utkarsh2299 commited on
Commit
9d41ce1
1 Parent(s): 1d0af52

Upload 12 files

Browse files
assamese/female/model/config.yaml ADDED
@@ -0,0 +1,272 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/tuning/train_fastspeech2.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ dry_run: false
5
+ iterator_type: sequence
6
+ output_dir: exp/tts_train_fastspeech2_raw_char_None
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: 4
14
+ dist_rank: 0
15
+ local_rank: 0
16
+ dist_master_addr: localhost
17
+ dist_master_port: 60699
18
+ dist_launcher: null
19
+ multiprocessing_distributed: true
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: 1000
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
+ grad_clip: 1.0
45
+ grad_clip_type: 2.0
46
+ grad_noise: false
47
+ accum_grad: 8
48
+ no_forward_run: false
49
+ resume: true
50
+ train_dtype: float32
51
+ use_amp: false
52
+ log_interval: null
53
+ use_tensorboard: true
54
+ use_wandb: false
55
+ wandb_project: null
56
+ wandb_id: null
57
+ wandb_entity: null
58
+ wandb_name: null
59
+ wandb_model_log_interval: -1
60
+ detect_anomaly: false
61
+ pretrain_path: null
62
+ init_param: []
63
+ ignore_init_mismatch: false
64
+ freeze_param: []
65
+ num_iters_per_epoch: 800
66
+ batch_size: 20
67
+ valid_batch_size: null
68
+ batch_bins: 3000000
69
+ valid_batch_bins: null
70
+ train_shape_file:
71
+ - exp/tts_stats_raw_char_None/train/text_shape.char
72
+ - exp/tts_stats_raw_char_None/train/speech_shape
73
+ valid_shape_file:
74
+ - exp/tts_stats_raw_char_None/valid/text_shape.char
75
+ - exp/tts_stats_raw_char_None/valid/speech_shape
76
+ batch_type: numel
77
+ valid_batch_type: null
78
+ fold_length:
79
+ - 150
80
+ - 204800
81
+ sort_in_batch: descending
82
+ sort_batch: descending
83
+ multiple_iterator: false
84
+ chunk_length: 500
85
+ chunk_shift_ratio: 0.5
86
+ num_cache_chunks: 1024
87
+ train_data_path_and_name_and_type:
88
+ - - dump/raw/tr_no_dev/text
89
+ - text
90
+ - text
91
+ - - duration_info/tr_no_dev/durations
92
+ - durations
93
+ - text_int
94
+ - - dump/raw/tr_no_dev/wav.scp
95
+ - speech
96
+ - sound
97
+ - - exp/tts_stats_raw_char_None/train/collect_feats/pitch.scp
98
+ - pitch
99
+ - npy
100
+ - - exp/tts_stats_raw_char_None/train/collect_feats/energy.scp
101
+ - energy
102
+ - npy
103
+ valid_data_path_and_name_and_type:
104
+ - - dump/raw/dev/text
105
+ - text
106
+ - text
107
+ - - duration_info/dev/durations
108
+ - durations
109
+ - text_int
110
+ - - dump/raw/dev/wav.scp
111
+ - speech
112
+ - sound
113
+ - - exp/tts_stats_raw_char_None/valid/collect_feats/pitch.scp
114
+ - pitch
115
+ - npy
116
+ - - exp/tts_stats_raw_char_None/valid/collect_feats/energy.scp
117
+ - energy
118
+ - npy
119
+ allow_variable_data_keys: false
120
+ max_cache_size: 0.0
121
+ max_cache_fd: 32
122
+ valid_max_cache_size: null
123
+ optim: adam
124
+ optim_conf:
125
+ lr: 1.0
126
+ scheduler: noamlr
127
+ scheduler_conf:
128
+ model_size: 384
129
+ warmup_steps: 4000
130
+ token_list:
131
+ - <blank>
132
+ - <unk>
133
+ - A
134
+ - a
135
+ - i
136
+ - ','
137
+ - r
138
+ - E
139
+ - k
140
+ - t
141
+ - n
142
+ - l
143
+ - o
144
+ - b
145
+ - u
146
+ - m
147
+ - y
148
+ - $
149
+ - .
150
+ - p
151
+ - h
152
+ - d
153
+ - s
154
+ - I
155
+ - g
156
+ - j
157
+ - ऐ
158
+ - c
159
+ - ट
160
+ - ख
161
+ - C
162
+ - w
163
+ - श
164
+ - M
165
+ - B
166
+ - थ
167
+ - ध
168
+ - ण
169
+ - ष
170
+ - ठ
171
+ - घ
172
+ - U
173
+ - P
174
+ - q
175
+ - ड
176
+ - ङ
177
+ - R
178
+ - औ
179
+ - ञ
180
+ - D
181
+ - ढ
182
+ - <sos/eos>
183
+ odim: null
184
+ model_conf: {}
185
+ use_preprocessor: true
186
+ token_type: char
187
+ bpemodel: null
188
+ non_linguistic_symbols: null
189
+ cleaner: null
190
+ g2p: g2p_en_no_space
191
+ feats_extract: fbank
192
+ feats_extract_conf:
193
+ n_fft: 1024
194
+ hop_length: 256
195
+ win_length: 1024
196
+ fs: 22050
197
+ fmin: 0
198
+ fmax: 8000
199
+ n_mels: 80
200
+ normalize: global_mvn
201
+ normalize_conf:
202
+ stats_file: /speech/arun/released_models/tts/female/assamese/fastspeech2_hs/feats_stats.npz
203
+ tts: fastspeech2
204
+ tts_conf:
205
+ adim: 384
206
+ aheads: 2
207
+ elayers: 4
208
+ eunits: 1536
209
+ dlayers: 4
210
+ dunits: 1536
211
+ positionwise_layer_type: conv1d
212
+ positionwise_conv_kernel_size: 3
213
+ duration_predictor_layers: 2
214
+ duration_predictor_chans: 256
215
+ duration_predictor_kernel_size: 3
216
+ postnet_layers: 5
217
+ postnet_filts: 5
218
+ postnet_chans: 256
219
+ use_masking: true
220
+ use_scaled_pos_enc: true
221
+ encoder_normalize_before: true
222
+ decoder_normalize_before: true
223
+ reduction_factor: 1
224
+ init_type: xavier_uniform
225
+ init_enc_alpha: 1.0
226
+ init_dec_alpha: 1.0
227
+ transformer_enc_dropout_rate: 0.2
228
+ transformer_enc_positional_dropout_rate: 0.2
229
+ transformer_enc_attn_dropout_rate: 0.2
230
+ transformer_dec_dropout_rate: 0.2
231
+ transformer_dec_positional_dropout_rate: 0.2
232
+ transformer_dec_attn_dropout_rate: 0.2
233
+ pitch_predictor_layers: 5
234
+ pitch_predictor_chans: 256
235
+ pitch_predictor_kernel_size: 5
236
+ pitch_predictor_dropout: 0.5
237
+ pitch_embed_kernel_size: 1
238
+ pitch_embed_dropout: 0.0
239
+ stop_gradient_from_pitch_predictor: true
240
+ energy_predictor_layers: 2
241
+ energy_predictor_chans: 256
242
+ energy_predictor_kernel_size: 3
243
+ energy_predictor_dropout: 0.5
244
+ energy_embed_kernel_size: 1
245
+ energy_embed_dropout: 0.0
246
+ stop_gradient_from_energy_predictor: false
247
+ pitch_extract: dio
248
+ pitch_extract_conf:
249
+ fs: 22050
250
+ n_fft: 1024
251
+ hop_length: 256
252
+ f0max: 400
253
+ f0min: 80
254
+ reduction_factor: 1
255
+ pitch_normalize: global_mvn
256
+ pitch_normalize_conf:
257
+ stats_file: /speech/arun/released_models/tts/female/assamese/fastspeech2_hs/pitch_stats.npz
258
+ energy_extract: energy
259
+ energy_extract_conf:
260
+ fs: 22050
261
+ n_fft: 1024
262
+ hop_length: 256
263
+ win_length: 1024
264
+ reduction_factor: 1
265
+ energy_normalize: global_mvn
266
+ energy_normalize_conf:
267
+ stats_file: /speech/arun/released_models/tts/female/assamese/fastspeech2_hs/energy_stats.npz
268
+ required:
269
+ - output_dir
270
+ - token_list
271
+ version: 0.10.3a3
272
+ distributed: true
assamese/female/model/energy_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80879ba3818e253e2cbfb78176b51de0c247fdb17f07a6b2db730c9d0026f31e
3
+ size 770
assamese/female/model/feats_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5be4c43d477c15c04189297ad17d1e1c436f31aed3caff14cacebe4fc13308ab
3
+ size 1402
assamese/female/model/feats_type ADDED
@@ -0,0 +1 @@
 
 
1
+ raw
assamese/female/model/model.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff9efd82aacade9fcd3fc17e350147d3a272356b1f419872ac9fa525b991b578
3
+ size 148685818
assamese/female/model/pitch_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0453690cf54a373b7a39ccbc4557f8487496bca2290c32d684369fe05f4ed4fe
3
+ size 770
assamese/male/model/config.yaml ADDED
@@ -0,0 +1,272 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/tuning/train_fastspeech2.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ dry_run: false
5
+ iterator_type: sequence
6
+ output_dir: exp/tts_train_fastspeech2_raw_char_None
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: 4
14
+ dist_rank: 0
15
+ local_rank: 0
16
+ dist_master_addr: localhost
17
+ dist_master_port: 33317
18
+ dist_launcher: null
19
+ multiprocessing_distributed: true
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: 1000
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
+ grad_clip: 1.0
45
+ grad_clip_type: 2.0
46
+ grad_noise: false
47
+ accum_grad: 8
48
+ no_forward_run: false
49
+ resume: true
50
+ train_dtype: float32
51
+ use_amp: false
52
+ log_interval: null
53
+ use_tensorboard: true
54
+ use_wandb: false
55
+ wandb_project: null
56
+ wandb_id: null
57
+ wandb_entity: null
58
+ wandb_name: null
59
+ wandb_model_log_interval: -1
60
+ detect_anomaly: false
61
+ pretrain_path: null
62
+ init_param: []
63
+ ignore_init_mismatch: false
64
+ freeze_param: []
65
+ num_iters_per_epoch: 800
66
+ batch_size: 20
67
+ valid_batch_size: null
68
+ batch_bins: 3000000
69
+ valid_batch_bins: null
70
+ train_shape_file:
71
+ - exp/tts_stats_raw_char_None/train/text_shape.char
72
+ - exp/tts_stats_raw_char_None/train/speech_shape
73
+ valid_shape_file:
74
+ - exp/tts_stats_raw_char_None/valid/text_shape.char
75
+ - exp/tts_stats_raw_char_None/valid/speech_shape
76
+ batch_type: numel
77
+ valid_batch_type: null
78
+ fold_length:
79
+ - 150
80
+ - 204800
81
+ sort_in_batch: descending
82
+ sort_batch: descending
83
+ multiple_iterator: false
84
+ chunk_length: 500
85
+ chunk_shift_ratio: 0.5
86
+ num_cache_chunks: 1024
87
+ train_data_path_and_name_and_type:
88
+ - - dump/raw/tr_no_dev/text
89
+ - text
90
+ - text
91
+ - - duration_info/tr_no_dev/durations
92
+ - durations
93
+ - text_int
94
+ - - dump/raw/tr_no_dev/wav.scp
95
+ - speech
96
+ - sound
97
+ - - exp/tts_stats_raw_char_None/train/collect_feats/pitch.scp
98
+ - pitch
99
+ - npy
100
+ - - exp/tts_stats_raw_char_None/train/collect_feats/energy.scp
101
+ - energy
102
+ - npy
103
+ valid_data_path_and_name_and_type:
104
+ - - dump/raw/dev/text
105
+ - text
106
+ - text
107
+ - - duration_info/dev/durations
108
+ - durations
109
+ - text_int
110
+ - - dump/raw/dev/wav.scp
111
+ - speech
112
+ - sound
113
+ - - exp/tts_stats_raw_char_None/valid/collect_feats/pitch.scp
114
+ - pitch
115
+ - npy
116
+ - - exp/tts_stats_raw_char_None/valid/collect_feats/energy.scp
117
+ - energy
118
+ - npy
119
+ allow_variable_data_keys: false
120
+ max_cache_size: 0.0
121
+ max_cache_fd: 32
122
+ valid_max_cache_size: null
123
+ optim: adam
124
+ optim_conf:
125
+ lr: 1.0
126
+ scheduler: noamlr
127
+ scheduler_conf:
128
+ model_size: 384
129
+ warmup_steps: 4000
130
+ token_list:
131
+ - <blank>
132
+ - <unk>
133
+ - A
134
+ - a
135
+ - i
136
+ - ','
137
+ - r
138
+ - E
139
+ - k
140
+ - n
141
+ - t
142
+ - l
143
+ - b
144
+ - o
145
+ - u
146
+ - m
147
+ - y
148
+ - $
149
+ - .
150
+ - p
151
+ - h
152
+ - d
153
+ - s
154
+ - I
155
+ - g
156
+ - j
157
+ - ट
158
+ - c
159
+ - ऐ
160
+ - C
161
+ - ख
162
+ - w
163
+ - श
164
+ - B
165
+ - थ
166
+ - M
167
+ - ध
168
+ - ष
169
+ - ण
170
+ - ठ
171
+ - घ
172
+ - q
173
+ - P
174
+ - U
175
+ - ड
176
+ - ङ
177
+ - R
178
+ - औ
179
+ - ञ
180
+ - D
181
+ - ढ
182
+ - <sos/eos>
183
+ odim: null
184
+ model_conf: {}
185
+ use_preprocessor: true
186
+ token_type: char
187
+ bpemodel: null
188
+ non_linguistic_symbols: null
189
+ cleaner: null
190
+ g2p: g2p_en_no_space
191
+ feats_extract: fbank
192
+ feats_extract_conf:
193
+ n_fft: 1024
194
+ hop_length: 256
195
+ win_length: 1024
196
+ fs: 22050
197
+ fmin: 0
198
+ fmax: 8000
199
+ n_mels: 80
200
+ normalize: global_mvn
201
+ normalize_conf:
202
+ stats_file: /speech/arun/released_models/tts/male/assamese/fastspeech2_hs/feats_stats.npz
203
+ tts: fastspeech2
204
+ tts_conf:
205
+ adim: 384
206
+ aheads: 2
207
+ elayers: 4
208
+ eunits: 1536
209
+ dlayers: 4
210
+ dunits: 1536
211
+ positionwise_layer_type: conv1d
212
+ positionwise_conv_kernel_size: 3
213
+ duration_predictor_layers: 2
214
+ duration_predictor_chans: 256
215
+ duration_predictor_kernel_size: 3
216
+ postnet_layers: 5
217
+ postnet_filts: 5
218
+ postnet_chans: 256
219
+ use_masking: true
220
+ use_scaled_pos_enc: true
221
+ encoder_normalize_before: true
222
+ decoder_normalize_before: true
223
+ reduction_factor: 1
224
+ init_type: xavier_uniform
225
+ init_enc_alpha: 1.0
226
+ init_dec_alpha: 1.0
227
+ transformer_enc_dropout_rate: 0.2
228
+ transformer_enc_positional_dropout_rate: 0.2
229
+ transformer_enc_attn_dropout_rate: 0.2
230
+ transformer_dec_dropout_rate: 0.2
231
+ transformer_dec_positional_dropout_rate: 0.2
232
+ transformer_dec_attn_dropout_rate: 0.2
233
+ pitch_predictor_layers: 5
234
+ pitch_predictor_chans: 256
235
+ pitch_predictor_kernel_size: 5
236
+ pitch_predictor_dropout: 0.5
237
+ pitch_embed_kernel_size: 1
238
+ pitch_embed_dropout: 0.0
239
+ stop_gradient_from_pitch_predictor: true
240
+ energy_predictor_layers: 2
241
+ energy_predictor_chans: 256
242
+ energy_predictor_kernel_size: 3
243
+ energy_predictor_dropout: 0.5
244
+ energy_embed_kernel_size: 1
245
+ energy_embed_dropout: 0.0
246
+ stop_gradient_from_energy_predictor: false
247
+ pitch_extract: dio
248
+ pitch_extract_conf:
249
+ fs: 22050
250
+ n_fft: 1024
251
+ hop_length: 256
252
+ f0max: 350
253
+ f0min: 40
254
+ reduction_factor: 1
255
+ pitch_normalize: global_mvn
256
+ pitch_normalize_conf:
257
+ stats_file: /speech/arun/released_models/tts/male/assamese/fastspeech2_hs/pitch_stats.npz
258
+ energy_extract: energy
259
+ energy_extract_conf:
260
+ fs: 22050
261
+ n_fft: 1024
262
+ hop_length: 256
263
+ win_length: 1024
264
+ reduction_factor: 1
265
+ energy_normalize: global_mvn
266
+ energy_normalize_conf:
267
+ stats_file: /speech/arun/released_models/tts/male/assamese/fastspeech2_hs/energy_stats.npz
268
+ required:
269
+ - output_dir
270
+ - token_list
271
+ version: 0.10.3a3
272
+ distributed: true
assamese/male/model/energy_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66cf788e284146143ddc0d81738e2a4cece116bcfa0ca105a518717fee76f681
3
+ size 770
assamese/male/model/feats_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69d6c9208c60aa8fbfce176574858e5fa363ba048fb2418bae339a990592f2c4
3
+ size 1402
assamese/male/model/feats_type ADDED
@@ -0,0 +1 @@
 
 
1
+ raw
assamese/male/model/model.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61e7dd206b37613770973d840f31906e19e51bb6ac771a40bbdd41afa3bce78a
3
+ size 148685818
assamese/male/model/pitch_stats.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:216a0a305e14a24fb5b1a90dc2dd44e31fe4f33efc022d1298bdd7d3890c7f2e
3
+ size 770