics)

[Intro]

[Verse 1]
交易命运的改变 源头在起心动念
这是改变的关键 但很多人视而不见
大赚大亏踏空卖飞 内在秩序不要乱
保持平和的心念 净值弹指间改变
当你大赚庆幸时 请你安静想想看
这笔盈利隐含多少潜在代价来交换
当你清楚知晓每次交易对应的代价
策略的赔率带给你先胜后战的心安

[Pre-Chorus]
当你选择策略保守会避开一些风险
模式内机会出现时进攻必然受限
当你选择激进策略博弈猛干一段
大赚的概率提升 大面概率也提升

[Chorus]
当你大亏懊悔时 请你冷静想想看
当下的首要任务是稳住还是继续猛干
如果内心秩序混乱还想力挽狂澜
连续重仓猛干会让你连续吃大面
此时赶紧降低期待调整心中的不甘
要警惕出其不意博弈小概率事件

[Verse 2]
因为赌性会在不顺之时出来作乱
回头发现 努力回血反而不如摆烂
当你交易卖飞时
请你转念想想看 后续行情是否出现一个更好的答案
其实时间拉长看 每条净值的曲线
那是模式画的图
更是心性的变现

[Bridge]
世间 一切都是万物因缘合和的结果
很多事
都不是我们想怎样就怎样的
走势无常才是交易基本的常态
于是先要接纳它 否则痛苦一直在
所有绊脚石
换个角度也是垫脚石
在你主动站上去看更高的风景时

[Verse 3]
当你后悔当初时 请你转念想想看
难道当初不那样 做现在真的会更棒 不那样现在会怎样
都是你的想象
况且日子还很长 不要过早下判

输入乐谱 (ABC Notation)

无固定乐谱输入(由模型自主规划或无乐谱生成)

本机实际生成 / 转谱的完整乐谱

X:1
T:
M:4/4
L:1/16
Q:1/4=104
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:F
% intro
V: Vocal
z8z2dz4z|Z3|
V: Ins
z12f2z2|a2z2a2z2a2z2f2d2|a2a2a2a2z4c2e2|g2z2g2z2g2z2e2c2|
V: Vocal
Z4|
V: Ins
g2g2g2g2z4cdfz|a3zf4d3za3z|a3zf4d3za3z|e4z12|
V: Vocal
z8z2A2A2A2|
V: Ins
Z|
% verse
V: Vocal
D2F2F3D2zA2A2A2|D2F2F3D2zA2A2A2|E2E2E3C2zC2C2D2|F2G2G2F2A4AGA2|
V: Ins
Z4|
V: Vocal
A2A2D2D2F4AGA2|A2A2D2D2G4GFGF|G2G2G2GFG2GFG2A2|z12cAcA|
V: Ins
Z4|
V: Vocal
cAcAcAcAcAc2cAcA|cAcAcAcAcAc2cAcA|cAcAcAcAcAc2cAcA|cAcAcAcAcAd2d2dd|
V: Ins
Z4|
V: Vocal
ffddccAAGFA2G2GF|ffddccAAGFA2G2GF|GFGFGFGFGFA2D2GF|GFGFGFGFGFA2D2F2|
V: Ins
Z4|
% chorus
V: Vocal
a2a2d2d2a3d3d2|a2a2d2d2a3d3g2-|gfgfgfgfgfgfgfa2|d4z12|
V: Ins
Z4|
V: Vocal
a2a2d2d2a3d3d2|a2a2d2d2a3c'3e2-|e4zdc6z2cd|f2f2f2f2f2edcAcA|
V: Ins
Z4|
V: Vocal
d4z6cccAcA|d2c2cAcAd2c2A2AA|d2c2cAcAd2c2A2A2|D4z12|
V: Ins
Z3|z4a3zg3zf3z|
V: Vocal
z8z2cccAcA|d2c2cAcAd2c2A2AA|d2c2cAcAd2c2A2A2|D4z8z2AG|
V: Ins
a4f4z8|Z3|
% verse
V: Vocal
AGAGAGAGAGA2D2AG|AGAGAGAGAGA2D2GF|GFGFGFGFGFGFGFGF|GFGFGFGFA2D2F4|
V: Ins
Z4|
V: Vocal
z2cAcAcAcAcAcAcA-|AzcAcAcAc2cAcAcA-|Azd2c2d2f2g2a2d'2-|d'4c'4a3ggfgf|
V: Ins
Z4|
% chorus
V: Vocal
a2a2d2d2a2a2d2d2|a3d3g4a2gfgf|g2g2c2c2g2g2c2c2|g3c3f4z2cdfd|
V: Ins
Z4|
V: Vocal
f2f2fdfdf2f2a2dd|fdfdfdfdfdfda2z2|z2e2e2e2A2A4cA|cAcAcAcAcAcAcAcA|
V: Ins
Z4|
V: Vocal
a3gg3ff4d4|a12c'4|a2a2d2d2a3d3d2|a2a2d2d2a3d3g2-|
V: Ins
Z4|
V: Vocal
gfgfgfgfgfgfgfa2|d4z12|a2a2d2d2a3d3d2|a2a2d2d2a3c'3e2-|
V: Ins
Z4|
V: Vocal
e4zdc6z2cd|f2f2f2f2f2eAcAcA|d4z6cAcAcA|d2c2cAcAd2c2A2AA|
V: Ins
Z4|
V: Vocal
d2c2cAcAd2c2A2A2|D4z12|z8z2cAcAcA|d2c2cAcAd2c2A2AA|
V: Ins
Z|z4a4g4f4|a4f4z8|Z|
V: Vocal
d2c2cAcAd2c2A2A2|D4z12|
V: Ins
Z|z12cdfz|
% outro
V: Vocal
Z4|
V: Ins
a3zf4d3za3z|a3zf4d3za3z|e16|Z|

32步标准模式详细指标

{
  "id": "melody_generated",
  "status": "complete",
  "case_sha256": "ad2f51849d639bc72b087a37888ca059d97bb36b30677a54627ffd0c0ab224f1",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 181.03866666666667,
  "official_seconds": 152.72,
  "duration_ratio": 1.1854286712065654,
  "e2e_seconds": 417.94476708298316,
  "rtf": 2.3085939306684935,
  "speed_multiplier": 0.4331640946965639,
  "load_seconds": 5.222712083952501,
  "transcription_seconds": 0,
  "save_seconds": 0.46523162501398474,
  "timing": {
    "abc": {
      "seconds": 18.721492583979852,
      "prefill_seconds": 0.2737319999723695,
      "ttft_seconds": 0.277456916985102,
      "output_tokens": 1303,
      "content_tokens": 1302,
      "output_tps": 69.59915157165346,
      "prefix_tokens": 630,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "semantic": {
      "seconds": 62.798967041948345,
      "prefill_seconds": 1.0575671669794247,
      "ttft_seconds": 1.0596303339698352,
      "output_tokens": 4527,
      "content_tokens": 4526,
      "output_tps": 72.08717297811702,
      "prefix_tokens": 1934,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 320.9904702079948,
    "vae_seconds": 9.49240120803006,
    "load": {
      "resolve_conversion_integrity_seconds": 5.222618916013744,
      "ar_load_events_seconds": [
        0.11920879100216553
      ],
      "ar_load_seconds": 0.11920879100216553,
      "conditioning_load_events_seconds": [
        0.1726755000418052
      ],
      "conditioning_load_seconds": 0.1726755000418052,
      "nar_load_events_seconds": [
        0.11611875001108274
      ],
      "nar_load_seconds": 0.11611875001108274,
      "vae_load_events_seconds": [
        0.1061998339719139
      ],
      "vae_load_seconds": 0.1061998339719139
    },
    "e2e_seconds": 412.14095762500074
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "melody",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "a86563a9199a737d2d27d95035ba6bf3287a02dea2eedd934e1cc403916de2ed",
  "audio_sha256": "e6edc8d011da73b514f87bd67b8e4df19e4a0f6025bb09733531d255f2417212",
  "result_sha256": "5dbc633cba22fec09a778ccfa4c27c98048aa1a68d43b10ee7cfee621a51ea3e",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.17611676890217365,
    "peak": 0.8792818784713745,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.011745583033189178,
      "rss_bytes": 51822592,
      "system_swap_used_bytes": 12382502912,
      "system_swap_in_bytes": 56332173312,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 70473826304,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 13708705792,
      "system_wired_bytes": 6456360960,
      "system_compressor_bytes": 2759802880,
      "system_file_backed_bytes": 30930026496,
      "physical_footprint_bytes": 31965928,
      "process_lifetime_peak_footprint_bytes": 31965928,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 418.47346291702706,
      "rss_bytes": 1018609664,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332369920,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 69795315712,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 13033472000,
      "system_wired_bytes": 7206256640,
      "system_compressor_bytes": 2742812672,
      "system_file_backed_bytes": 29968367616,
      "physical_footprint_bytes": 1316309224,
      "process_lifetime_peak_footprint_bytes": 12091726472,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 267354112,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11845720712,
      "process_lifetime_peak_footprint_bytes": 12091726472,
      "rss_bytes": 8174305280,
      "system_available_bytes": 70474907648,
      "system_compressor_bytes": 2759802880,
      "system_file_backed_bytes": 31175032832,
      "system_free_bytes": 13708705792,
      "system_swap_in_bytes": 56332369920,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12382502912,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11553832960
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/melody_generated/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power disconnected during acceptance execution",
      "receipt_path": "outputs/official_fullsong/melody_generated/measurement.json",
      "receipt_sha256": "7adccedb0ac39e78279f699e1f16669084ed9be9e0a0755b373be4ec51edc2e0"
    }
  ]
}

8步快速模式详细指标

{
  "id": "melody_generated",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 181.04,
  "e2e_seconds": 144.32,
  "rtf": 0.797,
  "speed_multiplier": 1.25,
  "timing": {
    "abc": {
      "seconds": 18.721492583979852,
      "prefill_seconds": 0.2737319999723695,
      "ttft_seconds": 0.277456916985102,
      "output_tokens": 1303,
      "content_tokens": 1302,
      "output_tps": 69.59915157165346,
      "prefix_tokens": 630,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "semantic": {
      "seconds": 62.798967041948345,
      "prefill_seconds": 1.0575671669794247,
      "ttft_seconds": 1.0596303339698352,
      "output_tokens": 4527,
      "content_tokens": 4526,
      "output_tps": 72.08717297811702,
      "prefix_tokens": 1934,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 51.28,
    "vae_seconds": 6.29,
    "load_seconds": 5.22,
    "transcription_seconds": 0,
    "e2e_seconds": 144.32,
    "32step_nar_seconds": 320.99,
    "nar_speedup": 6.26
  },
  "audio_sha256": "957541d71161ab5a1a6f07e97ab76c09337df23eeb8f5d35ab37c5e9ea243366",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1757972193267956,
    "peak": 0.8747795224189758,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.011652000015601516,
      "rss_bytes": 8472887296,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62914707456,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1533313024,
      "system_wired_bytes": 8739274752,
      "system_compressor_bytes": 2730852352,
      "system_file_backed_bytes": 28524134400,
      "physical_footprint_bytes": 8796298272,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 159255064,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 57.5879404580337,
      "rss_bytes": 8475590656,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64540033024,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3131080704,
      "system_wired_bytes": 7827423232,
      "system_compressor_bytes": 2730835968,
      "system_file_backed_bytes": 28588277760,
      "physical_footprint_bytes": 8969215008,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140083472,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11207531582,
      "mlx_cache_bytes": 185466880,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12146187296,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8476983296,
      "system_available_bytes": 64820985856,
      "system_compressor_bytes": 2730852352,
      "system_file_backed_bytes": 28588277760,
      "system_free_bytes": 3410116608,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 10354900992
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
英语 · MELODY · 官方完整乐谱 8步快速模式提速 2.13x (RTF 0.680)

Jingle Bells · Minor-Key Version

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:jingle-bells-minor-no-chord

官方原版 Demo

Official Reference

时长: 107.4s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 107.0s | 耗时: 154.7s | RTF: 1.446 (0.69x 实时)
NAR声学求解: 104.4s (32步) | VAE: 4.0s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 107.0s | 耗时: 72.7s | RTF: 0.680 (1.47x 实时)
NAR声学求解: 23.5s (8步 · 4.44x 提速) | VAE: 3.6s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Christmas pop in E minor, sleigh bells, 104 BPM

歌词 (Lyrics)

[Verse]
Dashing through the snow
In a one-horse open sleigh
O'er the fields we go
Laughing all the way
Bells on bobtails ring
Making spirits bright
What fun it is to ride and sing
A sleighing song tonight

[Chorus]
Jingle bells, jingle bells
Jingle all the way
Oh, what fun it is to ride
In a one-horse open sleigh, hey!
Jingle bells, jingle bells
Jingle all the way
Oh, what fun it is to ride
In a one-horse open sleigh

[Verse]
Now the ground is white
Go it while you're young
Take the girls tonight
And sing this sleighing song
Just get a bobtailed bay
Two forty as his speed
Hitch him to an open sleigh
And snap! You'll take the lead

[Chorus]
Jingle bells, jingle bells
Jingle all the way
Oh, what fun it is to ride
In a one-horse open sleigh, hey!
Jingle bells, jingle bells
Jingle all the way
Oh, what fun it is to ride
In a one-horse open sleigh
Oh, what fun it is to ride
In a one-horse open sleigh

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/32
Q:1/4=104
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:Em
% intro
V: Vocal
Z4|
V: Ins
Z|g4g4g8g4g4g8|g4b4e6f2g16|a4a4a8a4g4g4g2g2|
V: Vocal
Z|
V: Ins
b4b4a4f4e16|
% verse
V: Vocal
B4g4f4e4B8z4B2B2|B4g4f4e4c8z8|c4a4g4f4d8z8|b4b4a4f4g8z8|
V: Ins
Z4|
V: Vocal
B4g4f4e4B8z8|B4g4f4e4c8z4c4|c4a4g4f4b4b4b4b4|c'4b4a4f4e4z4b8|
V: Ins
Z4|
% chorus
V: Vocal
g4g4g8g4g4g8|g4b4e6f2g8z8|a4a4a6a2a4g4g4g2g2|g4f4f4g4f8b4z4|
V: Ins
Z4|
V: Vocal
g4g4g8g4g4g8|g4b4e6f2g8z8|a4a4a6a2a4g4g4g2g2|b4b4a4f4e12z4|
V: Ins
Z4|
V: Vocal
Z|
V: Ins
Z|
% interlude
V: Vocal
Z|
V: Ins
B4B2B2e4B2B2e2B2e2g2b8|
% verse
V: Vocal
B4g4f4e4B8z8|B4g4f4e4c8z8|c4a4g4f4d8z8|b4b4a4f4g8z8|
V: Ins
Z4|
V: Vocal
B4g4f4e4B8z4B4|B4g4f4e4c8z4c4|c4a4g4f4b4b4b4b4|c'4b4a4f4e4z4b8|
V: Ins
Z4|
% chorus
V: Vocal
g4g4g8g4g4g8|g4b4e6f2g8z8|a4a4a6a2a4g4g4g2g2|g4f4f4g4f8b4z4|
V: Ins
Z4|
V: Vocal
g4g4g8g4g4g8|g4b4e6f2g8z8|a4a4a6a2a4g4g4g2g2|b4b4a4f4e8z8|
V: Ins
Z4|
V: Vocal
a4a4a6a2a4g4g4g2g2|b8b8a8f8|e24z8|
V: Ins
Z3|
V: Vocal
M:1/4
Z|
V: Ins
M:1/4
Z|
% outro
V: Vocal
M:4/4
Z4|
V: Ins
M:4/4
g4g4g8g4g4g8|g4b4e6f2g16|a4a4a4a4a4g4g4g2g2|b4b4a4f4e16|
V: Vocal
Z|
V: Ins
e4g4g4e4z16|

32步标准模式详细指标

{
  "id": "melody_supplied",
  "status": "complete",
  "case_sha256": "81ff5bce14179bf0a6da751244f9aec6edc6e34e061c0b39eda3139f75861272",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 106.95866666666667,
  "official_seconds": 107.35866666666666,
  "duration_ratio": 0.9962741713136031,
  "e2e_seconds": 154.71270404098323,
  "rtf": 1.446471883602855,
  "speed_multiplier": 0.6913373231349731,
  "load_seconds": 5.201949458976742,
  "transcription_seconds": 0,
  "save_seconds": 0.3180820000125095,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 944
    },
    "semantic": {
      "seconds": 40.39203779201489,
      "prefill_seconds": 0.8464277079910971,
      "ttft_seconds": 0.8498130000079982,
      "output_tokens": 2675,
      "content_tokens": 2674,
      "output_tps": 66.22592338059313,
      "prefix_tokens": 1269,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 104.41671999997925,
    "vae_seconds": 3.958839250029996,
    "load": {
      "resolve_conversion_integrity_seconds": 5.201887417002581,
      "ar_load_events_seconds": [
        0.11900458298623562
      ],
      "ar_load_seconds": 0.11900458298623562,
      "conditioning_load_events_seconds": [
        0.21916387503733858
      ],
      "conditioning_load_seconds": 0.21916387503733858,
      "nar_load_events_seconds": [
        0.13244458299595863
      ],
      "nar_load_seconds": 0.13244458299595863,
      "vae_load_events_seconds": [
        0.10607362497830763
      ],
      "vae_load_seconds": 0.10607362497830763
    },
    "e2e_seconds": 148.9062277909834
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "melody",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "9d0b167724db8db0f005e8cf98a353533fe70e277508e799adec3d1661f82a64",
  "audio_sha256": "16759c15b81f802f51b4a676ae5bed9b16bc4e9334976e5d70c289ee5c983020",
  "result_sha256": "8271706dc45c46af193dcb0e1e8e4a040f345279acc987beeb066654d59e2869",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14550125703815084,
    "peak": 1.0,
    "clipping_fraction": 1.9477929168900137e-07
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.009916167007759213,
      "rss_bytes": 51953664,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332369920,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 70761709568,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 14330363904,
      "system_wired_bytes": 6635929600,
      "system_compressor_bytes": 2742812672,
      "system_file_backed_bytes": 29971595264,
      "physical_footprint_bytes": 32097000,
      "process_lifetime_peak_footprint_bytes": 32097000,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 155.07863366702804,
      "rss_bytes": 938688512,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332566528,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65943896064,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 5339480064,
      "system_wired_bytes": 7147044864,
      "system_compressor_bytes": 2742632448,
      "system_file_backed_bytes": 30022123520,
      "physical_footprint_bytes": 1260112176,
      "process_lifetime_peak_footprint_bytes": 11869723320,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196732990,
      "mlx_cache_bytes": 164937760,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11737373368,
      "process_lifetime_peak_footprint_bytes": 11869723320,
      "rss_bytes": 8234958848,
      "system_available_bytes": 70762905600,
      "system_compressor_bytes": 2742812672,
      "system_file_backed_bytes": 30022123520,
      "system_free_bytes": 14346682368,
      "system_swap_in_bytes": 56332566528,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11826184192
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/melody_supplied/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/melody_supplied/measurement.json",
      "receipt_sha256": "185c873093a44167d02daee1965c139df7240a2d49bb4990cb49e1a121fe8a0e"
    }
  ]
}

8步快速模式详细指标

{
  "id": "melody_supplied",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 106.96,
  "e2e_seconds": 72.73,
  "rtf": 0.68,
  "speed_multiplier": 1.47,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 944
    },
    "semantic": {
      "seconds": 40.39203779201489,
      "prefill_seconds": 0.8464277079910971,
      "ttft_seconds": 0.8498130000079982,
      "output_tokens": 2675,
      "content_tokens": 2674,
      "output_tps": 66.22592338059313,
      "prefix_tokens": 1269,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 23.49,
    "vae_seconds": 3.64,
    "load_seconds": 5.2,
    "transcription_seconds": 0,
    "e2e_seconds": 72.73,
    "32step_nar_seconds": 104.42,
    "nar_speedup": 4.44
  },
  "audio_sha256": "b8e4f6acc72ac699d7847970dfb3df370173951dccbd72a9ab4e8cc0e07a4c62",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14561454643051722,
    "peak": 1.0,
    "clipping_fraction": 1.9477929168900137e-07
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.013278500002343208,
      "rss_bytes": 8478752768,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64540033024,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3131080704,
      "system_wired_bytes": 7827423232,
      "system_compressor_bytes": 2730835968,
      "system_file_backed_bytes": 28588277760,
      "physical_footprint_bytes": 8782470200,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140083472,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 27.14811449998524,
      "rss_bytes": 8491614208,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64158646272,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3320397824,
      "system_wired_bytes": 7932329984,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28630089728,
      "physical_footprint_bytes": 9204849720,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 163420712,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11207565606,
      "mlx_cache_bytes": 163420712,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 12016294968,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8491614208,
      "system_available_bytes": 64540033024,
      "system_compressor_bytes": 2730835968,
      "system_file_backed_bytes": 28630089728,
      "system_free_bytes": 3426598912,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11573723136
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
日语 · OFF · 无乐谱直接生成 8步快速模式提速 1.99x (RTF 1.126)

Enka

pass · 完成

逐字使用官方歌词、风格;官方和本地均无乐谱直接生成。 官方案例 ID:bdfc142ef136070839953ffa

官方原版 Demo

Official Reference

时长: 243.3s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 251.1s | 耗时: 562.7s | RTF: 2.241 (0.45x 实时)
NAR声学求解: 355.0s (32步) | VAE: 13.2s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 251.1s | 耗时: 282.7s | RTF: 1.126 (0.89x 实时)
NAR声学求解: 79.9s (8步 · 4.44x 提速) | VAE: 9.0s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

retro,city pop,中森明菜,Japanese

歌词 (Lyrics)

风が麦浪をなびかせ 稲の花が香る
黒い土地が われらの父母を育てる
毎年冬には 大雪が舞い散る
暖かい炬燵の上で 世間話をする

青い空に 白い雲がゆらゆら
清らかな小川が せせらぎながら流れる
東北の人は ザワーキャベツと血腸が好き
ヤンコ踊りを踊れば 人々は喜びそう

大東北は わたしの故郷だ
ソナが奏でる 素敵な姿
仲間が集まれば 必ずお酒を二杯飲む
月を酔わせ わたしの心を温める

大東北は わたしの故郷だ
わたしはこの地で 生まれ育った
東北人の情 東北人の愛
この故郷を 心から愛してる

风が麦浪をなびかせ 稲の花が香る
黒い土地が われらの父母を育てる
毎年冬には 大雪が舞い散る
暖かい炬燵の上で 世間話をする

青い空に 白い雲がゆらゆら
清らかな小川が せせらぎながら流れる
東北の人は ザワーキャベツと血腸が好き
ヤンコ踊りを踊れば 人々は喜びそう

大東北は わたしの故郷だ
ソナが奏でる 素敵な姿
仲間が集まれば 必ずお酒を二杯飲む
月を酔わせ わたしの心を温める

大東北は わたしの故郷だ
わたしはこの地で 生まれ育った
東北人の情 東北人の愛
この故郷を 心から愛してる

大東北は わたしの故郷だ
ソナが奏でる 素敵な姿
仲間が集まれば 必ずお酒を二杯飲む
月を酔わせ わたしの心を温める

大東北は わたしの故郷だ
わたしはこの地で 生まれ育った
東北人の情 東北人の愛
この故郷を 心から愛してる
この故郷を 心から愛してる

输入乐谱 (ABC Notation)

无固定乐谱输入(由模型自主规划或无乐谱生成)

32步标准模式详细指标

{
  "id": "off_direct",
  "status": "complete",
  "case_sha256": "088ab9b8b866dce7532e5ecc6ed6aec999e48e744ed1dbcf6340b8f68a66f97d",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 251.11866666666666,
  "official_seconds": 243.28,
  "duration_ratio": 1.0322207607146772,
  "e2e_seconds": 562.6601044579875,
  "rtf": 2.2406144151954224,
  "speed_multiplier": 0.44630615299901205,
  "load_seconds": 5.205297333013732,
  "transcription_seconds": 0,
  "save_seconds": 0.5658685420057736,
  "timing": {
    "abc": {},
    "semantic": {
      "seconds": 188.53203529200982,
      "prefill_seconds": 0.27864099998259917,
      "ttft_seconds": 0.2826585419825278,
      "output_tokens": 6279,
      "content_tokens": 6278,
      "output_tps": 33.30468474641302,
      "prefix_tokens": 604,
      "cfg_branches": 2,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 355.02819666598225,
    "vae_seconds": 13.176016915997025,
    "load": {
      "resolve_conversion_integrity_seconds": 5.205235541972797,
      "ar_load_events_seconds": [
        0.12122329202247784
      ],
      "ar_load_seconds": 0.12122329202247784,
      "conditioning_load_events_seconds": [
        0.17318866698769853
      ],
      "conditioning_load_seconds": 0.17318866698769853,
      "nar_load_events_seconds": [
        0.11944195796968415
      ],
      "nar_load_seconds": 0.11944195796968415,
      "vae_load_events_seconds": [
        0.09848862502258271
      ],
      "vae_load_seconds": 0.09848862502258271
    },
    "e2e_seconds": 556.8740691249841
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "off",
    "cfg_scale": 1.01,
    "cfg_negative": "instruction_only",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "004582111a9d836a10ccb2cc7e31bb54cfe515a1b5936539b6c4e4763455c269",
  "audio_sha256": "bc71fd78a2fb6cc3fb7ed40ff8933e623a50c944fb2a7a09e7b5df59391b1d53",
  "result_sha256": "8d2e8a74360591e3b3c468b48bddd6a4b6c8c7aa62c8d52ba02a3e705705536e",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14199334393344099,
    "peak": 0.9785663485527039,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01051108300453052,
      "rss_bytes": 52101120,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332566528,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65943896064,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 5339480064,
      "system_wired_bytes": 7147044864,
      "system_compressor_bytes": 2742632448,
      "system_file_backed_bytes": 30022123520,
      "physical_footprint_bytes": 32244456,
      "process_lifetime_peak_footprint_bytes": 32244456,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 563.3207151250099,
      "rss_bytes": 1106231296,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332828672,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64992411648,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 4838768640,
      "system_wired_bytes": 8458977280,
      "system_compressor_bytes": 2738896896,
      "system_file_backed_bytes": 29403021312,
      "physical_footprint_bytes": 1411549368,
      "process_lifetime_peak_footprint_bytes": 12590766728,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 180060168,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 12035414688,
      "process_lifetime_peak_footprint_bytes": 12590766728,
      "rss_bytes": 8207761408,
      "system_available_bytes": 70676021248,
      "system_compressor_bytes": 2742779904,
      "system_file_backed_bytes": 30050975744,
      "system_free_bytes": 13948567552,
      "system_swap_in_bytes": 56332828672,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 12477759488
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/off_direct/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/off_direct/measurement.json",
      "receipt_sha256": "72176fb435f5c6546a2315df5c1a98d8b51af2d2f7fbc8d77d30326c8987ef16"
    }
  ]
}

8步快速模式详细指标

{
  "id": "off_direct",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 251.12,
  "e2e_seconds": 282.7,
  "rtf": 1.126,
  "speed_multiplier": 0.89,
  "timing": {
    "abc": {},
    "semantic": {
      "seconds": 188.53203529200982,
      "prefill_seconds": 0.27864099998259917,
      "ttft_seconds": 0.2826585419825278,
      "output_tokens": 6279,
      "content_tokens": 6278,
      "output_tps": 33.30468474641302,
      "prefix_tokens": 604,
      "cfg_branches": 2,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 79.92,
    "vae_seconds": 9.04,
    "load_seconds": 5.21,
    "transcription_seconds": 0,
    "e2e_seconds": 282.7,
    "32step_nar_seconds": 355.03,
    "nar_speedup": 4.44
  },
  "audio_sha256": "37692cb77dcaac0bd096e072c124b37909e0ba45e73a6de252c808d90c0b0bf1",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1421124804535521,
    "peak": 0.9784667491912842,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012283875024877489,
      "rss_bytes": 8487583744,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64158646272,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3320397824,
      "system_wired_bytes": 7932329984,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28630089728,
      "physical_footprint_bytes": 8815172664,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 163420712,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 88.97905654204078,
      "rss_bytes": 8493056000,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 62953586688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1449410560,
      "system_wired_bytes": 9355214848,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28695805952,
      "physical_footprint_bytes": 9175669816,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 147456544,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 163420712,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11491580984,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "rss_bytes": 8493056000,
      "system_available_bytes": 64688013312,
      "system_compressor_bytes": 2730819584,
      "system_file_backed_bytes": 28698525696,
      "system_free_bytes": 3333636096,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 10533879808
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
中文 · FULL · 官方完整乐谱 8步快速模式提速 2.22x (RTF 0.846)

Big Dog Goes Woof

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:dagoujiao-chord

官方原版 Demo

Official Reference

时长: 210.0s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 207.9s | 耗时: 390.5s | RTF: 1.878 (0.53x 实时)
NAR声学求解: 288.8s (32步) | VAE: 9.5s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 207.9s | 耗时: 175.8s | RTF: 0.846 (1.18x 实时)
NAR声学求解: 76.0s (8步 · 3.8x 提速) | VAE: 8.4s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

New Orleans brass-funk, second-line drums, sousaphone, honky-tonk piano, trombone and baritone sax, playful Mandarin duet, 115 BPM

歌词 (Lyrics)

[Chorus]
我们一起学狗叫 一起汪汪汪汪汪
在你门前摇尾巴 快来汪汪汪汪汪
我的脚步咚咚响 追着飞盘满街跑
不带我出门我就汪汪汪

[Verse]
每天我都等你带我跑
守在门边数着每分每秒
外面今天多热闹 风儿比我更知道
我就是你的小保镖

[Verse]
有时候我馋得像头熊
听见塑料袋就满屋乱跑
你总说等等 可我已经冲到大门口
今天也要当只快乐狗

[Chorus]
我们一起学狗叫 一起汪汪汪汪汪
在你门前摇尾巴 快来汪汪汪汪汪
我的脚步咚咚响 追着飞盘满街跑
不带我出门我就汪汪汪
我们一起学狗叫 一起汪汪汪汪汪
我要叼你的外套 闻见出门的味道
想要追着你的球 跑到草地上睡着
每天都惦记那根骨头

[Verse]
黄昏后我跑得一身土
你拿毛巾追我满屋乱跑
你说慢一点 我却摇着尾巴装不懂
今晚还要当只香香狗

[Chorus]
我们一起学狗叫 一起汪汪汪汪汪
在你门前摇尾巴 快来汪汪汪汪汪
我的脚步咚咚响 追着飞盘满街跑
不带我出门我就汪汪汪
我们一起学狗叫 一起汪汪汪汪汪
我要叼你的外套 闻见出门的味道
想要追着你的球 跑到草地上睡着
每天都惦记那根骨头

[Chorus]
我们终于跑回家 一起汪汪汪汪汪
在你脚边摇尾巴 晚安汪汪汪汪汪
我的呼噜轰轰响 梦里飞盘满天跑
不许我上床我就汪汪汪
我们明天还要跑 一起汪汪汪汪汪
我要守着你外套 闻着安心的味道
想要靠着你的脚 一直睡到天亮了
每天都等着新的骨头

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/32
Q:1/4=115
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:F#
% silence
V: Vocal
Z|
V: Ins
Z|
% intro
V: Vocal
"F#"z32|"C#"z32|"D#m7"z32|"Bsus2"z24f4g4|
V: Ins
z12f2z6f4z4f4|z4g4z4g6z2g4z4g4|z4f4z4f6z2f4z4f4|z4f4z4f8z12|
% chorus
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"C#/E#"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#sus4"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4d4d4d4d4z4c4A4|"A#m7"c4A4"D#m7"c4g4f8z4c4|"G#m7"a4a4a4b4c'4f4f4a4|"C#sus4"g16z16|
V: Ins
Z4|
% verse
V: Vocal
"F#"z8c4c4c8c4c4|"C#/E#"d4c4A4c4z4F4F4F4|"D#m7"A8A4A4A4F4F4A4|"F#maj7/C#"A8z12A4G4F4|
V: Ins
Z4|
V: Vocal
"B"F4F2F6C4z4F4F4F4|"F#/A#"F4C4C4A8F4z4C4|"G#m7"B8B4B4B4F4F4G4|"C#sus4"G16z16|
V: Ins
Z4|
V: Vocal
"F#"z8c4c4c8c4c4|"A#m7"d4c4A4c4z4c4c4c4|"D#m7"f8f4f4A4A4d4c4-|"C#6"c12z8c4c4A4|
V: Ins
Z4|
V: Vocal
"B"d6d6c4z4d4c4B4|"A#m7"c4c4"D#m7"c4g8f4z4c4|"G#m7"a8a4b4c'4a4f4a4-|"C#sus4"a4a4g8-"C#"g16|
V: Ins
Z4|
V: Vocal
"C#"z24f4g4|
V: Ins
Z|
% chorus
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4d4d4d4c4z4c4A4|"A#m"c4A4c4g4f8z4c4|"G#m7"a4a4a4b4c'4f4f4a4|"C#sus4"g16z8f4g4|
V: Ins
Z4|
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"Bmaj7"d4F4B4d4d4z4c4A4|"A#m"c4A4c4g4f8z4c4|"G#m7"b8b4a4"C#sus4"b4a4f4g4-|"F#"g4f16z12|
V: Ins
Z4|
% interlude
V: Vocal
"Bmaj7"z16"C#"z16|"F#"z32|"B"z16"C#"z16|"D#m7"z32|
V: Ins
a4g4f4g8f4e4f4-|f16-f4z4a4b4|c'8f4g8c'8a4-|a16-a4z8d2e2|
V: Vocal
"Bmaj7"z8"C#sus4"z24|"A#7"z8"D#m"z24|"G#m7"z16"C#sus4"z16|"F#sus4"z8"F#"z24|
V: Ins
f8a4g12z4a2c'2|^^c'4a4e'4f'12d'4^c'4-|c'4z4a4g8c4g4f4-|f16f4g8a4-|
% verse
V: Vocal
"F#"z8c4c4c8c4c4|"A#m7"d4c4A4c4z4c4c4c4|"D#m7"f8f4f4A4A4d4c4-|"C#"c12z8c4c4A4|
V: Ins
a8z24|Z3|
V: Vocal
"B"d4d2d6c4z4d4c4B4|"A#m7"c4c4"D#m7"c4g8f4z4c4|"G#m7"a4a4a4b4c'4a4f4a4-|"C#sus4"a4a4g16-"C#"g8|
V: Ins
Z4|
V: Vocal
"C#"z24f4g4|
V: Ins
z4c'2z2c'8a4c'4z8|
% chorus
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4d4d4d4d4z4c4A4|"A#m7"c4A4c4g4f8z4c4|"G#m7"a4a4a4b4c'4f4f4a4|"C#"g16z8f4g4|
V: Ins
Z4|
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4F4B4d4d4z4c4A4|"A#m"c4A4c4g4f8z4c4|"G#m7"b8b4a4"C#"b4a4f4g4-|"F#"g4f16z4f4g4|
V: Ins
Z4|
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4d4d4d4d4z4c4A4|"A#m"c4A4c4g4f4f4z4c4|"G#m7"a4a4a4b4c'4f4f4a4|"C#"a4g16z4f4g4|
V: Ins
Z4|
V: Vocal
"F#"a4c4f4a4a4z4g4f4|"A#m7"g4c'4c'4c'4c'4z4f4e4|"D#m7"f4f4f4f4f4z4e4f4|"C#6"e4f4e4d4c4z4c4c4|
V: Ins
Z4|
V: Vocal
"B"d4F4B4d4d4z4c4A4|"A#m"c4A4c4g4f4f4z4c4|"G#m7"b8b4a4"C#sus4"b4c'4f4g4|"F#"z4g4f16z8|
V: Ins
Z4|
% outro
V: Vocal
"F#"z32|"C#/E#"z32|"D#m7"z32|"Bsus2"z32|
V: Ins
a16z8a4g4-|g16-g4z4g4f4-|f16-f4z12|b8c'2z2c'16z4|
V: Vocal
"F#"z32|"F#"z32|
V: Ins
f'32|Z|

32步标准模式详细指标

{
  "id": "cover_dagoujiao",
  "status": "complete",
  "case_sha256": "6202abf61195ceb62850331159d65f1f79d39f9fca6c6ec514ced0a24babe31d",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 207.91866666666667,
  "official_seconds": 209.95866666666666,
  "duration_ratio": 0.9902838018911659,
  "e2e_seconds": 390.4508596250089,
  "rtf": 1.877901902146074,
  "speed_multiplier": 0.5325091789178102,
  "load_seconds": 5.236991917015985,
  "transcription_seconds": 0,
  "save_seconds": 0.46092441596556455,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 2398
    },
    "semantic": {
      "seconds": 86.1781507920241,
      "prefill_seconds": 1.704754917009268,
      "ttft_seconds": 1.7083948340150528,
      "output_tokens": 5199,
      "content_tokens": 5198,
      "output_tps": 60.32851659287605,
      "prefix_tokens": 2931,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 288.79281679098494,
    "vae_seconds": 9.522870707965922,
    "load": {
      "resolve_conversion_integrity_seconds": 5.236917499976698,
      "ar_load_events_seconds": [
        0.12041183398105204
      ],
      "ar_load_seconds": 0.12041183398105204,
      "conditioning_load_events_seconds": [
        0.19284049997804686
      ],
      "conditioning_load_seconds": 0.19284049997804686,
      "nar_load_events_seconds": [
        0.1329428749741055
      ],
      "nar_load_seconds": 0.1329428749741055,
      "vae_load_events_seconds": [
        0.1017137499875389
      ],
      "vae_load_seconds": 0.1017137499875389
    },
    "e2e_seconds": 384.63164837495424
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "1dfc2e74cf3d94c387b37ac8ed9971bd82efb12d37b6371cee329f6adea975c1",
  "audio_sha256": "ec9f59e58b06d8d91148c843af26fbc52db6c330c31e7c4cb7dbdcd10ba626ba",
  "result_sha256": "ca548e00759f8b74711a1516904e4776206829f1bafd06cb70e3ff340187b8b0",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.17102370424286245,
    "peak": 0.950627326965332,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01035045797470957,
      "rss_bytes": 52068352,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56332828672,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 70936838144,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 14734950400,
      "system_wired_bytes": 6459129856,
      "system_compressor_bytes": 2738896896,
      "system_file_backed_bytes": 29460611072,
      "physical_footprint_bytes": 32195280,
      "process_lifetime_peak_footprint_bytes": 32195280,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 390.99074724997627,
      "rss_bytes": 817184768,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333025280,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65638514688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 5577211904,
      "system_wired_bytes": 7952646144,
      "system_compressor_bytes": 2737373184,
      "system_file_backed_bytes": 29556326400,
      "physical_footprint_bytes": 975341680,
      "process_lifetime_peak_footprint_bytes": 12128754288,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 197705728,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11872508504,
      "process_lifetime_peak_footprint_bytes": 12128754288,
      "rss_bytes": 8184348672,
      "system_available_bytes": 70969376768,
      "system_compressor_bytes": 2738896896,
      "system_file_backed_bytes": 29571416064,
      "system_free_bytes": 14819852288,
      "system_swap_in_bytes": 56333025280,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11388895232
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/cover_dagoujiao/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/cover_dagoujiao/measurement.json",
      "receipt_sha256": "1c1d677994e2f85fa04dfc5ce5ebc3c58498deb5ecaa3faa3349c2423c6eb34a"
    }
  ]
}

8步快速模式详细指标

{
  "id": "cover_dagoujiao",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 207.92,
  "e2e_seconds": 175.84,
  "rtf": 0.846,
  "speed_multiplier": 1.18,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 2398
    },
    "semantic": {
      "seconds": 86.1781507920241,
      "prefill_seconds": 1.704754917009268,
      "ttft_seconds": 1.7083948340150528,
      "output_tokens": 5199,
      "content_tokens": 5198,
      "output_tps": 60.32851659287605,
      "prefix_tokens": 2931,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 75.97,
    "vae_seconds": 8.45,
    "load_seconds": 5.24,
    "transcription_seconds": 0,
    "e2e_seconds": 175.84,
    "32step_nar_seconds": 288.79,
    "nar_speedup": 3.8
  },
  "audio_sha256": "8685327bf758264132c5cdb230ae60a3bef7a24cffbe5cfb6cf591526cbd01f1",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.17103013179160736,
    "peak": 0.9562925100326538,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012668750016018748,
      "rss_bytes": 8692039680,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65228832768,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3737501696,
      "system_wired_bytes": 7104692224,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28713369600,
      "physical_footprint_bytes": 9003506888,
      "process_lifetime_peak_footprint_bytes": 12427844616,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 147456544,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 84.43346091703279,
      "rss_bytes": 8693448704,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64595066880,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3075407872,
      "system_wired_bytes": 7891238912,
      "system_compressor_bytes": 2730737664,
      "system_file_backed_bytes": 28774940672,
      "physical_footprint_bytes": 9281264816,
      "process_lifetime_peak_footprint_bytes": 12432284848,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 144386072,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11207531582,
      "mlx_cache_bytes": 170393600,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 12309355696,
      "process_lifetime_peak_footprint_bytes": 12432284848,
      "rss_bytes": 8696315904,
      "system_available_bytes": 65228832768,
      "system_compressor_bytes": 2730803200,
      "system_file_backed_bytes": 28774940672,
      "system_free_bytes": 3737501696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 10832625664
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
中文 · FULL · 官方完整乐谱 8步快速模式提速 2.2x (RTF 0.798)

Lake Baikal · Chinese Jazz Version

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:baikal-lake-jazz

官方原版 Demo

Official Reference

时长: 216.9s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 212.3s | 耗时: 372.7s | RTF: 1.756 (0.57x 实时)
NAR声学求解: 274.8s (32步) | VAE: 10.7s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 212.3s | 耗时: 169.4s | RTF: 0.798 (1.25x 实时)
NAR声学求解: 74.8s (8步 · 3.67x 提速) | VAE: 8.2s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

Jazz ballad, brushed drums, upright bass, warm piano, muted trumpet, intimate Mandarin male vocal, 60 BPM

歌词 (Lyrics)

[Verse]
在我的怀里
在你的眼里
那里春风沉醉
那里绿草如茵
月光把爱恋
洒满了湖面
两个人的篝火
照亮整个夜晚

[Chorus]
多少年以后
如云般游走
那变换的脚步
让我们难牵手
这一生一世
有多少你我
被吞没在月光
如水的夜里

[Post-Chorus]
多想某一天
往日又重现
我们流连忘返
在贝加尔湖畔

[Chorus]
多少年以后
往事随云走
那纷飞的冰雪
容不下那温柔
这一生一世
这时间太少
不够证明融化
冰雪的深情

[Post-Chorus]
就在某一天
你忽然出现
你清澈又神秘
在贝加尔湖畔
你清澈又神秘
像贝加尔湖畔

输入乐谱 (ABC Notation)

X:1
T:
M:2/4
L:1/16
Q:1/4=60
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:F
% intro
V: Vocal
Z2|"Gm"z8|"C"z8|
V: Ins
Z|z2def2c'2|b8|z2cde2b2|
V: Vocal
"Fmaj7"z8|"D"z8|"Gm"z8|"A7"z8|
V: Ins
a8|z2A2d2c2|B3AG4|G2E2A2G2|
V: Vocal
"Dm"z8|
V: Ins
d8|
% verse
V: Vocal
"Dm"z2DEF2c2|
V: Ins
Z|
V: Vocal
"Gm"B4z4|"C"z2CDE2B2|"Fmaj7"A4z4|"D"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G2z3F|"C"E2FG3B2|"Fmaj7"A4z4|"Dm"z2DEF2c2|
V: Ins
Z4|
V: Vocal
"Gm"B4z4|"C"z2CDE2B2|"Fmaj7"A4z4|"D"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G2z3F|"A7"E2AG3FE|"Dm"D2DDD2d2|
V: Ins
Z3|
% chorus
V: Vocal
"Bb"d6z2|"Bb"z2Ddc2dc|"Fmaj7"A4z4|"Fmaj7"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G2z3F|"C"E2FG3B2|"Fmaj7"A6z2|"Fmaj7"z2DDD2d2|
V: Ins
Z4|
V: Vocal
"Bb"d8|"Bb"z2deggf2|"Fmaj7"A6z2|"Fmaj7"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G4z2|"C"c2de3e2|"Fmaj7"a8|"A"e8|
V: Ins
Z4|
V: Vocal
"A"z8|"Dm"z2DEF2c2|
V: Ins
Z2|
% post-chorus
V: Vocal
"Gm"B4z4|"C"z2CDE2B2|"Fmaj7"A4z4|"D"z2AAd2cB-|
V: Ins
Z4|
V: Vocal
"Gm"BGG2z3E|"A"E2AG3F2|"Dm"D4z4|"Dm"z8|
V: Ins
Z2|z4fdAF|D2DEF2c2|
% interlude
V: Vocal
"Gm"z8|"C"z8|"Fmaj7"z8|"D"z8|
V: Ins
B8|z2cde2b2|a8|z2AAB2a^g|
V: Vocal
"Gm"z8|"C"z8|"Fmaj7"z8|"A"z8|
V: Ins
g6f2|e2fg3c'=b|a8|z2dfad'2^c'|
V: Vocal
"Gm"z8|"C"z8|"Fmaj7"z8|"D"z8|
V: Ins
d'2DGBd2^c|d3d'e'4|c'8|z2d^fad'2c'|
V: Vocal
"Gm"z8|"A7"z8|"Dm"z8|
V: Ins
b6a2|g2e2a2g2|d'd'^c'd'a'f'd'a|
% chorus
V: Vocal
"Dm"z2DD"Dm7/C"D2d2|
V: Ins
f^cz6|
V: Vocal
"Bb"d6z2|"Bb"z2Ddc2e2-|"Fmaj7"e2A4z2|"Fmaj7"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G2z3F|"C"E2FG3B2|"Fmaj7"A6z2|"Fmaj7"z2DDD2d2|
V: Ins
Z4|
V: Vocal
"Bb"d6z2|"Bb"z2deggf2|"Fmaj7"A6z2|"Fmaj7"z2AAd2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G4z2|"C"c2de3e2|"Fmaj7"a6gf|"A7"e8|
V: Ins
Z4|
V: Vocal
"A7"z8|"A7"z2DEF2c2|
V: Ins
Z2|
% post-chorus
V: Vocal
"Gm"B4z4|"C"z2CDE2B2|"Fmaj7"A4z4|"Dm7"z2AA"D7"d2c2|
V: Ins
Z4|
V: Vocal
"Gm"B2G2z3E|"A7"E2AG3E2|"Dm"D4z4|"Dm"z2AAd2c2|
V: Ins
Z2|z4fdAF|D2z6|
V: Vocal
"Gm"B2G2z3E|"A7"EE2GG4|"A7"z4FF3-|"Dm"FDD4z2|
V: Ins
Z4|
% outro
V: Vocal
"Dm"z8|"Dm"z8|"Dm"z4"D"z4|
V: Ins
Z3|
V: Vocal
"D"z4"Dsus2"z4|"Dsus2"z4"Dm"z4|"Dm"z8|"Dm"z8|
V: Ins
Z4|
V: Vocal
"Dm"z8|"Dm"z8|"Dm"z8|
V: Ins
Z3|

32步标准模式详细指标

{
  "id": "cover_baikal",
  "status": "complete",
  "case_sha256": "5e16e6ddaf10453f9c76aa556f8cd495b10f656dafebba79fbd24da51885aef4",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 212.27866666666668,
  "official_seconds": 216.91866666666667,
  "duration_ratio": 0.978609494188298,
  "e2e_seconds": 372.6630877080024,
  "rtf": 1.7555371604683265,
  "speed_multiplier": 0.5696262218301485,
  "load_seconds": 5.270500749989878,
  "transcription_seconds": 0,
  "save_seconds": 0.46892104198923334,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1455
    },
    "semantic": {
      "seconds": 81.14225450000959,
      "prefill_seconds": 0.984601583040785,
      "ttft_seconds": 0.9880465830210596,
      "output_tokens": 5308,
      "content_tokens": 5307,
      "output_tps": 65.4159787980672,
      "prefix_tokens": 1702,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 274.79719120799564,
    "vae_seconds": 10.727858000027481,
    "load": {
      "resolve_conversion_integrity_seconds": 5.270408125012182,
      "ar_load_events_seconds": [
        0.11972937500104308
      ],
      "ar_load_seconds": 0.11972937500104308,
      "conditioning_load_events_seconds": [
        0.16321333398809657
      ],
      "conditioning_load_seconds": 0.16321333398809657,
      "nar_load_events_seconds": [
        0.13335404195822775
      ],
      "nar_load_seconds": 0.13335404195822775,
      "vae_load_events_seconds": [
        0.10133187501924112
      ],
      "vae_load_seconds": 0.10133187501924112
    },
    "e2e_seconds": 366.80124545795843
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "e7b3263aec42b7863d934c3a09e582a76c5d7404071d068fb83206758e9ff3f3",
  "audio_sha256": "4d1dc12e8de200850fc5d7feae9b189bd02b9ee8584f8bd73f18ab29015422d7",
  "result_sha256": "2df9efe0b7a1dac21fea66831b611e54d484af4de0c940a13bc5010951c5fad1",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14606263065109223,
    "peak": 1.0,
    "clipping_fraction": 9.814143672782317e-08
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.010612874990329146,
      "rss_bytes": 51970048,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333025280,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 71012450304,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 14900953088,
      "system_wired_bytes": 6432079872,
      "system_compressor_bytes": 2737373184,
      "system_file_backed_bytes": 29603938304,
      "physical_footprint_bytes": 32129792,
      "process_lifetime_peak_footprint_bytes": 32129792,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 373.2065921250032,
      "rss_bytes": 817364992,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 69842173952,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 13065633792,
      "system_wired_bytes": 7090126848,
      "system_compressor_bytes": 2734948352,
      "system_file_backed_bytes": 29652516864,
      "physical_footprint_bytes": 974997760,
      "process_lifetime_peak_footprint_bytes": 12148841216,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 197296128,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11975809792,
      "process_lifetime_peak_footprint_bytes": 12148841216,
      "rss_bytes": 8187363328,
      "system_available_bytes": 71012450304,
      "system_compressor_bytes": 2737537024,
      "system_file_backed_bytes": 29652516864,
      "system_free_bytes": 14900953088,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11743920128
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/cover_baikal/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/cover_baikal/measurement.json",
      "receipt_sha256": "ca348f4fd8706a3364bb84fa36f426316bf933acff192e50b9c3fa35d01598c2"
    }
  ]
}

8步快速模式详细指标

{
  "id": "cover_baikal",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 212.28,
  "e2e_seconds": 169.42,
  "rtf": 0.798,
  "speed_multiplier": 1.25,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1455
    },
    "semantic": {
      "seconds": 81.14225450000959,
      "prefill_seconds": 0.984601583040785,
      "ttft_seconds": 0.9880465830210596,
      "output_tokens": 5308,
      "content_tokens": 5307,
      "output_tps": 65.4159787980672,
      "prefix_tokens": 1702,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 74.84,
    "vae_seconds": 8.16,
    "load_seconds": 5.27,
    "transcription_seconds": 0,
    "e2e_seconds": 169.42,
    "32step_nar_seconds": 274.8,
    "nar_speedup": 3.67
  },
  "audio_sha256": "8bf2e931175d88024e1ceb2d70b0b09696797435686554629d3a089dcb27b962",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.14606089197787198,
    "peak": 0.9878390431404114,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01329295797040686,
      "rss_bytes": 8696872960,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64595066880,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3075407872,
      "system_wired_bytes": 7891238912,
      "system_compressor_bytes": 2730737664,
      "system_file_backed_bytes": 28774940672,
      "physical_footprint_bytes": 9005767880,
      "process_lifetime_peak_footprint_bytes": 12432284848,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 144386072,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 83.01670549996197,
      "rss_bytes": 8699707392,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63285575680,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1811611648,
      "system_wired_bytes": 9141256192,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28466151424,
      "physical_footprint_bytes": 9013959880,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 149668112,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718142,
      "mlx_cache_bytes": 200541464,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 12385033416,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "rss_bytes": 8701263872,
      "system_available_bytes": 64851853312,
      "system_compressor_bytes": 2730737664,
      "system_file_backed_bytes": 28843098112,
      "system_free_bytes": 3362914304,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 9719758848
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
英语 · FULL · 官方完整乐谱 8步快速模式提速 2.18x (RTF 0.722)

Happy Birthday · Heavy-Metal Scream Version

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:happy-birthday-heavy-metal

官方原版 Demo

Official Reference

时长: 63.9s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 64.8s | 耗时: 102.2s | RTF: 1.576 (0.63x 实时)
NAR声学求解: 68.0s (32步) | VAE: 3.4s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 64.8s | 耗时: 46.8s | RTF: 0.722 (1.39x 实时)
NAR声学求解: 14.2s (8步 · 4.77x 提速) | VAE: 2.4s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

heavy metal, harsh screamed vocals, distorted guitars, double-kick drums, 160 BPM

歌词 (Lyrics)

[Chorus]
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY, DEAR FRIEND!!!
HAPPY BIRTHDAY TO YOU!!!

[Chorus]
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY, DEAR FRIEND!!!
HAPPY BIRTHDAY TO YOU!!!

[Chorus]
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY, DEAR FRIEND!!!
HAPPY BIRTHDAY TO YOU!!!

[Chorus]
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY, DEAR FRIEND!!!
HAPPY BIRTHDAY TO YOU!!!

[Chorus]
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY TO YOU!!!
HAPPY BIRTHDAY, DEAR FRIEND!!!
HAPPY BIRTHDAY TO YOU!!!

输入乐谱 (ABC Notation)

X:1
T:
M:3/4
L:1/32
Q:1/4=160
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:F
% silence
V: Vocal
Z3|
V: Ins
Z3|
% intro
V: Vocal
Z|"F"z24|"Bb"z24|"F/C"z16"C7"z8|
V: Ins
z16c4z4|c'8a8f8|e8d8b6b2|a8f8g8|
% chorus
V: Vocal
"F"z16c6c2|"F"d8c8f8|"C"e12z4c6c2|"C"d8c8g8|
V: Ins
f16z8|Z|Z|Z|
V: Vocal
"F"f12z4c6c2|"F"c'8a8f8|"Bb"e8d4z4b6b2|"F/C"a8f8"C7"g8|
V: Ins
Z4|
V: Vocal
"F"f8z8c6c2|
V: Ins
Z|
% chorus
V: Vocal
"F"d8c8f8|"C"e12z4c6c2|"C"d8c8g8|"F"f8z8c6c2|
V: Ins
Z4|
V: Vocal
"F"c'8a8f6f2|"Bb"e8d8b6b2|"F/C"a8f8"C7"g8|"F"f12z12|
V: Ins
Z|Z|Z|z16c4z2c2|
% interlude
V: Vocal
"F"z24|"C"z24|"C"z24|"F"z24|
V: Ins
d8c8f8|e16c4z2c2|d8c8g8|f16c4z2c2|
V: Vocal
"F"z24|"Bb"z24|"F/C"z16"C7"z8|
V: Ins
c'8a8f6f2|e8d8b6b2|a8f8g8|
% chorus
V: Vocal
"F"z16c6c2|"F"d8c8f8|"C"e12z4c6c2|"C"d8c8g8|
V: Ins
f16z8|Z|Z|Z|
V: Vocal
"F"f12z4c6c2|"F"c'8a8f8|"Bb"e8d4z4b6b2|"F/C"a8f8"C7"g8|
V: Ins
Z4|
V: Vocal
"F"f12z4c6c2|
V: Ins
Z|
% chorus
V: Vocal
"F"d8c8f8|"C"e8z8c6c2|"C"d8c8g8|"F"f8z8c6c2|
V: Ins
Z4|
V: Vocal
"F"c'8a8f6f2|"Bb"e8d8b6b2|"F/C"a8f8"C7"g8|"F"f12z4c6c2|
V: Ins
Z4|
% chorus
V: Vocal
"F"d8c8f8|"C"e8z8c6c2|"C"d8c8g8|"F"f12z4c6c2|
V: Ins
Z4|
V: Vocal
"F"c'8a8f8|"Bb"e8d4z4b6b2|"F/C"a8f8"C7"g8|"F"f24|
V: Ins
Z4|
% outro
V: Vocal
"F"z24|"F"z24|
V: Ins
Z2|

32步标准模式详细指标

{
  "id": "cover_birthday",
  "status": "complete",
  "case_sha256": "c6714902951c35a1221ef82166622d5e8063394694b46c6f50be8ed289bbc8e3",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 64.83866666666667,
  "official_seconds": 63.91866666666667,
  "duration_ratio": 1.0143932914745823,
  "e2e_seconds": 102.2056244580308,
  "rtf": 1.5763066964881676,
  "speed_multiplier": 0.634394310591896,
  "load_seconds": 5.280088083003648,
  "transcription_seconds": 0,
  "save_seconds": 0.22794137499295175,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 916
    },
    "semantic": {
      "seconds": 24.856704458012246,
      "prefill_seconds": 0.775047708011698,
      "ttft_seconds": 0.7778397499932908,
      "output_tokens": 1622,
      "content_tokens": 1621,
      "output_tps": 65.25402443191413,
      "prefix_tokens": 1178,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 67.96025833301246,
    "vae_seconds": 3.3562813330208883,
    "load": {
      "resolve_conversion_integrity_seconds": 5.280012957984582,
      "ar_load_events_seconds": [
        0.12474133301293477
      ],
      "ar_load_seconds": 0.12474133301293477,
      "conditioning_load_events_seconds": [
        0.21434220799710602
      ],
      "conditioning_load_seconds": 0.21434220799710602,
      "nar_load_events_seconds": [
        0.13605595898116007
      ],
      "nar_load_seconds": 0.13605595898116007,
      "vae_load_events_seconds": [
        0.09607491601491347
      ],
      "vae_load_seconds": 0.09607491601491347
    },
    "e2e_seconds": 96.31568462500582
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "29e8c2912f2135b57e5ca764903185e4ec122fb9858ac11173546e5898422058",
  "audio_sha256": "f5f5adf90d99f7a0a5486c30a703701c8008967dea18c8a6272861ac6f4b30f6",
  "result_sha256": "b290d050db7402ea10d0c7a160eea07b411aa37d0d36bdb8fa3f6f55c8b222a3",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.13923169661928095,
    "peak": 0.8563051223754883,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01033629197627306,
      "rss_bytes": 52019200,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 69842173952,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 13065633792,
      "system_wired_bytes": 7090126848,
      "system_compressor_bytes": 2734948352,
      "system_file_backed_bytes": 29652516864,
      "physical_footprint_bytes": 32146128,
      "process_lifetime_peak_footprint_bytes": 32146128,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 102.4645569169661,
      "rss_bytes": 884162560,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64413810688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 5084184576,
      "system_wired_bytes": 7852408832,
      "system_compressor_bytes": 2734784512,
      "system_file_backed_bytes": 29612359680,
      "physical_footprint_bytes": 1182746976,
      "process_lifetime_peak_footprint_bytes": 12073638632,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256153824,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196720200,
      "mlx_cache_bytes": 200541464,
      "mlx_peak_bytes": 11256153824,
      "physical_footprint_bytes": 11964685032,
      "process_lifetime_peak_footprint_bytes": 12073638632,
      "rss_bytes": 8132837376,
      "system_available_bytes": 70813024256,
      "system_compressor_bytes": 2734948352,
      "system_file_backed_bytes": 29655482368,
      "system_free_bytes": 14387118080,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11302895616
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/cover_birthday/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/cover_birthday/measurement.json",
      "receipt_sha256": "cb694d43c7ff60cd97b0a6b7bad326e2ccce28cbfff221ad8402f9a0ee1bf02a"
    }
  ]
}

8步快速模式详细指标

{
  "id": "cover_birthday",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 64.84,
  "e2e_seconds": 46.81,
  "rtf": 0.722,
  "speed_multiplier": 1.39,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 916
    },
    "semantic": {
      "seconds": 24.856704458012246,
      "prefill_seconds": 0.775047708011698,
      "ttft_seconds": 0.7778397499932908,
      "output_tokens": 1622,
      "content_tokens": 1621,
      "output_tps": 65.25402443191413,
      "prefix_tokens": 1178,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 14.24,
    "vae_seconds": 2.44,
    "load_seconds": 5.28,
    "transcription_seconds": 0,
    "e2e_seconds": 46.81,
    "32step_nar_seconds": 67.96,
    "nar_speedup": 4.77
  },
  "audio_sha256": "3d90709acca67112e636e4779bb95de272125df159f7c4a43f356383f9330f05",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.139536247135639,
    "peak": 0.868676483631134,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012288208992686123,
      "rss_bytes": 8702312448,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65363558400,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3860873216,
      "system_wired_bytes": 7103397888,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28499951616,
      "physical_footprint_bytes": 9016106184,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 149668112,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 16.69007233402226,
      "rss_bytes": 8704065536,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 61250945024,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 901758976,
      "system_wired_bytes": 9850748928,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28502966272,
      "physical_footprint_bytes": 9457327304,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140296248,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11130494006,
      "mlx_cache_bytes": 151318348,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11975023816,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "rss_bytes": 8706768896,
      "system_available_bytes": 65363558400,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28502966272,
      "system_free_bytes": 3860873216,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 12031361024
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
日语 · FULL · 录音转谱 → 生成 8步快速模式提速 2.19x (RTF 0.900)

Easy Listening

pass · 完成

输入为官网的原始乐谱录音,通过本机 MLX SheetSage2 完整转谱后生成翻唱。 官方案例 ID:1b91eef30046ca2f0cf5633e

官方原版 Demo

Official Reference

时长: 221.8s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 224.4s | 耗时: 442.2s | RTF: 1.971 (0.51x 实时)
NAR声学求解: 319.3s (32步) | VAE: 10.0s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 224.4s | 耗时: 201.9s | RTF: 0.900 (1.11x 实时)
NAR声学求解: 81.3s (8步 · 3.93x 提速) | VAE: 8.3s

实际转谱输入:官方乐谱录音

转谱输入来源

完整转谱记录与诊断
{
  "melody_notes": 482,
  "vocal_notes": 0,
  "instrumental_notes": 482,
  "events": 666,
  "abc_measures": 83,
  "abc_error": null,
  "diagnostics": [
    "measure 50: inferred 1/4 from downbeat span; declared numerators were [4]",
    "measure 59: inferred 1/4 from downbeat span; declared numerators were [4]",
    "measure 60: inferred 3/4 from downbeat span; declared numerators were [4, 4, 4]",
    "measure 72: inferred 3/4 from downbeat span; declared numerators were [4, 4, 4]",
    "measure 82: padded final 2/4 span to declared 4/4 with trailing rest"
  ],
  "backend": "mlx",
  "dtype": "float32",
  "model": {
    "model": "m-a-p/SheetSage2",
    "revision": "eab522a8168e8b8b8c4856bf8609cd86198f01fe",
    "model_sha256": "b235f68091a5f5b644000f2b5acb57d1e70432aca2b34ab1b9cf27236e1f4274",
    "config_sha256": "a986e63f5d831ecb823c11d19cfb371d763f25ae2e614ec9acd714c0b8bb87fd",
    "base_model": "m-a-p/MERT-v2-FullSong",
    "base_revision": "d8ba1c745e733b3908ce6ad16ebeb17ac7600a42",
    "base_sha256": "e6dd2ab187d6dd62b6521cd7d8f932e237acf0c5757745a7232082e28391350d",
    "backend": "mlx",
    "dtype": "float32",
    "lora_merge": "numpy_cpu_fp32",
    "implementation_sha256": {
      "transcription/model.py": "993c38cb38f0bd9bdf1c3c1ce4c513b5fe0ef025214a11d9599d7cc61bd08ee2",
      "transcription/pipeline.py": "804482c050bcab86d083d2ce8d54195236d194f8f6ba984f44b1789860480539",
      "transcription/control.py": "1143e58cfedc3bd5469083d34b600094478e0fcbab585a418564cc2471886719",
      "mert.py": "c7ad0c37fd0d90ce249dcd3c11c6301cc84c3a51c899eb8adc6d1f2291d2dfc8",
      "bart.py": "a3de735ac0760a19f450ef4fbba9f7104a293acf1985633c642fd202c3d93848"
    }
  },
  "source_audio_sha256": "3172a644320a11511d36c541ea2115be388720a3a8693aee57b65334653947f8",
  "duration_seconds": 225.79579166666667,
  "task": "full",
  "preset": "default",
  "audio_decode": "ffmpeg_float32_mono_24000",
  "warnings": [],
  "truncated": false,
  "elapsed_seconds": 13.900297292042524,
  "status": "complete"
}
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

A gentle and melancholic piece beginning with a delicate piano melody and a breathy, expressive flute line. A soft, clear female vocal enters, singing in a gentle, almost lullaby-like style. The arrangement gradually builds, introducing a subtle, downtempo lo-fi hip-hop beat with a soft kick and snare, complemented by a warm bassline and atmospheric synth pads. The track maintains a tranquil and introspective mood throughout, with the piano and flute weaving in and out, creating a serene soundscape perfect for relaxation or contemplation. The song concludes with a beautiful instrumental outro featuring the piano and flute, fading into silence.

歌词 (Lyrics)

[Intro: Piano & Flute]

[Verse 1]
静かな森 歩み寄る
木々のざわめき 聞こえる音
陽射しが揺れて 影踊る
心は軽く 自由になる

[Verse 2]
風が通り 抱きしめるように
葉は止ま 語る古い物語
小鳥の歌が 空を仰ぎ
その音色だけ 生きている

[Chorus]
森の中で 消えてゆく
小さな悩み 風に乗る
心地良い緑 包まれて
私は癒され 生まれ変わる

[Instrumental Break: Piano, Flute, Drums]

[Verse 3]
土の匂いが 深く染みる
地球の鼓動 感じている
目を閉じ夜明けを 夢見て
自然と溶ける 虚無する

[Verse 4]
月の光が ゆっくりと
静けさ抱え 手ら知ってる
心がささやく 小さな声
森のリズムで 調和化する

[Chorus]
森の中で 消えてゆく
小さな悩み 風に乗る
心地良い緑 包まれて
私は癒され 生まれ変わる

[Outro: Piano & Flute]

输入乐谱 (ABC Notation)

无固定乐谱输入(由模型自主规划或无乐谱生成)

本机实际生成 / 转谱的完整乐谱

X:1
T:
M:4/4
L:1/16
Q:1/4=85
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:D#m
% intro
V: Vocal
z8"Bmaj7"z8|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
z12z3a|a3c'3g6f2a2|g3c'3e8z2|e3f3c8Ac|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
e3f3c4f2e2c2|a3c'3g6f2a2|g3c'3e4f2e2c2|e3f3c8Ac|
V: Vocal
"D#m7"z16|
V: Ins
e3f3c4d2f2g2|
% verse
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a2dd3z4d2f2a2|g3c3z4c2d2e2|e2fe2c2cc2AA3GA-|A2z8d2f2g2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a3d3cd3d2f2a2|g3c3z4c2d2e2|e2fe2c2cc2AA3GA-|A2z8d2f2g2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a3d3cd3d2f2a2|g3c3z4c2d2e2|e2fe2c2cc2A2A2G2|G3c3A4d2f2g2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a3d3cd3d2f2a2|g3c3e8c2|d3e3f2g3f3e2|e3f3e4d2d4|
V: Vocal
"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|
V: Ins
z8z2d2a2g2|a3c'3g4a2g2f2|e3c4z3d2a2g2|a3c'3g4a2g2f2|
V: Vocal
"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|
V: Ins
g3a4z3d2a2g2|a3c'3g4a2g2f2|g3c'3a8de|f2f2f2a2g3f3e2|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
e3f3e4d2d4|z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8z2|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
Z|z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8-A2-|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
A6z4d2f2g2|a2dd2c3z2d2f2a2|g3c3e4c2d2e2|e2fe2c3c2AA3GA-|
V: Vocal
"D#m7"z16|"Bmaj7"z16|"C#"z16|"D#m7"z16|
V: Ins
A2z8d2f2g2|a3d3cd3d2f2a2|g3c4z4zd2e2|e3f3e4d4c2|
V: Vocal
"D#m7"z16|
V: Ins
e3f3e4d4z2|
V: Vocal
M:1/4
"D#m7"z4|
V: Ins
M:1/4
Z|
V: Vocal
M:4/4
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
M:4/4
z2aaa2gg2ff4z2|z2a2a2gg2c'3z4|z2aaa2gg2ff4e2|e3f3e4d4z2|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
z2aaa2gggff4z2|z2a2a2gg2c'c'2z4|z2aaa2gggff4e2|e3f3e4d2d4|
V: Vocal
M:1/4
"D#m7"z4|
V: Ins
M:1/4
Z|
V: Vocal
M:3/4
"D#m7"z12|
V: Ins
M:3/4
z8z2d2|
V: Vocal
M:4/4
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|
V: Ins
M:4/4
a3c'3g4a2g2f2|g3c4z3d2a2g2|a3c'3g4a2g2f2|g3a4z3d2a2g2|
V: Vocal
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|
V: Ins
a3c'3g4a2g2f2|g3c'3a8de|f2f2f2a2g3f3e2|e3f3e4d2d4|
V: Vocal
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|
V: Ins
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8z2|
V: Vocal
M:3/4
"D#m7"z12|
V: Ins
M:3/4
Z|
V: Vocal
M:4/4
"Bmaj7"z8"C#"z8|"D#m7"z16|"Bmaj7"z8"C#"z8|"D#m7"z16|
V: Ins
M:4/4
z2d'c'c'aaggffeeddc|cddccAAGGFFEEDEC|D3A3A8z2|Z|
V: Vocal
"Bmaj7"z16|"C#"z16|"D#m7"z16|"D#m7"z16|
V: Ins
a3c'3g6f2a2|g3c'3e8z2|e3f3c8Ac|e3f3c8z2|
V: Vocal
"D#m7"z16|Z|
V: Ins
Z2|

32步标准模式详细指标

{
  "id": "audio_to_cover",
  "status": "complete",
  "case_sha256": "340e01884cb6d1e6b446a91904a6e38967ec6ccdffa3f1f70a3e42d556e7bf9f",
  "attempt": 2,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 224.35866666666666,
  "official_seconds": 221.76,
  "duration_ratio": 1.0117183742183742,
  "e2e_seconds": 442.2056347919861,
  "rtf": 1.970976389554758,
  "speed_multiplier": 0.5073627493964549,
  "load_seconds": 5.256098542013206,
  "transcription_seconds": 15.441064957994968,
  "save_seconds": 0.4870239580050111,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1988
    },
    "semantic": {
      "seconds": 91.65822037495673,
      "prefill_seconds": 1.516166332992725,
      "ttft_seconds": 1.5191173749626614,
      "output_tokens": 5610,
      "content_tokens": 5609,
      "output_tps": 61.20563957112121,
      "prefix_tokens": 2450,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 319.2624067500001,
    "vae_seconds": 9.963404207956046,
    "load": {
      "resolve_conversion_integrity_seconds": 5.256009291042574,
      "ar_load_events_seconds": [
        0.1156300829607062
      ],
      "ar_load_seconds": 0.1156300829607062,
      "conditioning_load_events_seconds": [
        0.19259333302034065
      ],
      "conditioning_load_seconds": 0.19259333302034065,
      "nar_load_events_seconds": [
        0.134724291972816
      ],
      "nar_load_seconds": 0.134724291972816,
      "vae_load_events_seconds": [
        0.10129804094322026
      ],
      "vae_load_seconds": 0.10129804094322026
    },
    "e2e_seconds": 421.00786125002196
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "df69a5697e65d8617297ca13345ed57255c90e66d3e3f1608cc5b73161f03ce5",
  "audio_sha256": "bf9e9bce7f7ec16e85d0f9bcd805ab743442fc8c07581bb6370fbb046801fccc",
  "result_sha256": "ab1ad8a6d6ac756e3e169bd1cf7341a660244e91d9993256162daa1554522254",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.17828478451710916,
    "peak": 1.0,
    "clipping_fraction": 4.178577159191533e-07
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.009731750003993511,
      "rss_bytes": 51937280,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56333713408,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64413810688,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 5084184576,
      "system_wired_bytes": 7852408832,
      "system_compressor_bytes": 2734784512,
      "system_file_backed_bytes": 29612359680,
      "physical_footprint_bytes": 32080616,
      "process_lifetime_peak_footprint_bytes": 32080616,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 442.765825749957,
      "rss_bytes": 1139507200,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56335728640,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65310294016,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3876765696,
      "system_wired_bytes": 7238205440,
      "system_compressor_bytes": 2733244416,
      "system_file_backed_bytes": 31557681152,
      "physical_footprint_bytes": 1138821736,
      "process_lifetime_peak_footprint_bytes": 12167158840,
      "mlx_active_bytes": 20,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150242,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196718154,
      "mlx_cache_bytes": 244744192,
      "mlx_peak_bytes": 11256150242,
      "physical_footprint_bytes": 12009036856,
      "process_lifetime_peak_footprint_bytes": 12167158840,
      "rss_bytes": 8494088192,
      "system_available_bytes": 70310297600,
      "system_compressor_bytes": 2734784512,
      "system_file_backed_bytes": 32501121024,
      "system_free_bytes": 13867597824,
      "system_swap_in_bytes": 56335728640,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11352506368
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/audio_to_cover/attempt-2",
  "prior_attempts": [
    {
      "attempt": 1,
      "status": "failed",
      "error": "RuntimeError: AC power is required for an acceptance benchmark",
      "receipt_path": "outputs/official_fullsong/audio_to_cover/measurement.json",
      "receipt_sha256": "2c6a35f1a27cf8803212d3bf2a414f2911523022bfc63c86fb57a7897139bc73"
    }
  ]
}

8步快速模式详细指标

{
  "id": "audio_to_cover",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 224.36,
  "e2e_seconds": 201.94,
  "rtf": 0.9,
  "speed_multiplier": 1.11,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1988
    },
    "semantic": {
      "seconds": 91.65822037495673,
      "prefill_seconds": 1.516166332992725,
      "ttft_seconds": 1.5191173749626614,
      "output_tokens": 5610,
      "content_tokens": 5609,
      "output_tps": 61.20563957112121,
      "prefix_tokens": 2450,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 81.33,
    "vae_seconds": 8.26,
    "load_seconds": 5.26,
    "transcription_seconds": 15.44,
    "e2e_seconds": 201.94,
    "32step_nar_seconds": 319.26,
    "nar_speedup": 3.93
  },
  "audio_sha256": "1838f8280cba5add75cc8766ec029be68df3a72693f926808e5879cd0df01daa",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.1781342164851589,
    "peak": 1.0,
    "clipping_fraction": 3.2500044571489696e-07
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.01287879201117903,
      "rss_bytes": 8762310656,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 61250945024,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 901758976,
      "system_wired_bytes": 9850748928,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28502966272,
      "physical_footprint_bytes": 9515605240,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 140296248,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 89.59652404201915,
      "rss_bytes": 8765128704,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 64170033152,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2712895488,
      "system_wired_bytes": 8099119104,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28576841728,
      "physical_footprint_bytes": 9911999736,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 244744192,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11207531582,
      "mlx_cache_bytes": 244744192,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11811986680,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "rss_bytes": 8765603840,
      "system_available_bytes": 64800129024,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28576940032,
      "system_free_bytes": 3307438080,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 10101604352
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记
日语 · FULL · 官方完整乐谱 8步快速模式提速 2.14x (RTF 0.722)

Jazz-Funk

pass · 完成

逐字使用官方歌词、风格和完整 ABC 乐谱。 官方案例 ID:e3d76e66a92637ebb46e61ed

官方原版 Demo

Official Reference

时长: 130.0s | 官网音频源 ↗
Apple Silicon mlx-Yue (32步 标准模式)

32-Step Standard Mode

时长: 130.4s | 耗时: 201.3s | RTF: 1.543 (0.65x 实时)
NAR声学求解: 140.2s (32步) | VAE: 5.5s
Apple Silicon mlx-Yue (8步 快速模式 ⚡)

8-Step Fast Mode (⚡ Real-Time)

时长: 130.4s | 耗时: 94.2s | RTF: 0.722 (1.39x 实时)
NAR声学求解: 33.9s (8步 · 4.13x 提速) | VAE: 5.4s
展开查看完整风格、歌词、乐谱与实测指标

风格提示词 (Tags)

funk

歌词 (Lyrics)

[Verse]
姉が玉子焼き用の四角いフライパンで
目玉焼きを焼いている
合う蓋はないので
横から湯気は洩れるが
面白い発想に見えた

僕は丸い小さめの普通のフライパンで
目玉焼きを焼く

[Pre-Chorus]
かように遠くで暮らしていると
癖やメソッドや好みが違う

[Chorus]
当たり前、当たり前、当たり前
これはAが優れてBが劣るというような
問題ではないのだ

[Bridge]
君の一本道の進歩観を
多々野辺のどこへともない道に
考え直してみてはどうだろう

输入乐谱 (ABC Notation)

X:1
T:
M:4/4
L:1/32
Q:1/4=116
V: Vocal clef=treble name="Vocal Melody" snm="Vocal"
V: Ins clef=treble name="Ins Melody" snm="Inst."
K:Em
% intro
V: Vocal
z24z4"Em"z4|"Em"z32|"Em"z32|"Em"z32|
V: Ins
z24z4E4|E8G6A6^A4B4d2e2|E8G6A6^A4B4d2e2|E8G6A6^A4B4d2e2|
V: Vocal
"Em"z32|"Em"z16"G6"z16|"Am7"z16"C"z8"B7"z8|"Em"z16"G6"z16|
V: Ins
e4d4^A2=A2G4E4D4E8|E8z4D4G6G2z8|A6A2z8c4c4B2B6|E8z4D4G6G2z8|
V: Vocal
"Am7"z16"C"z8"B7"z8|"Em"z16"G6"z16|"Am7"z16"C"z8"B7"z8|"Em"z16"G6"z16|
V: Ins
A6A2z8c4c4B2B6|e8e2d2B2d4e12z2|A6A2z8c4c4B2B6|E8e2d2B2d4e12z2|
V: Vocal
"Am7"z16"C"z8"B7"z8|
V: Ins
A6A2z8c4c4B2B6|
% verse
V: Vocal
"Em"e4d4B4A4"G6"A4G4E4D4|"Am7"E4D4E4D4"C"E4G4-"B7"G4E4|"Em"z4e4d4B4"G6"A4G4E4A4-|"Am7"A8G2G2B4-"C"B8"B7"z8|
V: Ins
Z4|
V: Vocal
"Em"e8B2B2B2A2-"G6"A2G2G8E2D2|"Am7"E2D2E2D2E4D2E2-"C"E2E2G4-"B7"G4z4|"Em"E4D4E4D4"G6"E4z4G8|"Am7"A4G4A4G4"C"B8"B7"z8|
V: Ins
Z4|
% pre-chorus
V: Vocal
"Cmaj7"z8e8g8a8|"B7"z8e8g8a8|"Em7"b8a4g4a8g4a4-|"A7"a8g4e12z4B4|
V: Ins
Z4|
V: Vocal
"Cmaj7"c8d8e8g8|"B7"f16^d16|"Em"e8z24|"Em"z32|
V: Ins
Z2|z8G6A6^A4B4d2e2|E8G6A6^A4B4d2e2|
% chorus
V: Vocal
"Cmaj7"b16a4g4g4a4-|"B7"a8g4b12z8|"Em7"a16a4g4g4a4-|"A7"a8g4e8z8d4|
V: Ins
Z4|
V: Vocal
"Cmaj7"e8g8f8g8|"B7"f8^d8B8A8|"Em7"B16d16|"Em7"e16g16|
V: Ins
Z4|
V: Vocal
"Cmaj7"z4b4a2g2g2a6b4a2g2g2a2-|"B7"a4b4a2g2g2a6b4a4g4|"Em7"a4g4a4g4a4g4a4g4|"A7"a2g2a2g2a2g2b8e8z4|
V: Ins
Z4|
V: Vocal
"Cmaj7"e16g16|"B7"f8^d8B8A8|"Em7"B16d16|"Em7"e16g16|
V: Ins
Z4|
V: Vocal
"Cmaj7"z4e4e4e4e8b8|"B7"a4g4a4g4a2g2a8b4|"Em7"z4e4e4e4e4e4b8|"A7"a2g2a2g2a4g2a4g2b8z4|
V: Ins
Z4|
V: Vocal
"Cmaj7"e16g16|"B7"f16^d16|"Em7"e16d16|"Em7"B8A8G8A6G2|
V: Ins
Z4|
V: Vocal
"Em"E16z16|
V: Ins
z16e4e4e4z4|
% outro
V: Vocal
"Em"z16"G6"z16|"A7"z16"C"z8"B7"z8|"Em"z16"G6"z16|"A7"z16"C"z8"B7"z8|
V: Ins
e8e2d2B2d4e6z8|A6A2z8c4c4B8|E8e2d2B2d4e6z8|A6A2z8c4c4B8|
V: Vocal
"Em"z16"G6"z16|"A7"z16"C"z8"B7"z8|"Em"z16"G6"z16|"A7"z16"C"z8"B7"z8|
V: Ins
E8e2d2B2d4e6z8|A6A2z8c4c4B8|E8e2d2B2d4e6z8|A6A2z8c4c4B8|
V: Vocal
"Em"z32|"Em"z32|
V: Ins
E8G6A6^A4B4d2e2|e8z24|

32步标准模式详细指标

{
  "id": "jazz_funk_score",
  "status": "complete",
  "case_sha256": "d40f44a1b959680ae81011c7207830384050d24c9ec83495892201c941c5f796",
  "attempt": 1,
  "benchmark_script_sha256": "2c5868ef28d74bb713079a0c1cf886d419a4e3a73cce34be2413e4cda7ec9449",
  "retry_policy": "Only interrupted/failed execution, identical input and seed; no quality reroll",
  "truncated": {
    "abc": false,
    "semantic": false
  },
  "audio_seconds": 130.43866666666668,
  "official_seconds": 129.96,
  "duration_ratio": 1.0036831845696113,
  "e2e_seconds": 201.32383604202187,
  "rtf": 1.5434367828712998,
  "speed_multiplier": 0.6479047351325081,
  "load_seconds": 5.2746187500306405,
  "transcription_seconds": 0,
  "save_seconds": 0.3200644159805961,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1549
    },
    "semantic": {
      "seconds": 49.57729250000557,
      "prefill_seconds": 0.9960394580266438,
      "ttft_seconds": 1.0004081249935552,
      "output_tokens": 3262,
      "content_tokens": 3261,
      "output_tps": 65.79625137858494,
      "prefix_tokens": 1738,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 140.20773391699186,
    "vae_seconds": 5.536092333029956,
    "load": {
      "resolve_conversion_integrity_seconds": 5.274542875005864,
      "ar_load_events_seconds": [
        0.12086416594684124
      ],
      "ar_load_seconds": 0.12086416594684124,
      "conditioning_load_events_seconds": [
        0.18636520800646394
      ],
      "conditioning_load_seconds": 0.18636520800646394,
      "nar_load_events_seconds": [
        0.13901312503730878
      ],
      "nar_load_seconds": 0.13901312503730878,
      "vae_load_events_seconds": [
        0.10589120798977092
      ],
      "vae_load_seconds": 0.10589120798977092
    },
    "e2e_seconds": 195.45785833301488
  },
  "configuration": {
    "generation": {
      "abc": {
        "temperature": 0.7,
        "top_p": 0.9,
        "top_k": 30,
        "repetition_penalty": 1.005,
        "penalty_window": 100,
        "min_tokens": 32,
        "max_tokens": 4096
      },
      "semantic": {
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 100,
        "repetition_penalty": 1.2,
        "penalty_window": 50,
        "min_tokens": 200,
        "max_tokens": 9000
      },
      "ode_steps": 32,
      "ode_method": "midpoint",
      "context": 24576,
      "version": "yue2-native-v1"
    },
    "overrides": {},
    "cot": "full",
    "cfg_scale": 1.0,
    "cfg_negative": "same_instruction_and_exact_abc",
    "backend": "mlx",
    "quantization": "8bit",
    "model_dtype": "bfloat16",
    "vae_dtype": "float32",
    "vae_decode": "halo_crop",
    "vae_core_frames": 256,
    "vae_halo_frames": 16,
    "device": "DeviceType.gpu",
    "memory_budget_gib": 16.0,
    "offload_ar": false,
    "runtime_sha256": "ff4829d01380a2ac98fbc0aa0aed90f9b2bbe9c0a3473f1d62ed8ee3abc8efa9",
    "decoder_release": "standard",
    "validation_status": "unvalidated",
    "vae_backend": "mlx",
    "ar_precision": "8bit",
    "kv_dtype": "bfloat16",
    "nar_dtype": "bfloat16",
    "conditioning_precision": "bf16",
    "attention_opmath": "float32",
    "mlx_tf32_enabled": false,
    "attention_inputs": "native_steel_bfloat16",
    "query_chunk_size": 256,
    "runtime": {
      "lyra-yue2": "0.1.0",
      "mlx": "0.32.2",
      "mlx-lm": "0.31.3",
      "transformers": "5.0.0",
      "numpy": "2.2.6",
      "python": "3.12.10",
      "macos": "27.0"
    },
    "upstream_commit": "92a73cc7652fcc1f937855e4b765e0a0edd7ff2e",
    "execution_guard": "GPUExecution",
    "rng": {
      "ar": "request_local_mlx",
      "acoustic": "numpy_pcg64_fp32_full_song_v1"
    }
  },
  "identity": "3f218561eac2134fff51f3d807ff4e4683b4d88ca4eff5f99842f5ba6a4e8b60",
  "audio_sha256": "0ee90a8fbd7398b8fcd36af5975d5d98d9af45864efc72c9720bfbe4b6bac5fb",
  "result_sha256": "cb8afe79b78df5c0b02a74454295b5435126cf9b5a087a13688052bdc46da8ef",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.16763944066473796,
    "peak": 0.9980323910713196,
    "clipping_fraction": 0.0
  },
  "musical_quality_acceptance": "pending human listening; not established by duration or EOS",
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.013879666046705097,
      "rss_bytes": 52789248,
      "system_swap_used_bytes": 12374114304,
      "system_swap_in_bytes": 56335728640,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65310294016,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3876765696,
      "system_wired_bytes": 7238205440,
      "system_compressor_bytes": 2733244416,
      "system_file_backed_bytes": 31557681152,
      "physical_footprint_bytes": 32932584,
      "process_lifetime_peak_footprint_bytes": 32932584,
      "mlx_active_bytes": 0,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 0,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 201.7050155000179,
      "rss_bytes": 956547072,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56340054016,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63562334208,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 2313764864,
      "system_wired_bytes": 8933703680,
      "system_compressor_bytes": 2734473216,
      "system_file_backed_bytes": 31645499392,
      "physical_footprint_bytes": 1265600840,
      "process_lifetime_peak_footprint_bytes": 11908111032,
      "mlx_active_bytes": 8,
      "mlx_cache_bytes": 0,
      "mlx_peak_bytes": 11256150230,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 170590586,
      "mlx_peak_bytes": 11256150230,
      "physical_footprint_bytes": 11763456696,
      "process_lifetime_peak_footprint_bytes": 11908111032,
      "rss_bytes": 8251392000,
      "system_available_bytes": 70759759872,
      "system_compressor_bytes": 2734473216,
      "system_file_backed_bytes": 31645515776,
      "system_free_bytes": 12762710016,
      "system_swap_in_bytes": 56340054016,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12374114304,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 11263492096
    },
    "monitor_error": null,
    "exception": null
  },
  "artifact_directory": "outputs/official_fullsong/jazz_funk_score",
  "prior_attempts": []
}

8步快速模式详细指标

{
  "id": "jazz_funk_score",
  "mode": "fast_8step",
  "ode_steps": 8,
  "status": "complete",
  "audio_seconds": 130.44,
  "e2e_seconds": 94.16,
  "rtf": 0.722,
  "speed_multiplier": 1.39,
  "timing": {
    "abc": {
      "seconds": 0.0,
      "output_tokens": 0,
      "external_prefix_tokens": 1549
    },
    "semantic": {
      "seconds": 49.57729250000557,
      "prefill_seconds": 0.9960394580266438,
      "ttft_seconds": 1.0004081249935552,
      "output_tokens": 3262,
      "content_tokens": 3261,
      "output_tps": 65.79625137858494,
      "prefix_tokens": 1738,
      "cfg_branches": 1,
      "execution": "eager",
      "attention": "sdpa",
      "backend": "mlx"
    },
    "nar_seconds": 33.94,
    "vae_seconds": 5.37,
    "load_seconds": 5.27,
    "transcription_seconds": 0,
    "e2e_seconds": 94.16,
    "32step_nar_seconds": 140.21,
    "nar_speedup": 4.13
  },
  "audio_sha256": "c804544a94ba626b3373c0b04554a10dda82e995b43d136008e6fcacb130cea2",
  "audio_quality": {
    "sample_rate": 48000,
    "channels": 2,
    "finite": true,
    "rms": 0.16770223406866447,
    "peak": 0.9914978742599487,
    "clipping_fraction": 0.0
  },
  "resources": {
    "memory_counter_schema": 2,
    "swap_counter_source": "host_statistics64.swapins/swapouts",
    "sampling_interval_seconds": 0.1,
    "power_sampling_interval_seconds": 5.0,
    "metadata": {},
    "power_start": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "power_end": {
      "ac_connected": true,
      "pmset": "Now drawing from 'AC Power'\n -InternalBattery-0 (id=37224547)\t95%; AC attached; not charging present: true"
    },
    "start": {
      "elapsed_seconds": 0.012712624971754849,
      "rss_bytes": 8761589760,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 65159938048,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 3660283904,
      "system_wired_bytes": 7197179904,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28619997184,
      "physical_footprint_bytes": 9170033912,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 244744192,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "end": {
      "elapsed_seconds": 39.32432491600048,
      "rss_bytes": 8864759808,
      "system_swap_used_bytes": 12365725696,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_available_bytes": 63568969728,
      "system_total_bytes": 137438953472,
      "system_memory_pressure_level": 1,
      "system_free_bytes": 1938259968,
      "system_wired_bytes": 8683094016,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28772139008,
      "physical_footprint_bytes": 9686850880,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "mlx_active_bytes": 7526793954,
      "mlx_cache_bytes": 151142672,
      "mlx_peak_bytes": 11256166614,
      "ac_connected": true
    },
    "sampled_maxima": {
      "mlx_active_bytes": 11196716606,
      "mlx_cache_bytes": 244744192,
      "mlx_peak_bytes": 11256166614,
      "physical_footprint_bytes": 11642248440,
      "process_lifetime_peak_footprint_bytes": 12437249224,
      "rss_bytes": 8864759808,
      "system_available_bytes": 65160003584,
      "system_compressor_bytes": 2730721280,
      "system_file_backed_bytes": 28772253696,
      "system_free_bytes": 3660283904,
      "system_swap_in_bytes": 56345296896,
      "system_swap_out_bytes": 83892617216,
      "system_swap_used_bytes": 12365725696,
      "system_total_bytes": 137438953472,
      "system_wired_bytes": 10546626560
    },
    "monitor_error": null,
    "exception": null
  }
}

下载该案例全量输入与实测元数据 (JSON)

记录试听笔记