NeonBohdan commited on
Commit
a8b94a8
1 Parent(s): c03f373

Init configs

Browse files
Files changed (3) hide show
  1. config.json +255 -0
  2. language_ids.json +3 -0
  3. speaker_ids.json +4 -0
config.json ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_path": null,
3
+ "logger_uri": null,
4
+ "run_name": "",
5
+ "project_name": null,
6
+ "run_description": "",
7
+ "print_step": 25,
8
+ "plot_step": 100,
9
+ "model_param_stats": false,
10
+ "wandb_entity": null,
11
+ "dashboard_logger": "tensorboard",
12
+ "log_model_step": null,
13
+ "save_step": 10000,
14
+ "save_n_checkpoints": 5,
15
+ "save_checkpoints": true,
16
+ "save_all_best": false,
17
+ "save_best_after": 10000,
18
+ "target_loss": null,
19
+ "print_eval": false,
20
+ "test_delay_epochs": -1,
21
+ "run_eval": true,
22
+ "distributed_backend": "nccl",
23
+ "distributed_url": "tcp://localhost:54321",
24
+ "mixed_precision": true,
25
+ "epochs": 1000,
26
+ "batch_size": 32,
27
+ "eval_batch_size": 4,
28
+ "grad_clip": [
29
+ 1000,
30
+ 1000
31
+ ],
32
+ "scheduler_after_epoch": true,
33
+ "lr": 0.001,
34
+ "optimizer": "AdamW",
35
+ "optimizer_params": {
36
+ "betas": [
37
+ 0.8,
38
+ 0.99
39
+ ],
40
+ "eps": 1e-09,
41
+ "weight_decay": 0.01
42
+ },
43
+ "lr_scheduler": "",
44
+ "lr_scheduler_params": {},
45
+ "use_grad_scaler": false,
46
+ "cudnn_enable": true,
47
+ "cudnn_deterministic": false,
48
+ "cudnn_benchmark": false,
49
+ "training_seed": 54321,
50
+ "model": "vits",
51
+ "num_loader_workers": 8,
52
+ "num_eval_loader_workers": 8,
53
+ "use_noise_augment": false,
54
+ "audio": {
55
+ "fft_size": 1024,
56
+ "win_length": 1024,
57
+ "hop_length": 256,
58
+ "frame_shift_ms": null,
59
+ "frame_length_ms": null,
60
+ "stft_pad_mode": "reflect",
61
+ "sample_rate": 22050,
62
+ "resample": false,
63
+ "preemphasis": 0.0,
64
+ "ref_level_db": 20,
65
+ "do_sound_norm": false,
66
+ "log_func": "np.log",
67
+ "do_trim_silence": true,
68
+ "trim_db": 45,
69
+ "do_rms_norm": false,
70
+ "db_level": null,
71
+ "power": 1.5,
72
+ "griffin_lim_iters": 60,
73
+ "num_mels": 80,
74
+ "mel_fmin": 0.0,
75
+ "mel_fmax": null,
76
+ "spec_gain": 1.0,
77
+ "do_amp_to_db_linear": false,
78
+ "do_amp_to_db_mel": true,
79
+ "pitch_fmax": 640.0,
80
+ "pitch_fmin": 0.0,
81
+ "signal_norm": true,
82
+ "min_level_db": -100,
83
+ "symmetric_norm": true,
84
+ "max_norm": 4.0,
85
+ "clip_norm": true,
86
+ "stats_path": null
87
+ },
88
+ "use_phonemes": false,
89
+ "phonemizer": null,
90
+ "phoneme_language": null,
91
+ "compute_input_seq_cache": true,
92
+ "text_cleaner": "multilingual_cleaners",
93
+ "enable_eos_bos_chars": false,
94
+ "test_sentences_file": "",
95
+ "phoneme_cache_path": null,
96
+ "characters": {
97
+ "characters_class": "TTS.tts.utils.text.characters.Graphemes",
98
+ "vocab_dict": null,
99
+ "pad": "<PAD>",
100
+ "eos": "<EOS>",
101
+ "bos": "<BOS>",
102
+ "blank": "<BLNK>",
103
+ "characters": "abcdefghijklmnopqrstuvwxyz\u00af\u00b7\u00df\u00e0\u00e1\u00e2\u00e3\u00e4\u00e6\u00e7\u00e8\u00e9\u00ea\u00eb\u00ec\u00ed\u00ee\u00ef\u00f1\u00f2\u00f3\u00f4\u00f5\u00f6\u00f9\u00fa\u00fb\u00fc\u00ff\u0101\u0105\u0107\u0113\u0119\u011b\u012b\u0131\u0142\u0144\u014d\u0151\u0153\u015b\u016b\u0171\u017a\u017c\u01ce\u01d0\u01d2\u01d4\u0430\u0431\u0432\u0433\u0434\u0435\u0436\u0437\u0438\u0439\u043a\u043b\u043c\u043d\u043e\u043f\u0440\u0441\u0442\u0443\u0444\u0445\u0446\u0447\u0448\u0449\u044a\u044b\u044c\u044d\u044e\u044f\u0451\u0454\u0456\u0457\u0491",
104
+ "punctuations": "!'(),-.:;? ",
105
+ "phonemes": null,
106
+ "is_unique": true,
107
+ "is_sorted": true
108
+ },
109
+ "add_blank": true,
110
+ "batch_group_size": 0,
111
+ "loss_masking": null,
112
+ "sort_by_audio_len": false,
113
+ "min_audio_len": 32768,
114
+ "max_audio_len": 132300,
115
+ "min_text_len": 1,
116
+ "max_text_len": Infinity,
117
+ "compute_f0": false,
118
+ "compute_linear_spec": true,
119
+ "precompute_num_workers": 12,
120
+ "start_by_longest": false,
121
+ "datasets": [
122
+ {
123
+ "name": "css10",
124
+ "path": "./datasets/css10_hu/",
125
+ "meta_file_train": "transcript.txt",
126
+ "ignored_speakers": null,
127
+ "language": "hu",
128
+ "meta_file_val": "",
129
+ "meta_file_attn_mask": ""
130
+ }
131
+ ],
132
+ "test_sentences": [
133
+ [
134
+ "A sziv\u00e1rv\u00e1ny olyan optikai jelens\u00e9g, melyet es\u0151- vagy p\u00e1racseppek okoznak, mikor a f\u00e9ny prizmaszer\u0171en megt\u00f6rik rajtuk \u00e9s sz\u00edneire bomlik.",
135
+ "css10",
136
+ null,
137
+ "hu"
138
+ ]
139
+ ],
140
+ "eval_split_max_size": null,
141
+ "eval_split_size": 0.01,
142
+ "use_speaker_weighted_sampler": false,
143
+ "speaker_weighted_sampler_alpha": 1.0,
144
+ "use_language_weighted_sampler": true,
145
+ "language_weighted_sampler_alpha": 1.0,
146
+ "model_args": {
147
+ "num_chars": 129,
148
+ "out_channels": 513,
149
+ "spec_segment_size": 32,
150
+ "hidden_channels": 192,
151
+ "hidden_channels_ffn_text_encoder": 768,
152
+ "num_heads_text_encoder": 2,
153
+ "num_layers_text_encoder": 6,
154
+ "kernel_size_text_encoder": 3,
155
+ "dropout_p_text_encoder": 0.1,
156
+ "dropout_p_duration_predictor": 0.5,
157
+ "kernel_size_posterior_encoder": 5,
158
+ "dilation_rate_posterior_encoder": 1,
159
+ "num_layers_posterior_encoder": 16,
160
+ "kernel_size_flow": 5,
161
+ "dilation_rate_flow": 1,
162
+ "num_layers_flow": 4,
163
+ "resblock_type_decoder": "2",
164
+ "resblock_kernel_sizes_decoder": [
165
+ 3,
166
+ 5,
167
+ 7
168
+ ],
169
+ "resblock_dilation_sizes_decoder": [
170
+ [
171
+ 1,
172
+ 2
173
+ ],
174
+ [
175
+ 2,
176
+ 6
177
+ ],
178
+ [
179
+ 3,
180
+ 12
181
+ ]
182
+ ],
183
+ "upsample_rates_decoder": [
184
+ 8,
185
+ 8,
186
+ 4
187
+ ],
188
+ "upsample_initial_channel_decoder": 256,
189
+ "upsample_kernel_sizes_decoder": [
190
+ 16,
191
+ 16,
192
+ 8
193
+ ],
194
+ "use_sdp": true,
195
+ "noise_scale": 1.0,
196
+ "inference_noise_scale": 0.3,
197
+ "length_scale": 1,
198
+ "noise_scale_dp": 1.0,
199
+ "inference_noise_scale_dp": 0.3,
200
+ "max_inference_len": null,
201
+ "init_discriminator": false,
202
+ "use_spectral_norm_disriminator": false,
203
+ "use_speaker_embedding": true,
204
+ "num_speakers": 0,
205
+ "speakers_file": "./logs/run-June-03-2022_10+09PM-f3efb56/speakers.json",
206
+ "d_vector_file": null,
207
+ "speaker_embedding_channels": 256,
208
+ "use_d_vector_file": false,
209
+ "d_vector_dim": 0,
210
+ "detach_dp_input": true,
211
+ "use_language_embedding": true,
212
+ "embedded_language_dim": 4,
213
+ "num_languages": 1,
214
+ "language_ids_file": null,
215
+ "use_speaker_encoder_as_loss": false,
216
+ "speaker_encoder_config_path": "",
217
+ "speaker_encoder_model_path": "",
218
+ "condition_dp_on_speaker": true,
219
+ "freeze_encoder": false,
220
+ "freeze_DP": false,
221
+ "freeze_PE": false,
222
+ "freeze_flow_decoder": false,
223
+ "freeze_waveform_decoder": false
224
+ },
225
+ "lr_gen": 0.0002,
226
+ "lr_disc": 0.0002,
227
+ "lr_scheduler_gen": "ExponentialLR",
228
+ "lr_scheduler_gen_params": {
229
+ "gamma": 0.999875,
230
+ "last_epoch": -1
231
+ },
232
+ "lr_scheduler_disc": "ExponentialLR",
233
+ "lr_scheduler_disc_params": {
234
+ "gamma": 0.999875,
235
+ "last_epoch": -1
236
+ },
237
+ "kl_loss_alpha": 1.0,
238
+ "disc_loss_alpha": 1.0,
239
+ "gen_loss_alpha": 1.0,
240
+ "feat_loss_alpha": 1.0,
241
+ "mel_loss_alpha": 45.0,
242
+ "dur_loss_alpha": 1.0,
243
+ "speaker_encoder_loss_alpha": 1.0,
244
+ "return_wav": true,
245
+ "r": 1,
246
+ "num_speakers": 1,
247
+ "use_speaker_embedding": true,
248
+ "speakers_file": null,
249
+ "speaker_embedding_channels": 256,
250
+ "language_ids_file": null,
251
+ "use_language_embedding": true,
252
+ "use_d_vector_file": false,
253
+ "d_vector_file": null,
254
+ "d_vector_dim": 0
255
+ }
language_ids.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "hu": 0
3
+ }
speaker_ids.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "hu": 0,
3
+ "diana": 0
4
+ }