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
| # network architecture
| # encoder related
| encoder_conf:
| dropout_rate: 0.1
| positional_dropout_rate: 0.1
| attention_dropout_rate: 0.1
|
| # decoder related
| decoder_conf:
| dropout_rate: 0.1
| positional_dropout_rate: 0.1
| self_attention_dropout_rate: 0.1
| src_attention_dropout_rate: 0.1
|
| predictor_conf:
| threshold: 1.0
| l_order: 1
| r_order: 1
| tail_threshold: 0.45
|
| # hybrid CTC/attention
| model_conf:
| ctc_weight: 0.0
| lsm_weight: 0.1 # label smoothing option
| length_normalized_loss: true
| predictor_weight: 1.0
| predictor_bias: 1
| sampling_ratio: 0.75
|
| # minibatch related
| # dataset_type: small
| batch_type: length
| batch_bins: 2000
| num_workers: 16
| # dataset_type: large
| dataset_conf:
| filter_conf:
| min_length: 10
| max_length: 250
| min_token_length: 1
| max_token_length: 200
| shuffle: true
| shuffle_conf:
| shuffle_size: 10240
| sort_size: 500
| batch_conf:
| batch_type: 'token'
| batch_size: 6000
| num_workers: 16
|
| # optimization related
| accum_grad: 1
| grad_clip: 5
| max_epoch: 20
| val_scheduler_criterion:
| - valid
| - acc
| best_model_criterion:
| - - valid
| - acc
| - max
| keep_nbest_models: 10
|
| optim: adam
| optim_conf:
| lr: 0.0005
| scheduler: warmuplr
| scheduler_conf:
| warmup_steps: 30000
|
| specaug: specaug_lfr
| specaug_conf:
| apply_time_warp: false
| time_warp_window: 5
| time_warp_mode: bicubic
| apply_freq_mask: true
| freq_mask_width_range:
| - 0
| - 30
| lfr_rate: 6
| num_freq_mask: 1
| apply_time_mask: true
| time_mask_width_range:
| - 0
| - 12
| num_time_mask: 1
|
| unused_parameters: true
| log_interval: 50
| normalize: None
| split_with_space: true
|
|
