Commit
·
9b9e950
1
Parent(s):
d647987
Re-add model files after reset
Browse files- pupucodec/args.json +270 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/ckpts.json +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model.safetensors +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_1.safetensors +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_2.safetensors +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_3.safetensors +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_4.safetensors +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/optimizer.bin +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/optimizer_1.bin +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/random_states_0.pkl +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/scheduler.bin +3 -0
- pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/scheduler_1.bin +3 -0
- pupucodec_large/args.json +270 -0
- pupucodec_large/checkpoint/epoch-0053_step-2349317_loss-57.300222 +3 -0
- pupuvocoder/args.json +289 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/ckpts.json +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model.safetensors +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_1.safetensors +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_2.safetensors +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_3.safetensors +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_4.safetensors +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/optimizer.bin +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/optimizer_1.bin +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/random_states_0.pkl +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/scheduler.bin +3 -0
- pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/scheduler_1.bin +3 -0
- pupuvocoder_large/args.json +286 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/ckpts.json +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model.safetensors +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_1.safetensors +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_2.safetensors +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_3.safetensors +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_4.safetensors +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/optimizer.bin +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/optimizer_1.bin +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/random_states_0.pkl +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/scheduler.bin +3 -0
- pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/scheduler_1.bin +3 -0
pupucodec/args.json
ADDED
|
@@ -0,0 +1,270 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_config": "egs/afgen/exp_config_afcodec_base.json",
|
| 3 |
+
"dataset": [
|
| 4 |
+
"afgen_all_data",
|
| 5 |
+
],
|
| 6 |
+
"exp_name": "afcodec",
|
| 7 |
+
"inference": {
|
| 8 |
+
"batch_size": 1,
|
| 9 |
+
},
|
| 10 |
+
"log_dir": "/home/vocod/experiments/afgen",
|
| 11 |
+
"model": {
|
| 12 |
+
"afcodec": {
|
| 13 |
+
"codebook_dim": 8,
|
| 14 |
+
"codebook_size": 1024,
|
| 15 |
+
"decoder_dim": 512,
|
| 16 |
+
"decoder_rates": [
|
| 17 |
+
8,
|
| 18 |
+
8,
|
| 19 |
+
2,
|
| 20 |
+
2,
|
| 21 |
+
2,
|
| 22 |
+
],
|
| 23 |
+
"encoder_dim": 32,
|
| 24 |
+
"encoder_rates": [
|
| 25 |
+
2,
|
| 26 |
+
2,
|
| 27 |
+
2,
|
| 28 |
+
8,
|
| 29 |
+
8,
|
| 30 |
+
],
|
| 31 |
+
"n_codebooks": 9,
|
| 32 |
+
"quantizer_dropout": 0.5,
|
| 33 |
+
},
|
| 34 |
+
"discriminators": [
|
| 35 |
+
"msd",
|
| 36 |
+
"mpd",
|
| 37 |
+
"mssbstftd",
|
| 38 |
+
"mssbcqtd",
|
| 39 |
+
],
|
| 40 |
+
"generator": "afcodec",
|
| 41 |
+
"mpd": {
|
| 42 |
+
"discriminator_channel_mult_factor": 1,
|
| 43 |
+
"mpd_reshapes": [
|
| 44 |
+
2,
|
| 45 |
+
3,
|
| 46 |
+
5,
|
| 47 |
+
7,
|
| 48 |
+
11,
|
| 49 |
+
17,
|
| 50 |
+
23,
|
| 51 |
+
37,
|
| 52 |
+
],
|
| 53 |
+
"use_spectral_norm": false,
|
| 54 |
+
},
|
| 55 |
+
"mssbcqtd": {
|
| 56 |
+
"bins_per_octaves": [
|
| 57 |
+
24,
|
| 58 |
+
36,
|
| 59 |
+
48,
|
| 60 |
+
],
|
| 61 |
+
"dilations": [
|
| 62 |
+
1,
|
| 63 |
+
2,
|
| 64 |
+
4,
|
| 65 |
+
],
|
| 66 |
+
"filters": 32,
|
| 67 |
+
"filters_scale": 1,
|
| 68 |
+
"hop_lengths": [
|
| 69 |
+
1024,
|
| 70 |
+
512,
|
| 71 |
+
512,
|
| 72 |
+
],
|
| 73 |
+
"in_channels": 1,
|
| 74 |
+
"max_filters": 1024,
|
| 75 |
+
"n_octaves": [
|
| 76 |
+
10,
|
| 77 |
+
10,
|
| 78 |
+
10,
|
| 79 |
+
],
|
| 80 |
+
"out_channels": 1,
|
| 81 |
+
},
|
| 82 |
+
"msstftd": {
|
| 83 |
+
"filters": 32,
|
| 84 |
+
},
|
| 85 |
+
},
|
| 86 |
+
"model_type": "AFGenCodec",
|
| 87 |
+
"preprocess": {
|
| 88 |
+
"align_mel_duration": false,
|
| 89 |
+
"audio_dir": "audios",
|
| 90 |
+
"bits": 8,
|
| 91 |
+
"contentvec_dir": "contentvec",
|
| 92 |
+
"cut_mel_frame": 128,
|
| 93 |
+
"data_augment": false,
|
| 94 |
+
"dur_dir": "durs",
|
| 95 |
+
"duration_dir": "duration",
|
| 96 |
+
"emo2id": "emo2id.json",
|
| 97 |
+
"energy_dir": "energys",
|
| 98 |
+
"energy_extract_mode": "from_mel",
|
| 99 |
+
"energy_norm": false,
|
| 100 |
+
"energy_remove_outlier": false,
|
| 101 |
+
"extract_acoustic_token": false,
|
| 102 |
+
"extract_amplitude_phase": false,
|
| 103 |
+
"extract_audio": true,
|
| 104 |
+
"extract_contentvec_feature": false,
|
| 105 |
+
"extract_duration": false,
|
| 106 |
+
"extract_energy": false,
|
| 107 |
+
"extract_label": false,
|
| 108 |
+
"extract_linear_spec": false,
|
| 109 |
+
"extract_mcep": false,
|
| 110 |
+
"extract_mel": false,
|
| 111 |
+
"extract_mert_feature": false,
|
| 112 |
+
"extract_one_hot": false,
|
| 113 |
+
"extract_phone": false,
|
| 114 |
+
"extract_pitch": true,
|
| 115 |
+
"extract_uv": true,
|
| 116 |
+
"extract_wenet_feature": false,
|
| 117 |
+
"extract_whisper_feature": false,
|
| 118 |
+
"f0_max": 1975.5,
|
| 119 |
+
"f0_min": 32.7,
|
| 120 |
+
"file_lst": "file.lst",
|
| 121 |
+
"fmax": 22050,
|
| 122 |
+
"fmin": 0,
|
| 123 |
+
"hop_size": 512,
|
| 124 |
+
"imaginary_dir": "imaginarys",
|
| 125 |
+
"is_mu_law": false,
|
| 126 |
+
"lab_dir": "labs",
|
| 127 |
+
"label_dir": "labels",
|
| 128 |
+
"lexicon_path": "./text/lexicon/librispeech-lexicon.txt",
|
| 129 |
+
"linear_dir": "linears",
|
| 130 |
+
"log_amplitude_dir": "log_amplitudes",
|
| 131 |
+
"mcep_dir": "mcep",
|
| 132 |
+
"mel_dir": "mels",
|
| 133 |
+
"mel_extract_mode": "",
|
| 134 |
+
"mel_min_max_norm": false,
|
| 135 |
+
"min_level_db": -115,
|
| 136 |
+
"n_fft": 2048,
|
| 137 |
+
"n_mel": 128,
|
| 138 |
+
"num_silent_frames": 8,
|
| 139 |
+
"phase_dir": "phases",
|
| 140 |
+
"phone_dir": "phones",
|
| 141 |
+
"phone_energy_dir": "phone_energys",
|
| 142 |
+
"phone_extractor": "espeak",
|
| 143 |
+
"phone_pitch_dir": "phone_pitches",
|
| 144 |
+
"phone_seq_file": "phone_seq_file",
|
| 145 |
+
"pitch_bin": 256,
|
| 146 |
+
"pitch_dir": "pitches",
|
| 147 |
+
"pitch_extractor": "rmvpe",
|
| 148 |
+
"pitch_max": 1100.0,
|
| 149 |
+
"pitch_min": 50.0,
|
| 150 |
+
"pitch_norm": false,
|
| 151 |
+
"pitch_remove_outlier": false,
|
| 152 |
+
"processed_dir": "data/",
|
| 153 |
+
"raw_data": "raw_data",
|
| 154 |
+
"real_dir": "reals",
|
| 155 |
+
"ref_level_db": 20,
|
| 156 |
+
"sample_rate": 44100,
|
| 157 |
+
"spk2id": "singers.json",
|
| 158 |
+
"symbols_dict": "symbols.dict",
|
| 159 |
+
"train_file": "train.json",
|
| 160 |
+
"trim_fft_size": 512,
|
| 161 |
+
"trim_hop_size": 128,
|
| 162 |
+
"trim_silence": false,
|
| 163 |
+
"trim_top_db": 30,
|
| 164 |
+
"trimmed_wav_dir": "trimmed_wavs",
|
| 165 |
+
"use_amplitude_phase": false,
|
| 166 |
+
"use_audio": true,
|
| 167 |
+
"use_dur": false,
|
| 168 |
+
"use_emoid": false,
|
| 169 |
+
"use_frame_duration": false,
|
| 170 |
+
"use_frame_energy": false,
|
| 171 |
+
"use_frame_pitch": true,
|
| 172 |
+
"use_lab": false,
|
| 173 |
+
"use_label": false,
|
| 174 |
+
"use_linear": false,
|
| 175 |
+
"use_log_scale_energy": false,
|
| 176 |
+
"use_log_scale_pitch": false,
|
| 177 |
+
"use_mel": false,
|
| 178 |
+
"use_min_max_norm_mel": false,
|
| 179 |
+
"use_one_hot": false,
|
| 180 |
+
"use_phn_seq": false,
|
| 181 |
+
"use_phone": false,
|
| 182 |
+
"use_phone_duration": false,
|
| 183 |
+
"use_phone_energy": false,
|
| 184 |
+
"use_phone_pitch": false,
|
| 185 |
+
"use_spkid": false,
|
| 186 |
+
"use_text": false,
|
| 187 |
+
"use_uv": true,
|
| 188 |
+
"use_wav": false,
|
| 189 |
+
"use_wenet": false,
|
| 190 |
+
"utt2emo": "utt2emo",
|
| 191 |
+
"utt2spk": "utt2spk",
|
| 192 |
+
"uv_dir": "uvs",
|
| 193 |
+
"valid_file": "valid.json",
|
| 194 |
+
"wav_dir": "wavs",
|
| 195 |
+
"wenet_dir": "wenet",
|
| 196 |
+
"win_size": 2048,
|
| 197 |
+
},
|
| 198 |
+
"supported_model_type": [
|
| 199 |
+
"GANVocoder",
|
| 200 |
+
"Fastspeech2",
|
| 201 |
+
"DiffSVC",
|
| 202 |
+
"Transformer",
|
| 203 |
+
"EDM",
|
| 204 |
+
"Autotune",
|
| 205 |
+
"CD",
|
| 206 |
+
],
|
| 207 |
+
"task_type": "afgen",
|
| 208 |
+
"train": {
|
| 209 |
+
"adamw": {
|
| 210 |
+
"adam_b1": 0.8,
|
| 211 |
+
"adam_b2": 0.99,
|
| 212 |
+
"lr": 0.0001,
|
| 213 |
+
},
|
| 214 |
+
"batch_size": 16,
|
| 215 |
+
"criterions": [
|
| 216 |
+
"feature",
|
| 217 |
+
"discriminator",
|
| 218 |
+
"generator",
|
| 219 |
+
"multimel",
|
| 220 |
+
"codebook",
|
| 221 |
+
"commitment",
|
| 222 |
+
],
|
| 223 |
+
"dataloader": {
|
| 224 |
+
"num_worker": 32,
|
| 225 |
+
"persistent_workers": true,
|
| 226 |
+
"pin_memory": true,
|
| 227 |
+
"prefetch_factor": 4,
|
| 228 |
+
},
|
| 229 |
+
"ddp": true,
|
| 230 |
+
"exponential_lr": {
|
| 231 |
+
"lr_decay": 0.9999996,
|
| 232 |
+
},
|
| 233 |
+
"freeze_step": 0,
|
| 234 |
+
"gradient_accumulation_step": 1,
|
| 235 |
+
"keep_checkpoint_max": 5,
|
| 236 |
+
"keep_last": [
|
| 237 |
+
3,
|
| 238 |
+
-1,
|
| 239 |
+
],
|
| 240 |
+
"max_epoch": 1000000,
|
| 241 |
+
"max_steps": 1000000,
|
| 242 |
+
"multi_speaker_training": false,
|
| 243 |
+
"optimizer": "AdamW",
|
| 244 |
+
"random_seed": 114514,
|
| 245 |
+
"reducelronplateau": {
|
| 246 |
+
"factor": 0.8,
|
| 247 |
+
"min_lr": 0.0001,
|
| 248 |
+
"patience": 10,
|
| 249 |
+
},
|
| 250 |
+
"run_eval": [
|
| 251 |
+
true,
|
| 252 |
+
],
|
| 253 |
+
"sampler": {
|
| 254 |
+
"drop_last": true,
|
| 255 |
+
"holistic_shuffle": true,
|
| 256 |
+
},
|
| 257 |
+
"save_checkpoint_stride": [
|
| 258 |
+
1,
|
| 259 |
+
],
|
| 260 |
+
"save_checkpoints_steps": 10000,
|
| 261 |
+
"save_summary_steps": 500,
|
| 262 |
+
"scheduler": "ReduceLROnPlateau",
|
| 263 |
+
"total_training_steps": 50000,
|
| 264 |
+
"tracker": [
|
| 265 |
+
"tensorboard",
|
| 266 |
+
],
|
| 267 |
+
"valid_interval": 10000,
|
| 268 |
+
},
|
| 269 |
+
"use_custom_dataset": [],
|
| 270 |
+
}
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/ckpts.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
[]
|
| 3 |
+
]
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e99d196175c36eb76bb484dedf0735d9d2109d1fe8418a2cf4518c6d7ce4d794
|
| 3 |
+
size 127943572
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_1.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:45e01c794b9cce470553daa86860516a17629cc01195ad92fbabab7c34d9f539
|
| 3 |
+
size 118557612
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_2.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b66c8863bec06bf82ae0bf671b1aa3f09cd97f2863d4043704a8de94cf65f3d
|
| 3 |
+
size 263092072
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71abf18c28864cf67012f78e93c05944a9fde5cbbcd2a98b02443d25c804b6a2
|
| 3 |
+
size 5695720
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/model_4.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18aba7e99c90166928bac230681d34d0e9e67a45e49bc92adfa154d1c02f27bf
|
| 3 |
+
size 1560888
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/optimizer.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d69d069333d18162104bf55711403dc4ad1e2e1d3da39ca8f7ccb40c68cb7043
|
| 3 |
+
size 256144660
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/optimizer_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e93ff06259cd8bdd2da1e756f16abc2589cf45c36d946727c15f01ae812e0008
|
| 3 |
+
size 777186848
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/random_states_0.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0ad89a22587db88134843f6fe38faaeb909b4c43ba62d597ed1ac9ab1a15a5e
|
| 3 |
+
size 16513
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/scheduler.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf58b70e9b7d08140cc0ae8cbb43aabd0417766651455c7ce5d46c8ff150fe8b
|
| 3 |
+
size 1401
|
pupucodec/checkpoint/epoch-0102_step-3038789_loss-76.647869/scheduler_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66df678ca2b6e4ebcb98c1ad1ab38dc119808c8bee4265a823817957dfaf868b
|
| 3 |
+
size 1477
|
pupucodec_large/args.json
ADDED
|
@@ -0,0 +1,270 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_config": "egs/afgen/exp_config_afcodec_base.json",
|
| 3 |
+
"dataset": [
|
| 4 |
+
"afgen_all_data",
|
| 5 |
+
],
|
| 6 |
+
"exp_name": "afcodec_large",
|
| 7 |
+
"inference": {
|
| 8 |
+
"batch_size": 1,
|
| 9 |
+
},
|
| 10 |
+
"log_dir": "/home/vocod/experiments/afgen",
|
| 11 |
+
"model": {
|
| 12 |
+
"afcodec": {
|
| 13 |
+
"codebook_dim": 8,
|
| 14 |
+
"codebook_size": 1024,
|
| 15 |
+
"decoder_dim": 1536,
|
| 16 |
+
"decoder_rates": [
|
| 17 |
+
8,
|
| 18 |
+
8,
|
| 19 |
+
2,
|
| 20 |
+
2,
|
| 21 |
+
2,
|
| 22 |
+
],
|
| 23 |
+
"encoder_dim": 48,
|
| 24 |
+
"encoder_rates": [
|
| 25 |
+
2,
|
| 26 |
+
2,
|
| 27 |
+
2,
|
| 28 |
+
8,
|
| 29 |
+
8,
|
| 30 |
+
],
|
| 31 |
+
"n_codebooks": 9,
|
| 32 |
+
"quantizer_dropout": 0.5,
|
| 33 |
+
},
|
| 34 |
+
"discriminators": [
|
| 35 |
+
"msd",
|
| 36 |
+
"mpd",
|
| 37 |
+
"mssbstftd",
|
| 38 |
+
"mssbcqtd",
|
| 39 |
+
],
|
| 40 |
+
"generator": "afcodec",
|
| 41 |
+
"mpd": {
|
| 42 |
+
"discriminator_channel_mult_factor": 1,
|
| 43 |
+
"mpd_reshapes": [
|
| 44 |
+
2,
|
| 45 |
+
3,
|
| 46 |
+
5,
|
| 47 |
+
7,
|
| 48 |
+
11,
|
| 49 |
+
17,
|
| 50 |
+
23,
|
| 51 |
+
37,
|
| 52 |
+
],
|
| 53 |
+
"use_spectral_norm": false,
|
| 54 |
+
},
|
| 55 |
+
"mssbcqtd": {
|
| 56 |
+
"bins_per_octaves": [
|
| 57 |
+
24,
|
| 58 |
+
36,
|
| 59 |
+
48,
|
| 60 |
+
],
|
| 61 |
+
"dilations": [
|
| 62 |
+
1,
|
| 63 |
+
2,
|
| 64 |
+
4,
|
| 65 |
+
],
|
| 66 |
+
"filters": 32,
|
| 67 |
+
"filters_scale": 1,
|
| 68 |
+
"hop_lengths": [
|
| 69 |
+
1024,
|
| 70 |
+
512,
|
| 71 |
+
512,
|
| 72 |
+
],
|
| 73 |
+
"in_channels": 1,
|
| 74 |
+
"max_filters": 1024,
|
| 75 |
+
"n_octaves": [
|
| 76 |
+
10,
|
| 77 |
+
10,
|
| 78 |
+
10,
|
| 79 |
+
],
|
| 80 |
+
"out_channels": 1,
|
| 81 |
+
},
|
| 82 |
+
"msstftd": {
|
| 83 |
+
"filters": 32,
|
| 84 |
+
},
|
| 85 |
+
},
|
| 86 |
+
"model_type": "AFGenCodec",
|
| 87 |
+
"preprocess": {
|
| 88 |
+
"align_mel_duration": false,
|
| 89 |
+
"audio_dir": "audios",
|
| 90 |
+
"bits": 8,
|
| 91 |
+
"contentvec_dir": "contentvec",
|
| 92 |
+
"cut_mel_frame": 128,
|
| 93 |
+
"data_augment": false,
|
| 94 |
+
"dur_dir": "durs",
|
| 95 |
+
"duration_dir": "duration",
|
| 96 |
+
"emo2id": "emo2id.json",
|
| 97 |
+
"energy_dir": "energys",
|
| 98 |
+
"energy_extract_mode": "from_mel",
|
| 99 |
+
"energy_norm": false,
|
| 100 |
+
"energy_remove_outlier": false,
|
| 101 |
+
"extract_acoustic_token": false,
|
| 102 |
+
"extract_amplitude_phase": false,
|
| 103 |
+
"extract_audio": true,
|
| 104 |
+
"extract_contentvec_feature": false,
|
| 105 |
+
"extract_duration": false,
|
| 106 |
+
"extract_energy": false,
|
| 107 |
+
"extract_label": false,
|
| 108 |
+
"extract_linear_spec": false,
|
| 109 |
+
"extract_mcep": false,
|
| 110 |
+
"extract_mel": false,
|
| 111 |
+
"extract_mert_feature": false,
|
| 112 |
+
"extract_one_hot": false,
|
| 113 |
+
"extract_phone": false,
|
| 114 |
+
"extract_pitch": true,
|
| 115 |
+
"extract_uv": true,
|
| 116 |
+
"extract_wenet_feature": false,
|
| 117 |
+
"extract_whisper_feature": false,
|
| 118 |
+
"f0_max": 1975.5,
|
| 119 |
+
"f0_min": 32.7,
|
| 120 |
+
"file_lst": "file.lst",
|
| 121 |
+
"fmax": 22050,
|
| 122 |
+
"fmin": 0,
|
| 123 |
+
"hop_size": 512,
|
| 124 |
+
"imaginary_dir": "imaginarys",
|
| 125 |
+
"is_mu_law": false,
|
| 126 |
+
"lab_dir": "labs",
|
| 127 |
+
"label_dir": "labels",
|
| 128 |
+
"lexicon_path": "./text/lexicon/librispeech-lexicon.txt",
|
| 129 |
+
"linear_dir": "linears",
|
| 130 |
+
"log_amplitude_dir": "log_amplitudes",
|
| 131 |
+
"mcep_dir": "mcep",
|
| 132 |
+
"mel_dir": "mels",
|
| 133 |
+
"mel_extract_mode": "",
|
| 134 |
+
"mel_min_max_norm": false,
|
| 135 |
+
"min_level_db": -115,
|
| 136 |
+
"n_fft": 2048,
|
| 137 |
+
"n_mel": 128,
|
| 138 |
+
"num_silent_frames": 8,
|
| 139 |
+
"phase_dir": "phases",
|
| 140 |
+
"phone_dir": "phones",
|
| 141 |
+
"phone_energy_dir": "phone_energys",
|
| 142 |
+
"phone_extractor": "espeak",
|
| 143 |
+
"phone_pitch_dir": "phone_pitches",
|
| 144 |
+
"phone_seq_file": "phone_seq_file",
|
| 145 |
+
"pitch_bin": 256,
|
| 146 |
+
"pitch_dir": "pitches",
|
| 147 |
+
"pitch_extractor": "rmvpe",
|
| 148 |
+
"pitch_max": 1100.0,
|
| 149 |
+
"pitch_min": 50.0,
|
| 150 |
+
"pitch_norm": false,
|
| 151 |
+
"pitch_remove_outlier": false,
|
| 152 |
+
"processed_dir": "data/",
|
| 153 |
+
"raw_data": "raw_data",
|
| 154 |
+
"real_dir": "reals",
|
| 155 |
+
"ref_level_db": 20,
|
| 156 |
+
"sample_rate": 44100,
|
| 157 |
+
"spk2id": "singers.json",
|
| 158 |
+
"symbols_dict": "symbols.dict",
|
| 159 |
+
"train_file": "train.json",
|
| 160 |
+
"trim_fft_size": 512,
|
| 161 |
+
"trim_hop_size": 128,
|
| 162 |
+
"trim_silence": false,
|
| 163 |
+
"trim_top_db": 30,
|
| 164 |
+
"trimmed_wav_dir": "trimmed_wavs",
|
| 165 |
+
"use_amplitude_phase": false,
|
| 166 |
+
"use_audio": true,
|
| 167 |
+
"use_dur": false,
|
| 168 |
+
"use_emoid": false,
|
| 169 |
+
"use_frame_duration": false,
|
| 170 |
+
"use_frame_energy": false,
|
| 171 |
+
"use_frame_pitch": true,
|
| 172 |
+
"use_lab": false,
|
| 173 |
+
"use_label": false,
|
| 174 |
+
"use_linear": false,
|
| 175 |
+
"use_log_scale_energy": false,
|
| 176 |
+
"use_log_scale_pitch": false,
|
| 177 |
+
"use_mel": false,
|
| 178 |
+
"use_min_max_norm_mel": false,
|
| 179 |
+
"use_one_hot": false,
|
| 180 |
+
"use_phn_seq": false,
|
| 181 |
+
"use_phone": false,
|
| 182 |
+
"use_phone_duration": false,
|
| 183 |
+
"use_phone_energy": false,
|
| 184 |
+
"use_phone_pitch": false,
|
| 185 |
+
"use_spkid": false,
|
| 186 |
+
"use_text": false,
|
| 187 |
+
"use_uv": true,
|
| 188 |
+
"use_wav": false,
|
| 189 |
+
"use_wenet": false,
|
| 190 |
+
"utt2emo": "utt2emo",
|
| 191 |
+
"utt2spk": "utt2spk",
|
| 192 |
+
"uv_dir": "uvs",
|
| 193 |
+
"valid_file": "valid.json",
|
| 194 |
+
"wav_dir": "wavs",
|
| 195 |
+
"wenet_dir": "wenet",
|
| 196 |
+
"win_size": 2048,
|
| 197 |
+
},
|
| 198 |
+
"supported_model_type": [
|
| 199 |
+
"GANVocoder",
|
| 200 |
+
"Fastspeech2",
|
| 201 |
+
"DiffSVC",
|
| 202 |
+
"Transformer",
|
| 203 |
+
"EDM",
|
| 204 |
+
"Autotune",
|
| 205 |
+
"CD",
|
| 206 |
+
],
|
| 207 |
+
"task_type": "afgen",
|
| 208 |
+
"train": {
|
| 209 |
+
"adamw": {
|
| 210 |
+
"adam_b1": 0.8,
|
| 211 |
+
"adam_b2": 0.99,
|
| 212 |
+
"lr": 0.0001,
|
| 213 |
+
},
|
| 214 |
+
"batch_size": 8,
|
| 215 |
+
"criterions": [
|
| 216 |
+
"feature",
|
| 217 |
+
"discriminator",
|
| 218 |
+
"generator",
|
| 219 |
+
"multimel",
|
| 220 |
+
"codebook",
|
| 221 |
+
"commitment",
|
| 222 |
+
],
|
| 223 |
+
"dataloader": {
|
| 224 |
+
"num_worker": 32,
|
| 225 |
+
"persistent_workers": true,
|
| 226 |
+
"pin_memory": true,
|
| 227 |
+
"prefetch_factor": 4,
|
| 228 |
+
},
|
| 229 |
+
"ddp": true,
|
| 230 |
+
"exponential_lr": {
|
| 231 |
+
"lr_decay": 0.9999996,
|
| 232 |
+
},
|
| 233 |
+
"freeze_step": 0,
|
| 234 |
+
"gradient_accumulation_step": 1,
|
| 235 |
+
"keep_checkpoint_max": 5,
|
| 236 |
+
"keep_last": [
|
| 237 |
+
3,
|
| 238 |
+
-1,
|
| 239 |
+
],
|
| 240 |
+
"max_epoch": 1000000,
|
| 241 |
+
"max_steps": 1000000,
|
| 242 |
+
"multi_speaker_training": false,
|
| 243 |
+
"optimizer": "AdamW",
|
| 244 |
+
"random_seed": 114514,
|
| 245 |
+
"reducelronplateau": {
|
| 246 |
+
"factor": 0.8,
|
| 247 |
+
"min_lr": 0.0001,
|
| 248 |
+
"patience": 10,
|
| 249 |
+
},
|
| 250 |
+
"run_eval": [
|
| 251 |
+
true,
|
| 252 |
+
],
|
| 253 |
+
"sampler": {
|
| 254 |
+
"drop_last": true,
|
| 255 |
+
"holistic_shuffle": true,
|
| 256 |
+
},
|
| 257 |
+
"save_checkpoint_stride": [
|
| 258 |
+
1,
|
| 259 |
+
],
|
| 260 |
+
"save_checkpoints_steps": 10000,
|
| 261 |
+
"save_summary_steps": 500,
|
| 262 |
+
"scheduler": "ReduceLROnPlateau",
|
| 263 |
+
"total_training_steps": 50000,
|
| 264 |
+
"tracker": [
|
| 265 |
+
"tensorboard",
|
| 266 |
+
],
|
| 267 |
+
"valid_interval": 10000,
|
| 268 |
+
},
|
| 269 |
+
"use_custom_dataset": [],
|
| 270 |
+
}
|
pupucodec_large/checkpoint/epoch-0053_step-2349317_loss-57.300222
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:760d80a7c5b866dafb5a18eaedee6abcb53c66e949b5cae7cb1f579d4eb988c0
|
| 3 |
+
size 2598313053
|
pupuvocoder/args.json
ADDED
|
@@ -0,0 +1,289 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_config": "egs/afgen/exp_config_afvocoder_base.json",
|
| 3 |
+
"dataset": [
|
| 4 |
+
"afgen_all_data",
|
| 5 |
+
],
|
| 6 |
+
"exp_name": "afvocoder",
|
| 7 |
+
"inference": {
|
| 8 |
+
"batch_size": 1,
|
| 9 |
+
},
|
| 10 |
+
"log_dir": "/home/vocod/experiments/afgen",
|
| 11 |
+
"model": {
|
| 12 |
+
"afvocoder": {
|
| 13 |
+
"activation": "adaasnakebeta",
|
| 14 |
+
"oversampling": true,
|
| 15 |
+
"resblock": "1",
|
| 16 |
+
"resblock_dilation_sizes": [
|
| 17 |
+
[
|
| 18 |
+
1,
|
| 19 |
+
3,
|
| 20 |
+
5,
|
| 21 |
+
],
|
| 22 |
+
[
|
| 23 |
+
1,
|
| 24 |
+
3,
|
| 25 |
+
5,
|
| 26 |
+
],
|
| 27 |
+
[
|
| 28 |
+
1,
|
| 29 |
+
3,
|
| 30 |
+
5,
|
| 31 |
+
],
|
| 32 |
+
],
|
| 33 |
+
"resblock_kernel_sizes": [
|
| 34 |
+
3,
|
| 35 |
+
7,
|
| 36 |
+
11,
|
| 37 |
+
],
|
| 38 |
+
"upsample_initial_channel": 512,
|
| 39 |
+
"upsample_kernel_sizes": [
|
| 40 |
+
16,
|
| 41 |
+
16,
|
| 42 |
+
4,
|
| 43 |
+
4,
|
| 44 |
+
4,
|
| 45 |
+
],
|
| 46 |
+
"upsample_rates": [
|
| 47 |
+
8,
|
| 48 |
+
8,
|
| 49 |
+
2,
|
| 50 |
+
2,
|
| 51 |
+
2,
|
| 52 |
+
],
|
| 53 |
+
"upsample_type": "resample",
|
| 54 |
+
},
|
| 55 |
+
"discriminators": [
|
| 56 |
+
"msd",
|
| 57 |
+
"mpd",
|
| 58 |
+
"mssbstftd",
|
| 59 |
+
"mssbcqtd",
|
| 60 |
+
],
|
| 61 |
+
"generator": "afvocoder",
|
| 62 |
+
"mpd": {
|
| 63 |
+
"discriminator_channel_mult_factor": 1,
|
| 64 |
+
"mpd_reshapes": [
|
| 65 |
+
2,
|
| 66 |
+
3,
|
| 67 |
+
5,
|
| 68 |
+
7,
|
| 69 |
+
11,
|
| 70 |
+
17,
|
| 71 |
+
23,
|
| 72 |
+
37,
|
| 73 |
+
],
|
| 74 |
+
"use_spectral_norm": false,
|
| 75 |
+
},
|
| 76 |
+
"mssbcqtd": {
|
| 77 |
+
"bins_per_octaves": [
|
| 78 |
+
24,
|
| 79 |
+
36,
|
| 80 |
+
48,
|
| 81 |
+
],
|
| 82 |
+
"dilations": [
|
| 83 |
+
1,
|
| 84 |
+
2,
|
| 85 |
+
4,
|
| 86 |
+
],
|
| 87 |
+
"filters": 32,
|
| 88 |
+
"filters_scale": 1,
|
| 89 |
+
"hop_lengths": [
|
| 90 |
+
1024,
|
| 91 |
+
512,
|
| 92 |
+
512,
|
| 93 |
+
],
|
| 94 |
+
"in_channels": 1,
|
| 95 |
+
"max_filters": 1024,
|
| 96 |
+
"n_octaves": [
|
| 97 |
+
10,
|
| 98 |
+
10,
|
| 99 |
+
10,
|
| 100 |
+
],
|
| 101 |
+
"out_channels": 1,
|
| 102 |
+
},
|
| 103 |
+
"msstftd": {
|
| 104 |
+
"filters": 32,
|
| 105 |
+
},
|
| 106 |
+
},
|
| 107 |
+
"model_type": "AFGenVocoder",
|
| 108 |
+
"preprocess": {
|
| 109 |
+
"align_mel_duration": false,
|
| 110 |
+
"audio_dir": "audios",
|
| 111 |
+
"bits": 8,
|
| 112 |
+
"contentvec_dir": "contentvec",
|
| 113 |
+
"cut_mel_frame": 128,
|
| 114 |
+
"data_augment": false,
|
| 115 |
+
"dur_dir": "durs",
|
| 116 |
+
"duration_dir": "duration",
|
| 117 |
+
"emo2id": "emo2id.json",
|
| 118 |
+
"energy_dir": "energys",
|
| 119 |
+
"energy_extract_mode": "from_mel",
|
| 120 |
+
"energy_norm": false,
|
| 121 |
+
"energy_remove_outlier": false,
|
| 122 |
+
"extract_acoustic_token": false,
|
| 123 |
+
"extract_amplitude_phase": false,
|
| 124 |
+
"extract_audio": true,
|
| 125 |
+
"extract_contentvec_feature": false,
|
| 126 |
+
"extract_duration": false,
|
| 127 |
+
"extract_energy": false,
|
| 128 |
+
"extract_label": false,
|
| 129 |
+
"extract_linear_spec": false,
|
| 130 |
+
"extract_mcep": false,
|
| 131 |
+
"extract_mel": false,
|
| 132 |
+
"extract_mert_feature": false,
|
| 133 |
+
"extract_one_hot": false,
|
| 134 |
+
"extract_phone": false,
|
| 135 |
+
"extract_pitch": true,
|
| 136 |
+
"extract_uv": true,
|
| 137 |
+
"extract_wenet_feature": false,
|
| 138 |
+
"extract_whisper_feature": false,
|
| 139 |
+
"f0_max": 1975.5,
|
| 140 |
+
"f0_min": 32.7,
|
| 141 |
+
"file_lst": "file.lst",
|
| 142 |
+
"fmax": 22050,
|
| 143 |
+
"fmin": 0,
|
| 144 |
+
"hop_size": 512,
|
| 145 |
+
"imaginary_dir": "imaginarys",
|
| 146 |
+
"is_mu_law": false,
|
| 147 |
+
"lab_dir": "labs",
|
| 148 |
+
"label_dir": "labels",
|
| 149 |
+
"lexicon_path": "./text/lexicon/librispeech-lexicon.txt",
|
| 150 |
+
"linear_dir": "linears",
|
| 151 |
+
"log_amplitude_dir": "log_amplitudes",
|
| 152 |
+
"mcep_dir": "mcep",
|
| 153 |
+
"mel_dir": "mels",
|
| 154 |
+
"mel_extract_mode": "",
|
| 155 |
+
"mel_min_max_norm": false,
|
| 156 |
+
"min_level_db": -115,
|
| 157 |
+
"n_fft": 2048,
|
| 158 |
+
"n_mel": 128,
|
| 159 |
+
"num_silent_frames": 8,
|
| 160 |
+
"phase_dir": "phases",
|
| 161 |
+
"phone_dir": "phones",
|
| 162 |
+
"phone_energy_dir": "phone_energys",
|
| 163 |
+
"phone_extractor": "espeak",
|
| 164 |
+
"phone_pitch_dir": "phone_pitches",
|
| 165 |
+
"phone_seq_file": "phone_seq_file",
|
| 166 |
+
"pitch_bin": 256,
|
| 167 |
+
"pitch_dir": "pitches",
|
| 168 |
+
"pitch_extractor": "rmvpe",
|
| 169 |
+
"pitch_max": 1100.0,
|
| 170 |
+
"pitch_min": 50.0,
|
| 171 |
+
"pitch_norm": false,
|
| 172 |
+
"pitch_remove_outlier": false,
|
| 173 |
+
"processed_dir": "data/",
|
| 174 |
+
"raw_data": "raw_data",
|
| 175 |
+
"real_dir": "reals",
|
| 176 |
+
"ref_level_db": 20,
|
| 177 |
+
"sample_rate": 44100,
|
| 178 |
+
"spk2id": "singers.json",
|
| 179 |
+
"symbols_dict": "symbols.dict",
|
| 180 |
+
"train_file": "train.json",
|
| 181 |
+
"trim_fft_size": 512,
|
| 182 |
+
"trim_hop_size": 128,
|
| 183 |
+
"trim_silence": false,
|
| 184 |
+
"trim_top_db": 30,
|
| 185 |
+
"trimmed_wav_dir": "trimmed_wavs",
|
| 186 |
+
"use_amplitude_phase": false,
|
| 187 |
+
"use_audio": true,
|
| 188 |
+
"use_dur": false,
|
| 189 |
+
"use_emoid": false,
|
| 190 |
+
"use_frame_duration": false,
|
| 191 |
+
"use_frame_energy": false,
|
| 192 |
+
"use_frame_pitch": true,
|
| 193 |
+
"use_lab": false,
|
| 194 |
+
"use_label": false,
|
| 195 |
+
"use_linear": false,
|
| 196 |
+
"use_log_scale_energy": false,
|
| 197 |
+
"use_log_scale_pitch": false,
|
| 198 |
+
"use_mel": false,
|
| 199 |
+
"use_min_max_norm_mel": false,
|
| 200 |
+
"use_one_hot": false,
|
| 201 |
+
"use_phn_seq": false,
|
| 202 |
+
"use_phone": false,
|
| 203 |
+
"use_phone_duration": false,
|
| 204 |
+
"use_phone_energy": false,
|
| 205 |
+
"use_phone_pitch": false,
|
| 206 |
+
"use_spkid": false,
|
| 207 |
+
"use_text": false,
|
| 208 |
+
"use_uv": true,
|
| 209 |
+
"use_wav": false,
|
| 210 |
+
"use_wenet": false,
|
| 211 |
+
"utt2emo": "utt2emo",
|
| 212 |
+
"utt2spk": "utt2spk",
|
| 213 |
+
"uv_dir": "uvs",
|
| 214 |
+
"valid_file": "valid.json",
|
| 215 |
+
"wav_dir": "wavs",
|
| 216 |
+
"wenet_dir": "wenet",
|
| 217 |
+
"win_size": 2048,
|
| 218 |
+
},
|
| 219 |
+
"supported_model_type": [
|
| 220 |
+
"GANVocoder",
|
| 221 |
+
"Fastspeech2",
|
| 222 |
+
"DiffSVC",
|
| 223 |
+
"Transformer",
|
| 224 |
+
"EDM",
|
| 225 |
+
"Autotune",
|
| 226 |
+
"CD",
|
| 227 |
+
],
|
| 228 |
+
"task_type": "afgen",
|
| 229 |
+
"train": {
|
| 230 |
+
"adamw": {
|
| 231 |
+
"adam_b1": 0.8,
|
| 232 |
+
"adam_b2": 0.99,
|
| 233 |
+
"lr": 0.0001,
|
| 234 |
+
},
|
| 235 |
+
"batch_size": 12,
|
| 236 |
+
"criterions": [
|
| 237 |
+
"feature",
|
| 238 |
+
"discriminator",
|
| 239 |
+
"generator",
|
| 240 |
+
"multimel",
|
| 241 |
+
],
|
| 242 |
+
"dataloader": {
|
| 243 |
+
"num_worker": 32,
|
| 244 |
+
"persistent_workers": true,
|
| 245 |
+
"pin_memory": true,
|
| 246 |
+
"prefetch_factor": 4,
|
| 247 |
+
},
|
| 248 |
+
"ddp": true,
|
| 249 |
+
"exponential_lr": {
|
| 250 |
+
"lr_decay": 0.9999996,
|
| 251 |
+
},
|
| 252 |
+
"freeze_step": 0,
|
| 253 |
+
"gradient_accumulation_step": 1,
|
| 254 |
+
"keep_checkpoint_max": 5,
|
| 255 |
+
"keep_last": [
|
| 256 |
+
3,
|
| 257 |
+
-1,
|
| 258 |
+
],
|
| 259 |
+
"max_epoch": 1000000,
|
| 260 |
+
"max_steps": 1000000,
|
| 261 |
+
"multi_speaker_training": false,
|
| 262 |
+
"optimizer": "AdamW",
|
| 263 |
+
"random_seed": 114514,
|
| 264 |
+
"reducelronplateau": {
|
| 265 |
+
"factor": 0.8,
|
| 266 |
+
"min_lr": 0.0001,
|
| 267 |
+
"patience": 10,
|
| 268 |
+
},
|
| 269 |
+
"run_eval": [
|
| 270 |
+
true,
|
| 271 |
+
],
|
| 272 |
+
"sampler": {
|
| 273 |
+
"drop_last": true,
|
| 274 |
+
"holistic_shuffle": true,
|
| 275 |
+
},
|
| 276 |
+
"save_checkpoint_stride": [
|
| 277 |
+
1,
|
| 278 |
+
],
|
| 279 |
+
"save_checkpoints_steps": 10000,
|
| 280 |
+
"save_summary_steps": 500,
|
| 281 |
+
"scheduler": "ReduceLROnPlateau",
|
| 282 |
+
"total_training_steps": 50000,
|
| 283 |
+
"tracker": [
|
| 284 |
+
"tensorboard",
|
| 285 |
+
],
|
| 286 |
+
"valid_interval": 10000,
|
| 287 |
+
},
|
| 288 |
+
"use_custom_dataset": [],
|
| 289 |
+
}
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/ckpts.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
[]
|
| 3 |
+
]
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62320716a66f7312584a41b48e5e87bf7969cc88a44cdafed46eb2c6129f9dfe
|
| 3 |
+
size 60989884
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_1.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a632cbd40f28de473e27eeabe28588e2bf31f9a35b04117e5167cb6df1a47842
|
| 3 |
+
size 118557612
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_2.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a84d21c317403e74097235e6fe65788f325025a68be5e33ff6976f26a9a29c6
|
| 3 |
+
size 263092072
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6782f5789c6c7e5970784f3953099d1a2d4e7bee1d12b998d42c9df1f2fb5dba
|
| 3 |
+
size 5695720
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/model_4.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16de04741e454da73f0632ac4538c0fe67580fe6145006e46c1941a87dbf49f4
|
| 3 |
+
size 1560888
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/optimizer.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2f839cdb624e9d0f9aceda33851a1ee890883568ff1d44ca348f4e44e62f8bce
|
| 3 |
+
size 122237016
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/optimizer_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db95ba2fe046e4f26d11ec46ec4dbd7f363e0a8ff0926a042f827982916439ba
|
| 3 |
+
size 777186848
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/random_states_0.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3bf5a3c748fcda1d56dd26339af118162fc3504e101d72688001523829fc8f12
|
| 3 |
+
size 15537
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/scheduler.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:934394bfacc36e22bcce372e9032346f4b98c69e1d2edda63e443a90665cfd1d
|
| 3 |
+
size 1401
|
pupuvocoder/checkpoint/epoch-0051_step-2553605_loss-62.135194/scheduler_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02f72e0f40c53d323e413693fe69c4f5f2f1930a766f5a6608ae4eaf06666a7c
|
| 3 |
+
size 1477
|
pupuvocoder_large/args.json
ADDED
|
@@ -0,0 +1,286 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_config": "egs/afgen/exp_config_afvocoder_base.json",
|
| 3 |
+
"dataset": [
|
| 4 |
+
"afgen_all_data",
|
| 5 |
+
],
|
| 6 |
+
"exp_name": "afvocoder_large",
|
| 7 |
+
"inference": {
|
| 8 |
+
"batch_size": 1,
|
| 9 |
+
},
|
| 10 |
+
"log_dir": "/home/vocod/experiments/afgen",
|
| 11 |
+
"model": {
|
| 12 |
+
"afvocoder": {
|
| 13 |
+
"resblock": "1",
|
| 14 |
+
"resblock_dilation_sizes": [
|
| 15 |
+
[
|
| 16 |
+
1,
|
| 17 |
+
3,
|
| 18 |
+
5,
|
| 19 |
+
],
|
| 20 |
+
[
|
| 21 |
+
1,
|
| 22 |
+
3,
|
| 23 |
+
5,
|
| 24 |
+
],
|
| 25 |
+
[
|
| 26 |
+
1,
|
| 27 |
+
3,
|
| 28 |
+
5,
|
| 29 |
+
],
|
| 30 |
+
],
|
| 31 |
+
"resblock_kernel_sizes": [
|
| 32 |
+
3,
|
| 33 |
+
7,
|
| 34 |
+
11,
|
| 35 |
+
],
|
| 36 |
+
"upsample_initial_channel": 1536,
|
| 37 |
+
"upsample_kernel_sizes": [
|
| 38 |
+
16,
|
| 39 |
+
16,
|
| 40 |
+
4,
|
| 41 |
+
4,
|
| 42 |
+
4,
|
| 43 |
+
],
|
| 44 |
+
"upsample_rates": [
|
| 45 |
+
8,
|
| 46 |
+
8,
|
| 47 |
+
2,
|
| 48 |
+
2,
|
| 49 |
+
2,
|
| 50 |
+
],
|
| 51 |
+
},
|
| 52 |
+
"discriminators": [
|
| 53 |
+
"msd",
|
| 54 |
+
"mpd",
|
| 55 |
+
"mssbstftd",
|
| 56 |
+
"mssbcqtd",
|
| 57 |
+
],
|
| 58 |
+
"generator": "afvocoder",
|
| 59 |
+
"mpd": {
|
| 60 |
+
"discriminator_channel_mult_factor": 1,
|
| 61 |
+
"mpd_reshapes": [
|
| 62 |
+
2,
|
| 63 |
+
3,
|
| 64 |
+
5,
|
| 65 |
+
7,
|
| 66 |
+
11,
|
| 67 |
+
17,
|
| 68 |
+
23,
|
| 69 |
+
37,
|
| 70 |
+
],
|
| 71 |
+
"use_spectral_norm": false,
|
| 72 |
+
},
|
| 73 |
+
"mssbcqtd": {
|
| 74 |
+
"bins_per_octaves": [
|
| 75 |
+
24,
|
| 76 |
+
36,
|
| 77 |
+
48,
|
| 78 |
+
],
|
| 79 |
+
"dilations": [
|
| 80 |
+
1,
|
| 81 |
+
2,
|
| 82 |
+
4,
|
| 83 |
+
],
|
| 84 |
+
"filters": 32,
|
| 85 |
+
"filters_scale": 1,
|
| 86 |
+
"hop_lengths": [
|
| 87 |
+
1024,
|
| 88 |
+
512,
|
| 89 |
+
512,
|
| 90 |
+
],
|
| 91 |
+
"in_channels": 1,
|
| 92 |
+
"max_filters": 1024,
|
| 93 |
+
"n_octaves": [
|
| 94 |
+
10,
|
| 95 |
+
10,
|
| 96 |
+
10,
|
| 97 |
+
],
|
| 98 |
+
"out_channels": 1,
|
| 99 |
+
},
|
| 100 |
+
"msstftd": {
|
| 101 |
+
"filters": 32,
|
| 102 |
+
},
|
| 103 |
+
},
|
| 104 |
+
"model_type": "AFGenVocoder",
|
| 105 |
+
"preprocess": {
|
| 106 |
+
"align_mel_duration": false,
|
| 107 |
+
"audio_dir": "audios",
|
| 108 |
+
"bits": 8,
|
| 109 |
+
"contentvec_dir": "contentvec",
|
| 110 |
+
"cut_mel_frame": 128,
|
| 111 |
+
"data_augment": false,
|
| 112 |
+
"dur_dir": "durs",
|
| 113 |
+
"duration_dir": "duration",
|
| 114 |
+
"emo2id": "emo2id.json",
|
| 115 |
+
"energy_dir": "energys",
|
| 116 |
+
"energy_extract_mode": "from_mel",
|
| 117 |
+
"energy_norm": false,
|
| 118 |
+
"energy_remove_outlier": false,
|
| 119 |
+
"extract_acoustic_token": false,
|
| 120 |
+
"extract_amplitude_phase": false,
|
| 121 |
+
"extract_audio": true,
|
| 122 |
+
"extract_contentvec_feature": false,
|
| 123 |
+
"extract_duration": false,
|
| 124 |
+
"extract_energy": false,
|
| 125 |
+
"extract_label": false,
|
| 126 |
+
"extract_linear_spec": false,
|
| 127 |
+
"extract_mcep": false,
|
| 128 |
+
"extract_mel": false,
|
| 129 |
+
"extract_mert_feature": false,
|
| 130 |
+
"extract_one_hot": false,
|
| 131 |
+
"extract_phone": false,
|
| 132 |
+
"extract_pitch": true,
|
| 133 |
+
"extract_uv": true,
|
| 134 |
+
"extract_wenet_feature": false,
|
| 135 |
+
"extract_whisper_feature": false,
|
| 136 |
+
"f0_max": 1975.5,
|
| 137 |
+
"f0_min": 32.7,
|
| 138 |
+
"file_lst": "file.lst",
|
| 139 |
+
"fmax": 22050,
|
| 140 |
+
"fmin": 0,
|
| 141 |
+
"hop_size": 512,
|
| 142 |
+
"imaginary_dir": "imaginarys",
|
| 143 |
+
"is_mu_law": false,
|
| 144 |
+
"lab_dir": "labs",
|
| 145 |
+
"label_dir": "labels",
|
| 146 |
+
"lexicon_path": "./text/lexicon/librispeech-lexicon.txt",
|
| 147 |
+
"linear_dir": "linears",
|
| 148 |
+
"log_amplitude_dir": "log_amplitudes",
|
| 149 |
+
"mcep_dir": "mcep",
|
| 150 |
+
"mel_dir": "mels",
|
| 151 |
+
"mel_extract_mode": "",
|
| 152 |
+
"mel_min_max_norm": false,
|
| 153 |
+
"min_level_db": -115,
|
| 154 |
+
"n_fft": 2048,
|
| 155 |
+
"n_mel": 128,
|
| 156 |
+
"num_silent_frames": 8,
|
| 157 |
+
"phase_dir": "phases",
|
| 158 |
+
"phone_dir": "phones",
|
| 159 |
+
"phone_energy_dir": "phone_energys",
|
| 160 |
+
"phone_extractor": "espeak",
|
| 161 |
+
"phone_pitch_dir": "phone_pitches",
|
| 162 |
+
"phone_seq_file": "phone_seq_file",
|
| 163 |
+
"pitch_bin": 256,
|
| 164 |
+
"pitch_dir": "pitches",
|
| 165 |
+
"pitch_extractor": "rmvpe",
|
| 166 |
+
"pitch_max": 1100.0,
|
| 167 |
+
"pitch_min": 50.0,
|
| 168 |
+
"pitch_norm": false,
|
| 169 |
+
"pitch_remove_outlier": false,
|
| 170 |
+
"processed_dir": "data/",
|
| 171 |
+
"raw_data": "raw_data",
|
| 172 |
+
"real_dir": "reals",
|
| 173 |
+
"ref_level_db": 20,
|
| 174 |
+
"sample_rate": 44100,
|
| 175 |
+
"spk2id": "singers.json",
|
| 176 |
+
"symbols_dict": "symbols.dict",
|
| 177 |
+
"train_file": "train.json",
|
| 178 |
+
"trim_fft_size": 512,
|
| 179 |
+
"trim_hop_size": 128,
|
| 180 |
+
"trim_silence": false,
|
| 181 |
+
"trim_top_db": 30,
|
| 182 |
+
"trimmed_wav_dir": "trimmed_wavs",
|
| 183 |
+
"use_amplitude_phase": false,
|
| 184 |
+
"use_audio": true,
|
| 185 |
+
"use_dur": false,
|
| 186 |
+
"use_emoid": false,
|
| 187 |
+
"use_frame_duration": false,
|
| 188 |
+
"use_frame_energy": false,
|
| 189 |
+
"use_frame_pitch": true,
|
| 190 |
+
"use_lab": false,
|
| 191 |
+
"use_label": false,
|
| 192 |
+
"use_linear": false,
|
| 193 |
+
"use_log_scale_energy": false,
|
| 194 |
+
"use_log_scale_pitch": false,
|
| 195 |
+
"use_mel": false,
|
| 196 |
+
"use_min_max_norm_mel": false,
|
| 197 |
+
"use_one_hot": false,
|
| 198 |
+
"use_phn_seq": false,
|
| 199 |
+
"use_phone": false,
|
| 200 |
+
"use_phone_duration": false,
|
| 201 |
+
"use_phone_energy": false,
|
| 202 |
+
"use_phone_pitch": false,
|
| 203 |
+
"use_spkid": false,
|
| 204 |
+
"use_text": false,
|
| 205 |
+
"use_uv": true,
|
| 206 |
+
"use_wav": false,
|
| 207 |
+
"use_wenet": false,
|
| 208 |
+
"utt2emo": "utt2emo",
|
| 209 |
+
"utt2spk": "utt2spk",
|
| 210 |
+
"uv_dir": "uvs",
|
| 211 |
+
"valid_file": "valid.json",
|
| 212 |
+
"wav_dir": "wavs",
|
| 213 |
+
"wenet_dir": "wenet",
|
| 214 |
+
"win_size": 2048,
|
| 215 |
+
},
|
| 216 |
+
"supported_model_type": [
|
| 217 |
+
"GANVocoder",
|
| 218 |
+
"Fastspeech2",
|
| 219 |
+
"DiffSVC",
|
| 220 |
+
"Transformer",
|
| 221 |
+
"EDM",
|
| 222 |
+
"Autotune",
|
| 223 |
+
"CD",
|
| 224 |
+
],
|
| 225 |
+
"task_type": "afgen",
|
| 226 |
+
"train": {
|
| 227 |
+
"adamw": {
|
| 228 |
+
"adam_b1": 0.8,
|
| 229 |
+
"adam_b2": 0.99,
|
| 230 |
+
"lr": 0.0001,
|
| 231 |
+
},
|
| 232 |
+
"batch_size": 4,
|
| 233 |
+
"criterions": [
|
| 234 |
+
"feature",
|
| 235 |
+
"discriminator",
|
| 236 |
+
"generator",
|
| 237 |
+
"multimel",
|
| 238 |
+
],
|
| 239 |
+
"dataloader": {
|
| 240 |
+
"num_worker": 32,
|
| 241 |
+
"persistent_workers": true,
|
| 242 |
+
"pin_memory": true,
|
| 243 |
+
"prefetch_factor": 4,
|
| 244 |
+
},
|
| 245 |
+
"ddp": true,
|
| 246 |
+
"exponential_lr": {
|
| 247 |
+
"lr_decay": 0.9999996,
|
| 248 |
+
},
|
| 249 |
+
"freeze_step": 0,
|
| 250 |
+
"gradient_accumulation_step": 1,
|
| 251 |
+
"keep_checkpoint_max": 5,
|
| 252 |
+
"keep_last": [
|
| 253 |
+
3,
|
| 254 |
+
-1,
|
| 255 |
+
],
|
| 256 |
+
"max_epoch": 1000000,
|
| 257 |
+
"max_steps": 1000000,
|
| 258 |
+
"multi_speaker_training": false,
|
| 259 |
+
"optimizer": "AdamW",
|
| 260 |
+
"random_seed": 114514,
|
| 261 |
+
"reducelronplateau": {
|
| 262 |
+
"factor": 0.8,
|
| 263 |
+
"min_lr": 0.0001,
|
| 264 |
+
"patience": 10,
|
| 265 |
+
},
|
| 266 |
+
"run_eval": [
|
| 267 |
+
true,
|
| 268 |
+
],
|
| 269 |
+
"sampler": {
|
| 270 |
+
"drop_last": true,
|
| 271 |
+
"holistic_shuffle": true,
|
| 272 |
+
},
|
| 273 |
+
"save_checkpoint_stride": [
|
| 274 |
+
1,
|
| 275 |
+
],
|
| 276 |
+
"save_checkpoints_steps": 10000,
|
| 277 |
+
"save_summary_steps": 500,
|
| 278 |
+
"scheduler": "ReduceLROnPlateau",
|
| 279 |
+
"total_training_steps": 50000,
|
| 280 |
+
"tracker": [
|
| 281 |
+
"tensorboard",
|
| 282 |
+
],
|
| 283 |
+
"valid_interval": 10000,
|
| 284 |
+
},
|
| 285 |
+
"use_custom_dataset": [],
|
| 286 |
+
}
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/ckpts.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
[]
|
| 3 |
+
]
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf9a439dd1c17719a3a5f2c44ac209a08391b234a8d1e56fb8dcae0fc8904b43
|
| 3 |
+
size 536315052
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_1.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d54df842a9676f5b9bde274fb62d4629f751aa87aafcd673c66c99a119772d4
|
| 3 |
+
size 118557612
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_2.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c8395b3011862e165e3020ec40cb9959de475fc35b1b447436e563d0b94f74d
|
| 3 |
+
size 263092072
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:24f761b5c32671a1a3beac5574934ac5adc5769d1c6d89d2249cd40e1762a84f
|
| 3 |
+
size 5695720
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/model_4.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3292abaf7d17007de38fb6fa765e16707148dff97814eb2f8cd83354f12e532d
|
| 3 |
+
size 1560888
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/optimizer.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2fd47dffafc21204ac4ddb1e19ab0db318453b335715440e883f8e608a2ce7e6
|
| 3 |
+
size 1072885144
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/optimizer_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:56bbac969b78adfb0cd9e8e6ecf899f15cb70df11a0b4fe53e3a6bb99b5b2af1
|
| 3 |
+
size 777186848
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/random_states_0.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07f64d00ee76e090a6c362222c254f35f622a8d399a79380772ddfaaed44f236
|
| 3 |
+
size 16449
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/scheduler.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f9a058e0336350ebbda962a13b2283b707db64849ded3671653eac8424de7579
|
| 3 |
+
size 1401
|
pupuvocoder_large/checkpoint/epoch-0026_step-2315282_loss-46.095750/scheduler_1.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faed8f23fe895621c9fa0a495358083fdd927036b6f7429166a2ab86f5d08ad7
|
| 3 |
+
size 1477
|