- Fix voice_profiler threshold bug (HOST label overwrote Unknown unconditionally) - Audio preload optimization: single ffmpeg per episode, 149.5x realtime on 5070 Ti - WavLM threshold raised to 0.85 (Mike 0.90-0.99, callers 0.46-0.83) - Promo/bumper filter: weighted signature scoring, 42->27 clean Q&A pairs - Text-only Q&A fallback for episodes with no CALLER diarization labels - TRANSFORMERS_OFFLINE=1 to skip HuggingFace freshness checks - Add diarize_2018.py for targeted re-run + FTS5 rebuild - Add benchmark.py + BENCH_SETUP.md for GURU-BEAST-ROG (RTX 4090) comparison - Commit 9-episode training diarization.json outputs - Session log: 2026-04-27-diarization-pipeline.md Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
303 lines
5.5 KiB
JSON
303 lines
5.5 KiB
JSON
{
|
|
"num_speakers": 2,
|
|
"speaker_map": {
|
|
"CALLER": "CALLER",
|
|
"HOST": "HOST"
|
|
},
|
|
"turns": [
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 0.0,
|
|
"end": 20.0,
|
|
"confidence": 0.9
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 15.0,
|
|
"end": 25.0,
|
|
"confidence": 0.64
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 20.0,
|
|
"end": 140.0,
|
|
"confidence": 0.95
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 135.0,
|
|
"end": 145.0,
|
|
"confidence": 0.85
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 140.0,
|
|
"end": 760.0,
|
|
"confidence": 0.95
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 755.0,
|
|
"end": 795.0,
|
|
"confidence": 0.81
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 790.0,
|
|
"end": 1425.0,
|
|
"confidence": 0.86
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 1420.0,
|
|
"end": 1470.0,
|
|
"confidence": 0.71
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 1465.0,
|
|
"end": 2290.0,
|
|
"confidence": 0.94
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2285.0,
|
|
"end": 2335.0,
|
|
"confidence": 0.82
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2330.0,
|
|
"end": 2345.0,
|
|
"confidence": 0.95
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2340.0,
|
|
"end": 2390.0,
|
|
"confidence": 0.56
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2385.0,
|
|
"end": 2405.0,
|
|
"confidence": 0.93
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2400.0,
|
|
"end": 2420.0,
|
|
"confidence": 0.57
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2415.0,
|
|
"end": 2535.0,
|
|
"confidence": 0.95
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2530.0,
|
|
"end": 2540.0,
|
|
"confidence": 0.49
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2535.0,
|
|
"end": 2555.0,
|
|
"confidence": 0.93
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2550.0,
|
|
"end": 2565.0,
|
|
"confidence": 0.48
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2560.0,
|
|
"end": 2605.0,
|
|
"confidence": 0.94
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2600.0,
|
|
"end": 2610.0,
|
|
"confidence": 0.84
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2605.0,
|
|
"end": 2625.0,
|
|
"confidence": 0.92
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2620.0,
|
|
"end": 2630.0,
|
|
"confidence": 0.76
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2625.0,
|
|
"end": 2675.0,
|
|
"confidence": 0.97
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2670.0,
|
|
"end": 2720.0,
|
|
"confidence": 0.83
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2715.0,
|
|
"end": 2845.0,
|
|
"confidence": 0.88
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2840.0,
|
|
"end": 2930.0,
|
|
"confidence": 0.84
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 2925.0,
|
|
"end": 2955.0,
|
|
"confidence": 0.96
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 2950.0,
|
|
"end": 3005.0,
|
|
"confidence": 0.81
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3000.0,
|
|
"end": 3060.0,
|
|
"confidence": 0.94
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3055.0,
|
|
"end": 3065.0,
|
|
"confidence": 0.83
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3060.0,
|
|
"end": 3075.0,
|
|
"confidence": 0.85
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3070.0,
|
|
"end": 3110.0,
|
|
"confidence": 0.83
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3105.0,
|
|
"end": 3165.0,
|
|
"confidence": 0.96
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3160.0,
|
|
"end": 3180.0,
|
|
"confidence": 0.83
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3175.0,
|
|
"end": 3390.0,
|
|
"confidence": 0.92
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3385.0,
|
|
"end": 3425.0,
|
|
"confidence": 0.83
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3420.0,
|
|
"end": 3610.0,
|
|
"confidence": 0.97
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3605.0,
|
|
"end": 3625.0,
|
|
"confidence": 0.82
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3620.0,
|
|
"end": 3630.0,
|
|
"confidence": 0.85
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3625.0,
|
|
"end": 3685.0,
|
|
"confidence": 0.85
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3680.0,
|
|
"end": 3700.0,
|
|
"confidence": 0.89
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 3695.0,
|
|
"end": 3710.0,
|
|
"confidence": 0.84
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 3705.0,
|
|
"end": 4150.0,
|
|
"confidence": 0.92
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 4145.0,
|
|
"end": 4180.0,
|
|
"confidence": 0.69
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 4175.0,
|
|
"end": 4615.0,
|
|
"confidence": 0.88
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 4610.0,
|
|
"end": 4620.0,
|
|
"confidence": 0.84
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 4615.0,
|
|
"end": 4835.0,
|
|
"confidence": 0.85
|
|
},
|
|
{
|
|
"speaker": "CALLER",
|
|
"start": 4830.0,
|
|
"end": 4885.0,
|
|
"confidence": 0.81
|
|
},
|
|
{
|
|
"speaker": "HOST",
|
|
"start": 4880.0,
|
|
"end": 5300.0,
|
|
"confidence": 0.94
|
|
}
|
|
]
|
|
} |