| { | |
| "model_type": "pyannote-segmentation", | |
| "sample_rate": 16000, | |
| "sincnet": { | |
| "n_filters": [ | |
| 80, | |
| 60, | |
| 60 | |
| ], | |
| "kernel_sizes": [ | |
| 251, | |
| 5, | |
| 5 | |
| ], | |
| "strides": [ | |
| 10, | |
| 1, | |
| 1 | |
| ], | |
| "pool_sizes": [ | |
| 3, | |
| 3, | |
| 3 | |
| ] | |
| }, | |
| "lstm": { | |
| "hidden_size": 128, | |
| "num_layers": 4, | |
| "bidirectional": true | |
| }, | |
| "linear": { | |
| "hidden_size": 128, | |
| "num_layers": 2 | |
| }, | |
| "num_classes": 7, | |
| "max_speakers": 3, | |
| "powerset_max_classes": 2, | |
| "num_frames_per_chunk": 589, | |
| "chunk_duration": 10.0, | |
| "chunk_step_ratio": 0.1, | |
| "warm_up": [ | |
| 0.0, | |
| 0.0 | |
| ] | |
| } |