[15:54:40] === Mac voice-extract: idol=winter, urls=3 === [15:54:40] root: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data [15:54:40] device: mps [15:54:40] [1/7] fetch [15:54:40] fetch [skip] de6FhGIFoxA [15:54:40] fetch [skip] OrFTzGP57MQ [15:54:40] fetch [skip] yOIZ53fviVk [15:54:40] [2/7] vocals (demucs-mlx) [15:54:40] vocals [skip] de6FhGIFoxA [15:54:40] vocals [skip] OrFTzGP57MQ [15:54:40] vocals [skip] yOIZ53fviVk [15:54:40] [3/7] diarize (pyannote 3.1 / MPS) [15:54:46] diarize [skip] de6FhGIFoxA [15:54:46] diarize [skip] OrFTzGP57MQ [15:54:46] diarize [skip] yOIZ53fviVk [15:54:46] [4/7] select [15:54:46] select auto_dominant: SPEAKER_01 (total 1627.0s) [15:54:46] SPEAKER_01: 1627.0s [15:54:46] SPEAKER_00: 122.9s [15:54:46] SPEAKER_02: 45.8s [15:54:46] SPEAKER_03: 25.3s [15:54:46] [5/8] match (WeSpeaker ResNet34 / MPS) [15:54:46] match [load WeSpeaker] on mps ... [15:54:46] loaded (0.7s) [15:54:47] match [centroid] from de6FhGIFoxA SPEAKER_01 ... [15:55:05] centroid: 648 segments [15:55:12] match [de6FhGIFoxA] 497/1430 segments ≥ 0.55 (6.7s) [15:55:21] match [OrFTzGP57MQ] 343/1100 segments ≥ 0.55 (8.4s) [15:55:24] match [yOIZ53fviVk] 194/548 segments ≥ 0.55 (3.3s) [15:55:24] matched total: 1034 [15:55:24] [6/8] stitch (인접 matched → 10-15s clips) [15:55:26] stitch [de6FhGIFoxA]: 24 clips (10.0-15.0s) [15:55:26] stitch [OrFTzGP57MQ]: 12 clips (10.0-15.0s) [15:55:26] stitch [yOIZ53fviVk]: 8 clips (10.0-15.0s) [15:55:26] stitched total: 44 [15:55:26] [7/8] filter (VAD + SNR) [15:55:26] filter [silero-vad] loading ... [15:55:28] filter kept: 38 / 44 stitched clips [15:55:28] rejected: {'no_voice': 0, 'voice_dur': 0, 'voice_ratio': 0, 'snr': 6} [15:55:28] [8/8] concat + transcribe [15:55:28] prompt prompt.wav: 13.42s (de6FhGIFoxA_04609420_13415_0.794.wav, cos=0.794) [15:55:28] transcribe [docksvc whisper] prompt.wav ... [15:55:29] done (1.3s) — 87 chars [15:55:29] "학교 다닐 때 봉사 시간을 채워야 했단 말이에요 그 무슨 무조건 채워야 되는 모두가 이제 모두가 봉사 시간 10시간인가 무조건 그게 수행평가에요 수행평가였어" [15:55:29] === done === [15:55:29] dataset_dir: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/filtered [15:55:29] prompt.wav: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/prompt.wav [15:55:29] prompt.txt: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/prompt.txt [15:55:29] manifest: /Users/inhohur/ai/experiments/voice-extract-mac/winter/data/manifest.json