| { | |
| "do_normalize": true, | |
| "feature_extractor_type": "Wav2Vec2FeatureExtractor", | |
| "feature_size": 1, | |
| "padding_side": "right", | |
| "padding_value": 0.0, | |
| "return_attention_mask": true, | |
| "sampling_rate": 16000, | |
| "preprocessing": { | |
| "audio_normalization": { | |
| "method": "peak", | |
| "target_level": -23.0, | |
| "headroom_db": 3.0 | |
| }, | |
| "spectral_features": { | |
| "mel_filters": 128, | |
| "window_size_ms": 25, | |
| "stride_ms": 10, | |
| "fmin": 50, | |
| "fmax": 8000, | |
| "htk_compat": true | |
| }, | |
| "augmentation": { | |
| "time_masking": { | |
| "enabled": true, | |
| "time_mask_param": 100, | |
| "num_masks": 2 | |
| }, | |
| "freq_masking": { | |
| "enabled": true, | |
| "freq_mask_param": 27, | |
| "num_masks": 2 | |
| }, | |
| "noise": { | |
| "enabled": true, | |
| "noise_types": ["gaussian", "pink"], | |
| "snr_range": [5, 20] | |
| } | |
| }, | |
| "signal_enhancement": { | |
| "vad": { | |
| "enabled": true, | |
| "threshold": 0.5, | |
| "min_speech_duration_ms": 250 | |
| }, | |
| "noise_reduction": { | |
| "enabled": true, | |
| "method": "spectral_gating", | |
| "stationary_threshold": 1.5 | |
| } | |
| } | |
| }, | |
| "advanced_settings": { | |
| "feature_extraction": { | |
| "normalize_means": true, | |
| "normalize_vars": true, | |
| "deltas_order": 2, | |
| "cmvn_window": 300 | |
| }, | |
| "resampling": { | |
| "method": "kaiser_best", | |
| "lowpass_filter_width": 64, | |
| "rolloff": 0.945, | |
| "beta": 14.0 | |
| }, | |
| "performance": { | |
| "num_workers": 4, | |
| "pin_memory": true, | |
| "prefetch_factor": 2, | |
| "persistent_workers": true | |
| } | |
| } | |
| } | |