File size: 1,286 Bytes
b4d3e53
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# @package _group_

common:
  memory_efficient_fp16: true
  log_format: json
  log_interval: 100

checkpoint:
  save_interval_updates: 10000
  keep_interval_updates: 1
  no_epoch_checkpoints: true
  save_dir: /esat/spchtemp/scratch/jponcele/selfsupervised_exps/result/pretrain_w2v2_cgn-unsup-VW_base

task:
  _name: audio_pretraining
  data: /users/spraak/jponcele/BenchmarkingSS/data/cgn_unsup_VW_w2v2
  max_sample_size: 250000
  min_sample_size: 4000
  segments: true
  normalize: true

dataset:
  num_workers: 6
  #batch_size: 4
  max_tokens: 1400000
  skip_invalid_size_inputs_valid_test: true
  valid_subset: test
  data_buffer_size: 1  #2
  required_batch_size_multiple: 1  #default=8

distributed_training:
  distributed_world_size: 1
  ddp_backend: legacy_ddp

criterion:
  _name: wav2vec
  infonce: true
  log_keys: ["prob_perplexity","code_perplexity","temp"]
  loss_weights: [0.1, 10]

optimization:
  max_update: 400000
  lr: [0.0005]
  update_freq: [32]

optimizer:
  _name: adam
  adam_betas: (0.9,0.98)
  adam_eps: 1e-06
  weight_decay: 0.01

lr_scheduler:
  _name: polynomial_decay
  warmup_updates: 50000

model:
  _name: wav2vec2
  quantize_targets: true
  final_dim: 256
  encoder_layerdrop: 0.05
  dropout_input: 0.1
  dropout_features: 0.1
  feature_grad_mult: 0.1