rakhlin commited on
Commit
d7c9865
·
1 Parent(s): 90ecf69

Delete tts/tts_models--en--ljspeech--fast_pitch

Browse files
tts/tts_models--en--ljspeech--fast_pitch/config.json DELETED
@@ -1,163 +0,0 @@
1
- {
2
- "model": "fast_pitch",
3
- "run_name": "fast_pitch_ljspeech",
4
- "run_description": "",
5
- "epochs": 1000,
6
- "batch_size": 32,
7
- "eval_batch_size": 16,
8
- "mixed_precision": false,
9
- "scheduler_after_epoch": false,
10
- "run_eval": true,
11
- "test_delay_epochs": -1,
12
- "print_eval": false,
13
- "dashboard_logger": "tensorboard",
14
- "print_step": 50,
15
- "plot_step": 100,
16
- "model_param_stats": false,
17
- "project_name": null,
18
- "log_model_step": null,
19
- "wandb_entity": null,
20
- "save_step": 10000,
21
- "checkpoint": true,
22
- "keep_all_best": false,
23
- "keep_after": 10000,
24
- "num_loader_workers": 0,
25
- "num_eval_loader_workers": 4,
26
- "use_noise_augment": false,
27
- "output_path": "/home/ubuntu/TTS/recipes/ljspeech/fast_pitch",
28
- "distributed_backend": "nccl",
29
- "distributed_url": "tcp://localhost:54321",
30
- "audio": {
31
- "fft_size": 1024,
32
- "win_length": 1024,
33
- "hop_length": 256,
34
- "frame_shift_ms": null,
35
- "frame_length_ms": null,
36
- "stft_pad_mode": "reflect",
37
- "sample_rate": 22050,
38
- "resample": false,
39
- "preemphasis": 0.0,
40
- "ref_level_db": 20,
41
- "do_sound_norm": false,
42
- "log_func": "np.log",
43
- "do_trim_silence": true,
44
- "trim_db": 60.0,
45
- "power": 1.5,
46
- "griffin_lim_iters": 60,
47
- "num_mels": 80,
48
- "mel_fmin": 0.0,
49
- "mel_fmax": 8000,
50
- "spec_gain": 1.0,
51
- "do_amp_to_db_linear": true,
52
- "do_amp_to_db_mel": true,
53
- "signal_norm": false,
54
- "min_level_db": -100,
55
- "symmetric_norm": true,
56
- "max_norm": 4.0,
57
- "clip_norm": true,
58
- "stats_path": null
59
- },
60
- "use_phonemes": true,
61
- "phonemizer": "gruut",
62
- "phoneme_language": "en-us",
63
- "compute_input_seq_cache": true,
64
- "text_cleaner": "english_cleaners",
65
- "enable_eos_bos_chars": false,
66
- "test_sentences_file": "",
67
- "phoneme_cache_path": "/home/ubuntu/TTS/recipes/ljspeech/fast_pitch/phoneme_cache",
68
- "characters": {
69
- "pad": "_",
70
- "eos": "~",
71
- "bos": "^",
72
- "characters": "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!'(),-.:;? ",
73
- "punctuations": "!'(),-.:;? ",
74
- "phonemes": "iy\u0268\u0289\u026fu\u026a\u028f\u028ae\u00f8\u0258\u0259\u0275\u0264o\u025b\u0153\u025c\u025e\u028c\u0254\u00e6\u0250a\u0276\u0251\u0252\u1d7b\u0298\u0253\u01c0\u0257\u01c3\u0284\u01c2\u0260\u01c1\u029bpbtd\u0288\u0256c\u025fk\u0261q\u0262\u0294\u0274\u014b\u0272\u0273n\u0271m\u0299r\u0280\u2c71\u027e\u027d\u0278\u03b2fv\u03b8\u00f0sz\u0283\u0292\u0282\u0290\u00e7\u029dx\u0263\u03c7\u0281\u0127\u0295h\u0266\u026c\u026e\u028b\u0279\u027bj\u0270l\u026d\u028e\u029f\u02c8\u02cc\u02d0\u02d1\u028dw\u0265\u029c\u02a2\u02a1\u0255\u0291\u027a\u0267\u02b2\u025a\u02de\u026b"
75
- },
76
- "batch_group_size": 0,
77
- "loss_masking": null,
78
- "sort_by_audio_len": true,
79
- "min_seq_len": 13,
80
- "max_seq_len": 500000,
81
- "compute_f0": true,
82
- "compute_linear_spec": false,
83
- "add_blank": false,
84
- "datasets": [
85
- {
86
- "name": "ljspeech",
87
- "path": "/home/ubuntu/TTS/recipes/ljspeech/fast_pitch/../LJSpeech-1.1/",
88
- "meta_file_train": "metadata.csv",
89
- "ununsed_speakers": null,
90
- "meta_file_val": "",
91
- "meta_file_attn_mask": ""
92
- }
93
- ],
94
- "optimizer": "Adam",
95
- "optimizer_params": {
96
- "betas": [
97
- 0.9,
98
- 0.998
99
- ],
100
- "weight_decay": 1e-06
101
- },
102
- "lr_scheduler": "NoamLR",
103
- "lr_scheduler_params": {
104
- "warmup_steps": 4000
105
- },
106
- "test_sentences": [
107
- "It took me quite a long time to develop a voice, and now that I have it I'm not going to be silent.",
108
- "Be a voice, not an echo.",
109
- "I'm sorry Dave. I'm afraid I can't do that.",
110
- "This cake is great. It's so delicious and moist.",
111
- "Prior to November 22, 1963."
112
- ],
113
- "use_speaker_embedding": false,
114
- "use_d_vector_file": false,
115
- "d_vector_dim": 0,
116
- "model_args": {
117
- "num_chars": null,
118
- "out_channels": 80,
119
- "hidden_channels": 384,
120
- "num_speakers": 0,
121
- "duration_predictor_hidden_channels": 256,
122
- "duration_predictor_kernel_size": 3,
123
- "duration_predictor_dropout_p": 0.1,
124
- "pitch_predictor_hidden_channels": 256,
125
- "pitch_predictor_kernel_size": 3,
126
- "pitch_predictor_dropout_p": 0.1,
127
- "pitch_embedding_kernel_size": 3,
128
- "positional_encoding": true,
129
- "poisitonal_encoding_use_scale": true,
130
- "length_scale": 1,
131
- "encoder_type": "fftransformer",
132
- "encoder_params": {
133
- "hidden_channels_ffn": 1024,
134
- "num_heads": 1,
135
- "num_layers": 6,
136
- "dropout_p": 0.1
137
- },
138
- "decoder_type": "fftransformer",
139
- "decoder_params": {
140
- "hidden_channels_ffn": 1024,
141
- "num_heads": 1,
142
- "num_layers": 6,
143
- "dropout_p": 0.1
144
- },
145
- "use_d_vector": false,
146
- "d_vector_dim": 0,
147
- "detach_duration_predictor": false,
148
- "max_duration": 75,
149
- "use_aligner": true
150
- },
151
- "d_vector_file": false,
152
- "lr": 0.0001,
153
- "grad_clip": 5.0,
154
- "ssim_loss_alpha": 1.0,
155
- "dur_loss_alpha": 1.0,
156
- "spec_loss_alpha": 1.0,
157
- "pitch_loss_alpha": 1.0,
158
- "aligner_loss_alpha": 1.0,
159
- "binary_align_loss_alpha": 1.0,
160
- "binary_align_loss_start_step": 20000,
161
- "r": 1,
162
- "f0_cache_path": "/home/ubuntu/TTS/recipes/ljspeech/fast_pitch/f0_cache"
163
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
tts/tts_models--en--ljspeech--fast_pitch/model_file.pth DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:1779ef4ef9f9f3c016efee5925c0742393eb7c7183f6daae1928b88cbef294b8
3
- size 458065485