File size: 1,539 Bytes
9188d4c b0b7f70 9188d4c b0b7f70 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 | {
"model_type": " laughter_detection",
"architectures": ["WavLMForLaughterDetection"],
"task": "audio_classification",
"num_labels": 2,
"label_mapping": {"0": "no_laugh", "1": "laughter"},
"dataset": {
"name": "standup4ai",
"num_videos": 221,
"total_duration_minutes": 540,
"language": "en",
"class_balance": {
"positive": 0.132,
"negative": 0.868
}
},
"audio_config": {
"sample_rate": 16000,
"hop_length": 512,
"n_fft": 2048,
"n_mels": 128,
"fmin": 80,
"fmax": 8000
},
"prosody_features": {
"dimension": 23,
"source": "eGeMAPS v02",
"extraction": "librosa + openSMILE",
"features": [
"F0_mean", "F0_std", "F0_slope", "F0_median", "F0_range",
"pause_duration",
"speech_rate", "word_duration_mean", "word_duration_std", "syllable_rate", "phonation_time", "total_time",
"RMS_energy_mean", "RMS_energy_std", "RMS_energy_range",
"MFCC1", "MFCC2", "MFCC3",
"spectral_centroid", "spectral_bandwidth", "spectral_rolloff",
"voice_activity_ratio",
"formant_emphasis"
]
},
"validation": {
"method": "5-fold_stratified_cv",
"n_folds": 5,
"prosody_only_f1": "0.960 ± 0.021",
"fusion_f1": "0.879 ± 0.034",
"delta_f1": "-0.081",
"p_value": 0.024,
"confidence_interval_95": [-0.152, -0.010]
},
"fusion_analysis": {
"finding": "NEGATIVE_RESULT",
"cause": "frozen_WavLM_mean_pool_destoys_temporal_cues",
"regression": "prosody_only_better_than_fusion"
}
} |