zzb1420 commited on
Commit
01ca460
·
1 Parent(s): bfa18a3

Upload 3 files

Browse files
tools/xiaopin/G_5000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f968bc7d98b98243acd41ae976979d63d7d34e4ca8fd94a2ada68827e034249
3
+ size 728369839
tools/xiaopin/config.json ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "train": {
3
+ "log_interval": 200,
4
+ "eval_interval": 1000,
5
+ "seed": 42,
6
+ "epochs": 1000,
7
+ "learning_rate": 0.0002,
8
+ "betas": [
9
+ 0.8,
10
+ 0.99
11
+ ],
12
+ "eps": 1e-09,
13
+ "batch_size": 16,
14
+ "bf16_run": false,
15
+ "lr_decay": 0.99995,
16
+ "segment_size": 16384,
17
+ "init_lr_ratio": 1,
18
+ "warmup_epochs": 0,
19
+ "c_mel": 45,
20
+ "c_kl": 1.0,
21
+ "c_commit": 100,
22
+ "skip_optimizer": true,
23
+ "freeze_ZH_bert": false,
24
+ "freeze_JP_bert": false,
25
+ "freeze_EN_bert": false,
26
+ "freeze_emo": false
27
+ },
28
+ "data": {
29
+ "training_files": "filelists/train.list",
30
+ "validation_files": "filelists/val.list",
31
+ "max_wav_value": 32768.0,
32
+ "sampling_rate": 44100,
33
+ "filter_length": 2048,
34
+ "hop_length": 512,
35
+ "win_length": 2048,
36
+ "n_mel_channels": 128,
37
+ "mel_fmin": 0.0,
38
+ "mel_fmax": null,
39
+ "add_blank": true,
40
+ "n_speakers": 1,
41
+ "cleaned_text": true,
42
+ "spk2id": {
43
+ "xiapinTwo": 0
44
+ }
45
+ },
46
+ "model": {
47
+ "use_spk_conditioned_encoder": true,
48
+ "use_noise_scaled_mas": true,
49
+ "use_mel_posterior_encoder": false,
50
+ "use_duration_discriminator": true,
51
+ "inter_channels": 192,
52
+ "hidden_channels": 192,
53
+ "filter_channels": 768,
54
+ "n_heads": 2,
55
+ "n_layers": 6,
56
+ "kernel_size": 3,
57
+ "p_dropout": 0.1,
58
+ "resblock": "1",
59
+ "resblock_kernel_sizes": [
60
+ 3,
61
+ 7,
62
+ 11
63
+ ],
64
+ "resblock_dilation_sizes": [
65
+ [
66
+ 1,
67
+ 3,
68
+ 5
69
+ ],
70
+ [
71
+ 1,
72
+ 3,
73
+ 5
74
+ ],
75
+ [
76
+ 1,
77
+ 3,
78
+ 5
79
+ ]
80
+ ],
81
+ "upsample_rates": [
82
+ 8,
83
+ 8,
84
+ 2,
85
+ 2,
86
+ 2
87
+ ],
88
+ "upsample_initial_channel": 512,
89
+ "upsample_kernel_sizes": [
90
+ 16,
91
+ 16,
92
+ 8,
93
+ 2,
94
+ 2
95
+ ],
96
+ "n_layers_q": 3,
97
+ "use_spectral_norm": false,
98
+ "gin_channels": 512,
99
+ "slm": {
100
+ "model": "./slm/wavlm-base-plus",
101
+ "sr": 16000,
102
+ "hidden": 768,
103
+ "nlayers": 13,
104
+ "initial_channel": 64
105
+ }
106
+ },
107
+ "version": "2.3"
108
+ }
tools/xiaopin/config.yml ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ bert_gen:
2
+ config_path: config.json
3
+ device: cuda
4
+ num_processes: 4
5
+ use_multi_device: false
6
+ dataset_path: Data\xiaopinTwo
7
+ emo_gen:
8
+ config_path: config.json
9
+ device: cuda
10
+ num_processes: 4
11
+ use_multi_device: false
12
+ mirror: ''
13
+ openi_token: ''
14
+ preprocess_text:
15
+ clean: true
16
+ cleaned_path: filelists/cleaned.list
17
+ config_path: config.json
18
+ max_val_total: 12
19
+ train_path: filelists/train.list
20
+ transcription_path: filelists/short_character_anno.list
21
+ val_path: filelists/val.list
22
+ val_per_lang: 4
23
+ resample:
24
+ in_dir: raw
25
+ out_dir: wavs
26
+ sampling_rate: 44100
27
+ server:
28
+ device: cuda
29
+ models:
30
+ - config: ''
31
+ device: cuda
32
+ language: ZH
33
+ model: ''
34
+ speakers:
35
+ - length_scale: 1
36
+ noise_scale: 0.6
37
+ noise_scale_w: 0.8
38
+ sdp_ratio: 0.2
39
+ speaker: "\u79D1\u6BD4"
40
+ - length_scale: 0.5
41
+ noise_scale: 0.7
42
+ noise_scale_w: 0.8
43
+ sdp_ratio: 0.3
44
+ speaker: "\u4E94\u6761\u609F"
45
+ - length_scale: 1.2
46
+ noise_scale: 0.6
47
+ noise_scale_w: 0.8
48
+ sdp_ratio: 0.2
49
+ speaker: "\u5B89\u500D\u664B\u4E09"
50
+ - config: ''
51
+ device: cpu
52
+ language: JP
53
+ model: ''
54
+ speakers: []
55
+ port: 5000
56
+ train_ms:
57
+ base:
58
+ model_image: "Bert-VITS2_2.3\u5E95\u6A21"
59
+ repo_id: Stardust_minus/Bert-VITS2
60
+ use_base_model: false
61
+ config_path: config.json
62
+ env:
63
+ LOCAL_RANK: 0
64
+ MASTER_ADDR: localhost
65
+ MASTER_PORT: 10086
66
+ RANK: 0
67
+ WORLD_SIZE: 1
68
+ keep_ckpts: 10
69
+ model: models
70
+ num_workers: 2
71
+ spec_cache: true
72
+ translate:
73
+ app_key: ''
74
+ secret_key: ''
75
+ webui:
76
+ config_path: config.json
77
+ debug: false
78
+ device: cuda
79
+ language_identification_library: langid
80
+ model: models/G_5000.pth
81
+ port: 7860
82
+ share: false