[15:47:54] === Mac voice-extract: idol=winter, urls=3 === [15:47:54] root: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data [15:47:54] device: mps [15:47:54] [1/7] fetch [15:47:54] fetch [skip] de6FhGIFoxA [15:47:54] fetch [skip] OrFTzGP57MQ [15:47:54] fetch [skip] yOIZ53fviVk [15:47:54] [2/7] vocals (demucs-mlx) [15:47:54] vocals [skip] de6FhGIFoxA [15:47:54] vocals [skip] OrFTzGP57MQ [15:47:54] vocals [skip] yOIZ53fviVk [15:47:54] [3/7] diarize (pyannote 3.1 / MPS) [15:47:58] diarize [skip] de6FhGIFoxA [15:47:58] diarize [skip] OrFTzGP57MQ [15:47:58] diarize [skip] yOIZ53fviVk [15:47:58] [4/7] select [15:47:58] select auto_dominant: SPEAKER_01 (total 1627.0s) [15:47:58] SPEAKER_01: 1627.0s [15:47:58] SPEAKER_00: 122.9s [15:47:58] SPEAKER_02: 45.8s [15:47:58] SPEAKER_03: 25.3s [15:47:58] [5/7] match (WeSpeaker ResNet34 / MPS) [15:47:58] match [load WeSpeaker] on mps ... [15:47:59] loaded (1.4s) [15:48:00] match [centroid] from de6FhGIFoxA SPEAKER_01 ... [15:48:45] centroid: 648 segments [15:49:28] match [de6FhGIFoxA] 497/1430 segments ≥ 0.55 (41.5s) [15:50:06] match [OrFTzGP57MQ] 343/1100 segments ≥ 0.55 (37.4s) [15:50:24] match [yOIZ53fviVk] 194/548 segments ≥ 0.55 (18.0s) [15:50:24] matched total: 1034 [15:50:24] [6/7] filter [15:50:24] filter [silero-vad] loading ... [15:50:30] filter kept: 298 segments (2.0s ≤ dur ≤ 15.0s) [15:50:30] rejected: {'too_short': 482, 'too_long': 0, 'no_voice': 1, 'voice_dur': 6, 'voice_ratio': 1, 'snr': 246} [15:50:30] [7/7] concat + transcribe [15:50:30] concat prompt.wav: 12.00s (3 chunks) [15:50:30] transcribe [docksvc whisper] prompt.wav ... [15:50:31] done (1.6s) — 85 chars [15:50:31] "와 학교 다닐 때 빼빼로 교환했었는데 막 친구들이랑 공연할 때는 긴 머리 좋아하고 근데 나는 개인적으로 나는 나는 지금 손톱도 많이 자라서 나 네일 하러" [15:50:31] === done === [15:50:31] dataset_dir: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/filtered [15:50:31] prompt.wav: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/prompt.wav [15:50:31] prompt.txt: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/prompt.txt [15:50:31] manifest: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/manifest.json