data_type: shard
dataloader_args:
  batch_size: 16
  drop_last: true
  num_workers: 6
  pin_memory: false
  prefetch_factor: 8
dataset_args:
  aug_prob: 0
  cmvn: true
  cmvn_args:
    norm_mean: true
    norm_var: false
  filter: true
  filter_args:
    max_num_frames: 400
    min_num_frames: 50
  frontend: s3prl
  num_frms: 150
  resample_rate: 16000
  s3prl_args:
    download_dir: ./s3prl_hub
    frame_length: 20
    frame_shift: 20
    frozen: false
    layer: -1
    layerwise_feature: true
    multilayer_feature: true
    upstream_args:
      name: wav2vec2_large_960
  sample_num_per_epoch: 0
  shuffle: true
  shuffle_args:
    shuffle_size: 2500
  spec_aug: false
  spec_aug_args:
    max_f: 8
    max_t: 10
    num_f_mask: 1
    num_t_mask: 1
    prob: 0.6
  speed_perturb: false
enable_amp: false
exp_dir: exp/MHFA_wav2vec2_large_960-FT-1stage5
gpus:
- 1
log_batch_interval: 100
loss: CrossEntropyLoss
loss_args: {}
margin_scheduler: MarginScheduler
margin_update:
  epoch_iter: 11397
  final_margin: 0.0
  fix_start_epoch: 1
  increase_start_epoch: 1
  increase_type: exp
  initial_margin: 0.0
  update_margin: true
model: SSL_BACKEND_MHFA
model_args:
  compression_dim: 128
  embed_dim: 256
  feat_dim: 1024
  feature_grad_mult: 0.05
  head_nb: 32
  nb_layer: 25
model_init: null
num_avg: 2
num_epochs: 5
optimizer: AdamW
optimizer_args:
  lr: 1.0e-05
  weight_decay: 1.0e-08
projection_args:
  do_lm: false
  easy_margin: false
  embed_dim: 256
  num_class: 2
  project_type: softmax
  scale: 32.0
save_epoch_interval: 1
scheduler: ExponentialDecrease
scheduler_args:
  epoch_iter: 11397
  final_lr: 5.0e-07
  initial_lr: 1.0e-05
  num_epochs: 5
  scale_ratio: 0.25
  warm_from_zero: true
  warm_up_epoch: 1
seed: 42
train_data: data/asv5/train/shard.list
train_label: data/asv5/train/utt2cls