Hugging Face's logo

ahnafsamin
/
Tacotron2-gronings

Text-to-Speech
Gronings
gronings
Tacotron 2
Model card Files Community
ahnafsamin commited on
Commit
b5ca63b
•
1 Parent(s): 3800212

Upload config.yaml

Browse files
Files changed (1) hide show
  1. config.yaml +215 -0
config.yaml ADDED
@@ -0,0 +1,215 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/train.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ dry_run: false
5
+ iterator_type: sequence
6
+ output_dir: exp/tts_train_raw_char_tacotron
7
+ ngpu: 1
8
+ seed: 0
9
+ num_workers: 1
10
+ num_att_plot: 3
11
+ dist_backend: nccl
12
+ dist_init_method: env://
13
+ dist_world_size: null
14
+ dist_rank: null
15
+ local_rank: 0
16
+ dist_master_addr: null
17
+ dist_master_port: null
18
+ dist_launcher: null
19
+ multiprocessing_distributed: false
20
+ unused_parameters: false
21
+ sharded_ddp: false
22
+ cudnn_enabled: true
23
+ cudnn_benchmark: false
24
+ cudnn_deterministic: true
25
+ collect_stats: false
26
+ write_collected_feats: false
27
+ max_epoch: 200
28
+ patience: null
29
+ val_scheduler_criterion:
30
+ - valid
31
+ - loss
32
+ early_stopping_criterion:
33
+ - valid
34
+ - loss
35
+ - min
36
+ best_model_criterion:
37
+ - - valid
38
+ - loss
39
+ - min
40
+ - - train
41
+ - loss
42
+ - min
43
+ keep_nbest_models: 5
44
+ nbest_averaging_interval: 0
45
+ grad_clip: 1.0
46
+ grad_clip_type: 2.0
47
+ grad_noise: false
48
+ accum_grad: 2
49
+ no_forward_run: false
50
+ resume: true
51
+ train_dtype: float32
52
+ use_amp: false
53
+ log_interval: null
54
+ use_matplotlib: true
55
+ use_tensorboard: true
56
+ use_wandb: false
57
+ wandb_project: null
58
+ wandb_id: null
59
+ wandb_entity: null
60
+ wandb_name: null
61
+ wandb_model_log_interval: -1
62
+ detect_anomaly: false
63
+ pretrain_path: null
64
+ init_param: []
65
+ ignore_init_mismatch: false
66
+ freeze_param: []
67
+ num_iters_per_epoch: 1000
68
+ batch_size: 20
69
+ valid_batch_size: null
70
+ batch_bins: 2000000
71
+ valid_batch_bins: null
72
+ train_shape_file:
73
+ - exp/tts_stats_raw_char_tacotron/train/text_shape.char
74
+ - exp/tts_stats_raw_char_tacotron/train/speech_shape
75
+ valid_shape_file:
76
+ - exp/tts_stats_raw_char_tacotron/valid/text_shape.char
77
+ - exp/tts_stats_raw_char_tacotron/valid/speech_shape
78
+ batch_type: numel
79
+ valid_batch_type: null
80
+ fold_length:
81
+ - 150
82
+ - 204800
83
+ sort_in_batch: descending
84
+ sort_batch: descending
85
+ multiple_iterator: false
86
+ chunk_length: 500
87
+ chunk_shift_ratio: 0.5
88
+ num_cache_chunks: 1024
89
+ train_data_path_and_name_and_type:
90
+ - - dump/raw/tr_no_dev/text
91
+ - text
92
+ - text
93
+ - - dump/raw/tr_no_dev/wav.scp
94
+ - speech
95
+ - sound
96
+ valid_data_path_and_name_and_type:
97
+ - - dump/raw/dev/text
98
+ - text
99
+ - text
100
+ - - dump/raw/dev/wav.scp
101
+ - speech
102
+ - sound
103
+ allow_variable_data_keys: false
104
+ max_cache_size: 0.0
105
+ max_cache_fd: 32
106
+ valid_max_cache_size: null
107
+ optim: adam
108
+ optim_conf:
109
+ lr: 0.001
110
+ eps: 1.0e-06
111
+ weight_decay: 0.0
112
+ scheduler: null
113
+ scheduler_conf: {}
114
+ token_list:
115
+ - <blank>
116
+ - <unk>
117
+ - <space>
118
+ - E
119
+ - N
120
+ - A
121
+ - O
122
+ - T
123
+ - I
124
+ - R
125
+ - D
126
+ - L
127
+ - S
128
+ - K
129
+ - M
130
+ - G
131
+ - U
132
+ - H
133
+ - .
134
+ - W
135
+ - V
136
+ - Z
137
+ - P
138
+ - B
139
+ - ','
140
+ - J
141
+ - C
142
+ - F
143
+ - '?'
144
+ - ''''
145
+ - '!'
146
+ - Y
147
+ - X
148
+ - '`'
149
+ - <sos/eos>
150
+ odim: null
151
+ model_conf: {}
152
+ use_preprocessor: true
153
+ token_type: char
154
+ bpemodel: null
155
+ non_linguistic_symbols: null
156
+ cleaner: tacotron
157
+ g2p: g2p_en
158
+ feats_extract: fbank
159
+ feats_extract_conf:
160
+ n_fft: 1024
161
+ hop_length: 256
162
+ win_length: null
163
+ fs: 22050
164
+ fmin: 80
165
+ fmax: 7600
166
+ n_mels: 80
167
+ normalize: global_mvn
168
+ normalize_conf:
169
+ stats_file: exp/tts_stats_raw_char_tacotron/train/feats_stats.npz
170
+ tts: tacotron2
171
+ tts_conf:
172
+ embed_dim: 512
173
+ elayers: 1
174
+ eunits: 512
175
+ econv_layers: 3
176
+ econv_chans: 512
177
+ econv_filts: 5
178
+ atype: location
179
+ adim: 512
180
+ aconv_chans: 32
181
+ aconv_filts: 15
182
+ cumulate_att_w: true
183
+ dlayers: 2
184
+ dunits: 1024
185
+ prenet_layers: 2
186
+ prenet_units: 256
187
+ postnet_layers: 5
188
+ postnet_chans: 512
189
+ postnet_filts: 5
190
+ output_activation: null
191
+ use_batch_norm: true
192
+ use_concate: true
193
+ use_residual: false
194
+ dropout_rate: 0.5
195
+ zoneout_rate: 0.1
196
+ reduction_factor: 1
197
+ spk_embed_dim: null
198
+ use_masking: true
199
+ bce_pos_weight: 5.0
200
+ use_guided_attn_loss: true
201
+ guided_attn_loss_sigma: 0.4
202
+ guided_attn_loss_lambda: 1.0
203
+ pitch_extract: null
204
+ pitch_extract_conf: {}
205
+ pitch_normalize: null
206
+ pitch_normalize_conf: {}
207
+ energy_extract: null
208
+ energy_extract_conf: {}
209
+ energy_normalize: null
210
+ energy_normalize_conf: {}
211
+ required:
212
+ - output_dir
213
+ - token_list
214
+ version: 0.10.7a1
215
+ distributed: false