imenLa commited on
Commit
99dc793
·
verified ·
1 Parent(s): 4fc5ccd

Upload 2 files

Browse files
exp/tts_finetune_tacotron2_phn_notoken_r3_lr03_sr22050/checkpoint.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90723dc2070a3dd9e033bc33371d9f11e81d1687bed250cf8bba6d746aafd7af
3
+ size 323288504
exp/tts_finetune_tacotron2_phn_notoken_r3_lr03_sr22050/config.yaml ADDED
@@ -0,0 +1,264 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/tuning/finetune_tacotron2.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ dry_run: false
5
+ iterator_type: sequence
6
+ output_dir: exp/tts_finetune_tacotron2_phn_notoken_r3_lr03_sr22050
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: 120
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: 1
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
+ create_graph_in_tensorboard: false
57
+ use_wandb: false
58
+ wandb_project: null
59
+ wandb_id: null
60
+ wandb_entity: null
61
+ wandb_name: null
62
+ wandb_model_log_interval: -1
63
+ detect_anomaly: false
64
+ pretrain_path: null
65
+ init_param:
66
+ - downloads/0123456cl789/exp/tts_train_tacotron2_r3_raw_phn_tacotron_g2p_en_no_space/train.loss.ave_5best.pth:::tts.enc.embed
67
+ ignore_init_mismatch: false
68
+ freeze_param: []
69
+ num_iters_per_epoch: 200
70
+ batch_size: 20
71
+ valid_batch_size: null
72
+ batch_bins: 1600000
73
+ valid_batch_bins: null
74
+ train_shape_file:
75
+ - exp/tts_stats_raw_phn_none/train/text_shape.phn
76
+ - exp/tts_stats_raw_phn_none/train/speech_shape
77
+ valid_shape_file:
78
+ - exp/tts_stats_raw_phn_none/valid/text_shape.phn
79
+ - exp/tts_stats_raw_phn_none/valid/speech_shape
80
+ batch_type: numel
81
+ valid_batch_type: null
82
+ fold_length:
83
+ - 150
84
+ - 204800
85
+ sort_in_batch: descending
86
+ sort_batch: descending
87
+ multiple_iterator: false
88
+ chunk_length: 500
89
+ chunk_shift_ratio: 0.5
90
+ num_cache_chunks: 1024
91
+ chunk_excluded_key_prefixes: []
92
+ train_data_path_and_name_and_type:
93
+ - - dump/raw/train/text
94
+ - text
95
+ - text
96
+ - - dump/raw/train/wav.scp
97
+ - speech
98
+ - sound
99
+ valid_data_path_and_name_and_type:
100
+ - - dump/raw/dev/text
101
+ - text
102
+ - text
103
+ - - dump/raw/dev/wav.scp
104
+ - speech
105
+ - sound
106
+ allow_variable_data_keys: false
107
+ max_cache_size: 0.0
108
+ max_cache_fd: 32
109
+ valid_max_cache_size: null
110
+ exclude_weight_decay: false
111
+ exclude_weight_decay_conf: {}
112
+ optim: adam
113
+ optim_conf:
114
+ lr: 0.001
115
+ eps: 1.0e-06
116
+ weight_decay: 0.0
117
+ scheduler: null
118
+ scheduler_conf: {}
119
+ token_list:
120
+ - <blank>
121
+ - <unk>
122
+ - a
123
+ - sil
124
+ - l
125
+ - aa
126
+ - m
127
+ - ii0
128
+ - t
129
+ - <
130
+ - n
131
+ - r
132
+ - E
133
+ - i0
134
+ - b
135
+ - uu0
136
+ - f
137
+ - i1
138
+ - k
139
+ - w
140
+ - A
141
+ - s
142
+ - y
143
+ - d
144
+ - q
145
+ - h
146
+ - H
147
+ - $
148
+ - u0
149
+ - AA
150
+ - j
151
+ - T
152
+ - x
153
+ - S
154
+ - z
155
+ - ll
156
+ - I1
157
+ - D
158
+ - II0
159
+ - g
160
+ - tt
161
+ - rr
162
+ - I0
163
+ - UU0
164
+ - dd
165
+ - u1
166
+ - U0
167
+ - mm
168
+ - nn
169
+ - '*'
170
+ - $$
171
+ - bb
172
+ - yy
173
+ - ss
174
+ - jj
175
+ - ww
176
+ - ^
177
+ - SS
178
+ - TT
179
+ - Z
180
+ - zz
181
+ - kk
182
+ - U1
183
+ - HH
184
+ - ff
185
+ - qq
186
+ - xx
187
+ - ^^
188
+ - DD
189
+ - hh
190
+ - EE
191
+ - ZZ
192
+ - '**'
193
+ - aaaa
194
+ - ssss
195
+ - v
196
+ - uu1
197
+ - jjjj
198
+ - <sos/eos>
199
+ odim: null
200
+ model_conf: {}
201
+ use_preprocessor: true
202
+ token_type: phn
203
+ bpemodel: null
204
+ non_linguistic_symbols: null
205
+ cleaner: null
206
+ g2p: null
207
+ feats_extract: fbank
208
+ feats_extract_conf:
209
+ n_fft: 1024
210
+ hop_length: 256
211
+ win_length: null
212
+ fs: 22050
213
+ fmin: 80
214
+ fmax: 7600
215
+ n_mels: 80
216
+ normalize: global_mvn
217
+ normalize_conf:
218
+ stats_file: exp/tts_stats_raw_phn_none/train/feats_stats.npz
219
+ tts: tacotron2
220
+ tts_conf:
221
+ embed_dim: 512
222
+ elayers: 1
223
+ eunits: 512
224
+ econv_layers: 3
225
+ econv_chans: 512
226
+ econv_filts: 5
227
+ atype: location
228
+ adim: 512
229
+ aconv_chans: 32
230
+ aconv_filts: 15
231
+ cumulate_att_w: true
232
+ dlayers: 2
233
+ dunits: 1024
234
+ prenet_layers: 2
235
+ prenet_units: 256
236
+ postnet_layers: 5
237
+ postnet_chans: 512
238
+ postnet_filts: 5
239
+ output_activation: null
240
+ use_batch_norm: true
241
+ use_concate: true
242
+ use_residual: false
243
+ dropout_rate: 0.5
244
+ zoneout_rate: 0.1
245
+ reduction_factor: 3
246
+ spk_embed_dim: null
247
+ use_masking: true
248
+ bce_pos_weight: 20.0
249
+ use_guided_attn_loss: true
250
+ guided_attn_loss_sigma: 0.4
251
+ guided_attn_loss_lambda: 1.0
252
+ pitch_extract: null
253
+ pitch_extract_conf: {}
254
+ pitch_normalize: null
255
+ pitch_normalize_conf: {}
256
+ energy_extract: null
257
+ energy_extract_conf: {}
258
+ energy_normalize: null
259
+ energy_normalize_conf: {}
260
+ required:
261
+ - output_dir
262
+ - token_list
263
+ version: '202304'
264
+ distributed: false