Hugging Face's logo

salmanshahid
/
segmentation

Voice Activity Detection
PyTorch
TensorBoard
pyannote
audio
voice
speech
speaker
speaker-segmentation
overlapped-speech-detection
resegmentation
Model card Files Metrics Community
segmentation
File size: 1,980 Bytes 
07d5bbe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
protocol: X.SpeakerDiarization.Custom
patience: 20
task:
  _target_: pyannote.audio.tasks.Segmentation
  duration: 5.0
  warm_up: 0.0
  balance: null
  overlap:
    probability: 0.5
    snr_min: 0.0
    snr_max: 10.0
  weight: null
  batch_size: 32
  num_workers: 10
  pin_memory: false
  loss: bce
  vad_loss: bce
model:
  _target_: pyannote.audio.models.segmentation.PyanNet
  sincnet:
    stride: 10
  lstm:
    num_layers: 4
    monolithic: true
    dropout: 0.5
  linear:
    num_layers: 2
optimizer:
  _target_: torch.optim.Adam
  lr: 0.001
  betas:
  - 0.9
  - 0.999
  eps: 1.0e-08
  weight_decay: 0
  amsgrad: false
trainer:
  _target_: pytorch_lightning.Trainer
  accelerator: ddp
  accumulate_grad_batches: 1
  amp_backend: native
  amp_level: O2
  auto_lr_find: false
  auto_scale_batch_size: false
  auto_select_gpus: true
  benchmark: true
  check_val_every_n_epoch: 1
  checkpoint_callback: true
  deterministic: false
  fast_dev_run: false
  flush_logs_every_n_steps: 100
  gpus: -1
  gradient_clip_val: 0.5
  limit_test_batches: 1.0
  limit_train_batches: 1.0
  limit_val_batches: 1.0
  log_every_n_steps: 50
  log_gpu_memory: null
  max_epochs: 1000
  max_steps: null
  min_epochs: 1
  min_steps: null
  num_nodes: 1
  num_processes: 1
  num_sanity_val_steps: 2
  overfit_batches: 0.0
  precision: 32
  prepare_data_per_node: true
  process_position: 0
  profiler: null
  progress_bar_refresh_rate: 1
  reload_dataloaders_every_epoch: false
  replace_sampler_ddp: true
  sync_batchnorm: false
  terminate_on_nan: false
  tpu_cores: null
  track_grad_norm: -1
  truncated_bptt_steps: null
  val_check_interval: 1.0
  weights_save_path: null
  weights_summary: top
augmentation:
  transform: Compose
  params:
    shuffle: false
    transforms:
    - transform: AddBackgroundNoise
      params:
        background_paths: /gpfswork/rech/eie/commun/data/background/musan
        min_snr_in_db: 5.0
        max_snr_in_db: 15.0
        mode: per_example
        p: 0.9