Upload checkpoint
Browse files- args.json +209 -0
- checkpoint/epoch-0125_step-0097902_loss-1.074404/optimizer.bin +3 -0
- checkpoint/epoch-0125_step-0097902_loss-1.074404/pytorch_model.bin +3 -0
- checkpoint/epoch-0125_step-0097902_loss-1.074404/random_states_0.pkl +3 -0
- checkpoint/epoch-0125_step-0097902_loss-1.074404/scheduler.bin +3 -0
args.json
ADDED
@@ -0,0 +1,209 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"dataset_path": {
|
3 |
+
"LJSpeech": "/home/datasets/LJSpeech-1.1",
|
4 |
+
},
|
5 |
+
"base_config": "config/base.json",
|
6 |
+
"dataset": [
|
7 |
+
"LJSpeech",
|
8 |
+
],
|
9 |
+
"preprocess": {
|
10 |
+
"trim_silence": false,
|
11 |
+
"num_silent_frames": 8,
|
12 |
+
"trim_fft_size": 512,
|
13 |
+
"trim_hop_size": 128,
|
14 |
+
"trim_top_db": 30,
|
15 |
+
"extract_mel": true,
|
16 |
+
"extract_mcep": false,
|
17 |
+
"extract_pitch": true,
|
18 |
+
"extract_uv": true,
|
19 |
+
"pitch_norm": false,
|
20 |
+
"extract_audio": true,
|
21 |
+
"extract_label": false,
|
22 |
+
"pitch_extractor": "parselmouth",
|
23 |
+
"extract_energy": false,
|
24 |
+
"energy_norm": false,
|
25 |
+
"energy_extract_mode": "from_mel",
|
26 |
+
"extract_duration": false,
|
27 |
+
"mel_min_max_norm": false,
|
28 |
+
"mu_law_norm": false,
|
29 |
+
"extract_whisper_feature": false,
|
30 |
+
"extract_contentvec_feature": false,
|
31 |
+
"extract_mert_feature": false,
|
32 |
+
"extract_wenet_feature": false,
|
33 |
+
"n_mel": 80,
|
34 |
+
"win_size": 1024,
|
35 |
+
"hop_size": 256,
|
36 |
+
"sample_rate": 22050,
|
37 |
+
"n_fft": 1024,
|
38 |
+
"fmin": 0,
|
39 |
+
"fmax": 8000,
|
40 |
+
"min_level_db": -115,
|
41 |
+
"ref_level_db": 20,
|
42 |
+
"bits": 8,
|
43 |
+
"processed_dir": "processed_data",
|
44 |
+
"trimmed_wav_dir": "trimmed_wavs",
|
45 |
+
"wav_dir": "wavs",
|
46 |
+
"audio_dir": "audios",
|
47 |
+
"label_dir": "labels",
|
48 |
+
"mel_dir": "mels",
|
49 |
+
"mcep_dir": "mcep",
|
50 |
+
"dur_dir": "durs",
|
51 |
+
"lab_dir": "labs",
|
52 |
+
"wenet_dir": "wenet",
|
53 |
+
"contentvec_dir": "contentvec",
|
54 |
+
"pitch_dir": "pitches",
|
55 |
+
"energy_dir": "energys",
|
56 |
+
"uv_dir": "uvs",
|
57 |
+
"duration_dir": "duration",
|
58 |
+
"phone_seq_file": "phone_seq_file",
|
59 |
+
"file_lst": "file.lst",
|
60 |
+
"train_file": "train.json",
|
61 |
+
"valid_file": "test.json",
|
62 |
+
"spk2id": "spk2id.json",
|
63 |
+
"utt2spk": "utt2spk",
|
64 |
+
"emo2id": "emo2id.json",
|
65 |
+
"utt2emo": "utt2emo",
|
66 |
+
"use_phn_seq": false,
|
67 |
+
"use_lab": false,
|
68 |
+
"use_mel": true,
|
69 |
+
"use_wav": false,
|
70 |
+
"use_phone_pitch": false,
|
71 |
+
"use_log_scale_pitch": false,
|
72 |
+
"use_phone_energy": false,
|
73 |
+
"use_phone_duration": false,
|
74 |
+
"use_log_scale_energy": false,
|
75 |
+
"use_wenet": false,
|
76 |
+
"use_dur": false,
|
77 |
+
"use_spkid": false,
|
78 |
+
"use_emoid": false,
|
79 |
+
"use_frame_pitch": false,
|
80 |
+
"use_uv": true,
|
81 |
+
"use_frame_energy": false,
|
82 |
+
"use_frame_duration": false,
|
83 |
+
"use_audio": true,
|
84 |
+
"use_label": false,
|
85 |
+
"use_one_hot": false,
|
86 |
+
"data_augment": false,
|
87 |
+
"align_mel_duration": false,
|
88 |
+
"f0_min": 50,
|
89 |
+
"f0_max": 1100,
|
90 |
+
"pitch_bin": 256,
|
91 |
+
"pitch_max": 1100.0,
|
92 |
+
"pitch_min": 50.0,
|
93 |
+
"cut_mel_frame": 32,
|
94 |
+
"use_min_max_norm_mel": false,
|
95 |
+
},
|
96 |
+
"train": {
|
97 |
+
"ddp": false,
|
98 |
+
"random_seed": 970227,
|
99 |
+
"batch_size": 16,
|
100 |
+
"epochs": 50000,
|
101 |
+
"max_steps": 1000000,
|
102 |
+
"total_training_steps": 50000,
|
103 |
+
"save_summary_steps": 500,
|
104 |
+
"save_checkpoints_steps": 10000,
|
105 |
+
"valid_interval": 10000,
|
106 |
+
"keep_checkpoint_max": 15,
|
107 |
+
"multi_speaker_training": false,
|
108 |
+
"adamw": {
|
109 |
+
"lr": 0.0002,
|
110 |
+
"adam_b1": 0.8,
|
111 |
+
"adam_b2": 0.99,
|
112 |
+
},
|
113 |
+
"exponential_lr": {
|
114 |
+
"lr_decay": 0.999,
|
115 |
+
},
|
116 |
+
"criterions": [
|
117 |
+
"feature",
|
118 |
+
"discriminator",
|
119 |
+
"generator",
|
120 |
+
"mel",
|
121 |
+
"wav",
|
122 |
+
],
|
123 |
+
},
|
124 |
+
"model_type": "GANVocoder",
|
125 |
+
"model": {
|
126 |
+
"generator": "hifigan",
|
127 |
+
"discriminators": [
|
128 |
+
"msd",
|
129 |
+
"mpd",
|
130 |
+
"msstftd",
|
131 |
+
"mscqtd",
|
132 |
+
],
|
133 |
+
"hifigan": {
|
134 |
+
"resblock": "2",
|
135 |
+
"upsample_rates": [
|
136 |
+
8,
|
137 |
+
8,
|
138 |
+
4,
|
139 |
+
],
|
140 |
+
"upsample_kernel_sizes": [
|
141 |
+
16,
|
142 |
+
16,
|
143 |
+
8,
|
144 |
+
],
|
145 |
+
"upsample_initial_channel": 256,
|
146 |
+
"resblock_kernel_sizes": [
|
147 |
+
3,
|
148 |
+
5,
|
149 |
+
7,
|
150 |
+
],
|
151 |
+
"resblock_dilation_sizes": [
|
152 |
+
[
|
153 |
+
1,
|
154 |
+
2,
|
155 |
+
],
|
156 |
+
[
|
157 |
+
2,
|
158 |
+
6,
|
159 |
+
],
|
160 |
+
[
|
161 |
+
3,
|
162 |
+
12,
|
163 |
+
],
|
164 |
+
],
|
165 |
+
},
|
166 |
+
"mpd": {
|
167 |
+
"mpd_reshapes": [
|
168 |
+
2,
|
169 |
+
3,
|
170 |
+
5,
|
171 |
+
7,
|
172 |
+
11,
|
173 |
+
],
|
174 |
+
"use_spectral_norm": false,
|
175 |
+
"discriminator_channel_multi": 1,
|
176 |
+
},
|
177 |
+
"msstftd": {
|
178 |
+
"filters": 32,
|
179 |
+
},
|
180 |
+
"mscqtd": {
|
181 |
+
"hop_lengths": [
|
182 |
+
512,
|
183 |
+
256,
|
184 |
+
256,
|
185 |
+
],
|
186 |
+
"filters": 32,
|
187 |
+
"max_filters": 1024,
|
188 |
+
"filters_scale": 1,
|
189 |
+
"dilations": [
|
190 |
+
1,
|
191 |
+
2,
|
192 |
+
4,
|
193 |
+
],
|
194 |
+
"in_channels": 1,
|
195 |
+
"out_channels": 1,
|
196 |
+
"n_octaves": [
|
197 |
+
9,
|
198 |
+
9,
|
199 |
+
9,
|
200 |
+
],
|
201 |
+
"bins_per_octaves": [
|
202 |
+
24,
|
203 |
+
36,
|
204 |
+
48,
|
205 |
+
],
|
206 |
+
},
|
207 |
+
},
|
208 |
+
"exp_name": "hifigan",
|
209 |
+
}
|
checkpoint/epoch-0125_step-0097902_loss-1.074404/optimizer.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b5d9b9b39ccf380e221a5a6d478a7ff25180d06de2bf950a8dae87acf0ed193a
|
3 |
+
size 276922053
|
checkpoint/epoch-0125_step-0097902_loss-1.074404/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8e43d61b22f26605009fa3158fd5180e372478b2da60ef10ff01e5757f27507a
|
3 |
+
size 140532853
|
checkpoint/epoch-0125_step-0097902_loss-1.074404/random_states_0.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3bd5cd54e288a2fd338b25270761631a6f1dc8a9341e57ae2ce4530f0f020b46
|
3 |
+
size 14663
|
checkpoint/epoch-0125_step-0097902_loss-1.074404/scheduler.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:41d54eb84f01111fe50f4fa27b78a012dfbe425d9f708a28f54191fcab8bd569
|
3 |
+
size 627
|