Hugging Face's logo

PorYoung
/
sovits-Kasugano-Sora

Model card Files
xet
 Community
PorYoung commited on
Commit
b874297
·
1 Parent(s): 92df858

add pth

Browse files
Files changed (17) hide show
  1. sovits4.0/mixed/D_7000.pth +3 -0
  2. sovits4.0/mixed/G_7000.pth +3 -0
  3. sovits4.0/mixed/config.json +94 -0
  4. sovits4.0/mixed/trained_from_G7000/G_242000.pth +3 -0
  5. sovits4.0/speak/D_20000.pth +3 -0
  6. sovits4.0/speak/G_20000.pth +3 -0
  7. sovits4.0/speak/config.json +94 -0
  8. svs-fork/mixed/D_42000.pth +3 -0
  9. svs-fork/mixed/D_9000.pth +3 -0
  10. svs-fork/mixed/G_42000.pth +3 -0
  11. svs-fork/mixed/G_9000.pth +3 -0
  12. svs-fork/mixed/config.json +95 -0
  13. svs-fork/mixed/trained_from_speak_G40000/D_110000.pth +3 -0
  14. svs-fork/mixed/trained_from_speak_G40000/G_110000.pth +3 -0
  15. svs-fork/speak/G_26000.pth +3 -0
  16. svs-fork/speak/G_40000.pth +3 -0
  17. svs-fork/speak/G_6000.pth +3 -0
sovits4.0/mixed/D_7000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:065d2626327697d64f2a11b3be7487c87fcf71561e6580a9e45b65611278b505
3
+ size 561093195
sovits4.0/mixed/G_7000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbd1df14aa2ed6146b512861bea5aa7e6fe61862442d8783485aa4a9dd11c642
3
+ size 542170079
sovits4.0/mixed/config.json ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "train": {
3
+ "log_interval": 200,
4
+ "eval_interval": 500,
5
+ "seed": 1234,
6
+ "epochs": 100000,
7
+ "learning_rate": 0.0001,
8
+ "betas": [
9
+ 0.8,
10
+ 0.99
11
+ ],
12
+ "eps": 1e-09,
13
+ "batch_size": 40,
14
+ "fp16_run": false,
15
+ "lr_decay": 0.999875,
16
+ "segment_size": 10240,
17
+ "init_lr_ratio": 1,
18
+ "warmup_epochs": 0,
19
+ "c_mel": 45,
20
+ "c_kl": 1.0,
21
+ "use_sr": true,
22
+ "max_speclen": 512,
23
+ "port": "8001",
24
+ "keep_ckpts": 4,
25
+ "all_in_mem": false
26
+ },
27
+ "data": {
28
+ "training_files": "filelists/train.txt",
29
+ "validation_files": "filelists/val.txt",
30
+ "max_wav_value": 32768.0,
31
+ "sampling_rate": 44100,
32
+ "filter_length": 2048,
33
+ "hop_length": 512,
34
+ "win_length": 2048,
35
+ "n_mel_channels": 80,
36
+ "mel_fmin": 0.0,
37
+ "mel_fmax": 22050
38
+ },
39
+ "model": {
40
+ "inter_channels": 192,
41
+ "hidden_channels": 192,
42
+ "filter_channels": 768,
43
+ "n_heads": 2,
44
+ "n_layers": 6,
45
+ "kernel_size": 3,
46
+ "p_dropout": 0.1,
47
+ "resblock": "1",
48
+ "resblock_kernel_sizes": [
49
+ 3,
50
+ 7,
51
+ 11
52
+ ],
53
+ "resblock_dilation_sizes": [
54
+ [
55
+ 1,
56
+ 3,
57
+ 5
58
+ ],
59
+ [
60
+ 1,
61
+ 3,
62
+ 5
63
+ ],
64
+ [
65
+ 1,
66
+ 3,
67
+ 5
68
+ ]
69
+ ],
70
+ "upsample_rates": [
71
+ 8,
72
+ 8,
73
+ 2,
74
+ 2,
75
+ 2
76
+ ],
77
+ "upsample_initial_channel": 512,
78
+ "upsample_kernel_sizes": [
79
+ 16,
80
+ 16,
81
+ 4,
82
+ 4,
83
+ 4
84
+ ],
85
+ "n_layers_q": 3,
86
+ "use_spectral_norm": false,
87
+ "gin_channels": 256,
88
+ "ssl_dim": 256,
89
+ "n_speakers": 1
90
+ },
91
+ "spk": {
92
+ "kasugano_sora_mix": 0
93
+ }
94
+ }
sovits4.0/mixed/trained_from_G7000/G_242000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:065cbb61ee7adbb285ce779888546fa64623ccc0719ef4fee6652967f6d50812
3
+ size 542181531
sovits4.0/speak/D_20000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:942ea831f36477abd82dd1886b5e364c9964a498ee59c8e0c9e2b58e09a50758
3
+ size 561098185
sovits4.0/speak/G_20000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11a2a2c2678c3d002c80d8a2373cc2308e8b1ae922be54f16f1ec4ec2bfcfee0
3
+ size 542178077
sovits4.0/speak/config.json ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "train": {
3
+ "log_interval": 200,
4
+ "eval_interval": 500,
5
+ "seed": 1234,
6
+ "epochs": 100000,
7
+ "learning_rate": 0.0002,
8
+ "betas": [
9
+ 0.8,
10
+ 0.99
11
+ ],
12
+ "eps": 1e-09,
13
+ "batch_size": 40,
14
+ "fp16_run": false,
15
+ "lr_decay": 0.999875,
16
+ "segment_size": 10240,
17
+ "init_lr_ratio": 1,
18
+ "warmup_epochs": 0,
19
+ "c_mel": 45,
20
+ "c_kl": 1.0,
21
+ "use_sr": true,
22
+ "max_speclen": 512,
23
+ "port": "8002",
24
+ "keep_ckpts": 4,
25
+ "all_in_mem": false
26
+ },
27
+ "data": {
28
+ "training_files": "filelists/train.txt",
29
+ "validation_files": "filelists/val.txt",
30
+ "max_wav_value": 32768.0,
31
+ "sampling_rate": 44100,
32
+ "filter_length": 2048,
33
+ "hop_length": 512,
34
+ "win_length": 2048,
35
+ "n_mel_channels": 80,
36
+ "mel_fmin": 0.0,
37
+ "mel_fmax": 22050
38
+ },
39
+ "model": {
40
+ "inter_channels": 192,
41
+ "hidden_channels": 192,
42
+ "filter_channels": 768,
43
+ "n_heads": 2,
44
+ "n_layers": 6,
45
+ "kernel_size": 3,
46
+ "p_dropout": 0.1,
47
+ "resblock": "1",
48
+ "resblock_kernel_sizes": [
49
+ 3,
50
+ 7,
51
+ 11
52
+ ],
53
+ "resblock_dilation_sizes": [
54
+ [
55
+ 1,
56
+ 3,
57
+ 5
58
+ ],
59
+ [
60
+ 1,
61
+ 3,
62
+ 5
63
+ ],
64
+ [
65
+ 1,
66
+ 3,
67
+ 5
68
+ ]
69
+ ],
70
+ "upsample_rates": [
71
+ 8,
72
+ 8,
73
+ 2,
74
+ 2,
75
+ 2
76
+ ],
77
+ "upsample_initial_channel": 512,
78
+ "upsample_kernel_sizes": [
79
+ 16,
80
+ 16,
81
+ 4,
82
+ 4,
83
+ 4
84
+ ],
85
+ "n_layers_q": 3,
86
+ "use_spectral_norm": false,
87
+ "gin_channels": 256,
88
+ "ssl_dim": 256,
89
+ "n_speakers": 1
90
+ },
91
+ "spk": {
92
+ "kasugano_sora_speak": 0
93
+ }
94
+ }
svs-fork/mixed/D_42000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8d32eee45f0431e73ceae765679eb0d75b33f4166a784915f698402aac59bd3
3
+ size 561098249
svs-fork/mixed/D_9000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a5670cfeb6e1e50aa942a99dd0a84374f4de336ae69b393e1140af9a8dcf423
3
+ size 561098249
svs-fork/mixed/G_42000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e28254386e3e83d7a1eaea197bf1bacc50d53f97dcb83428b0724dd7bc87083
3
+ size 542789469
svs-fork/mixed/G_9000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab937ec28ee831dc6b666d75080def78d5cedfb872890b59079ed7eae48b4674
3
+ size 542789469
svs-fork/mixed/config.json ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "train": {
3
+ "log_interval": 200,
4
+ "eval_interval": 500,
5
+ "seed": 1234,
6
+ "epochs": 100000,
7
+ "learning_rate": 0.00015,
8
+ "betas": [
9
+ 0.8,
10
+ 0.99
11
+ ],
12
+ "eps": 1e-09,
13
+ "batch_size": 20,
14
+ "fp16_run": false,
15
+ "bf16_run": false,
16
+ "lr_decay": 0.9995,
17
+ "segment_size": 10240,
18
+ "init_lr_ratio": 1,
19
+ "warmup_epochs": 0,
20
+ "c_mel": 45,
21
+ "c_kl": 1.0,
22
+ "use_sr": true,
23
+ "max_speclen": 512,
24
+ "port": "8001",
25
+ "keep_ckpts": 4,
26
+ "num_workers": 4
27
+ },
28
+ "data": {
29
+ "training_files": "filelists/44k/train.txt",
30
+ "validation_files": "filelists/44k/val.txt",
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": 80,
37
+ "mel_fmin": 0.0,
38
+ "mel_fmax": 22050
39
+ },
40
+ "model": {
41
+ "inter_channels": 192,
42
+ "hidden_channels": 192,
43
+ "filter_channels": 768,
44
+ "n_heads": 2,
45
+ "n_layers": 6,
46
+ "kernel_size": 3,
47
+ "p_dropout": 0.1,
48
+ "resblock": "1",
49
+ "resblock_kernel_sizes": [
50
+ 3,
51
+ 7,
52
+ 11
53
+ ],
54
+ "resblock_dilation_sizes": [
55
+ [
56
+ 1,
57
+ 3,
58
+ 5
59
+ ],
60
+ [
61
+ 1,
62
+ 3,
63
+ 5
64
+ ],
65
+ [
66
+ 1,
67
+ 3,
68
+ 5
69
+ ]
70
+ ],
71
+ "upsample_rates": [
72
+ 8,
73
+ 8,
74
+ 2,
75
+ 2,
76
+ 2
77
+ ],
78
+ "upsample_initial_channel": 512,
79
+ "upsample_kernel_sizes": [
80
+ 16,
81
+ 16,
82
+ 4,
83
+ 4,
84
+ 4
85
+ ],
86
+ "n_layers_q": 3,
87
+ "use_spectral_norm": false,
88
+ "gin_channels": 256,
89
+ "ssl_dim": 256,
90
+ "n_speakers": 200
91
+ },
92
+ "spk": {
93
+ "kasugano_sora_mix": 0
94
+ }
95
+ }
svs-fork/mixed/trained_from_speak_G40000/D_110000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:535c133dc26837c48dcb838d7d637c9bc3d2ca9d93387c90e25f3d495454ceed
3
+ size 561098249
svs-fork/mixed/trained_from_speak_G40000/G_110000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e91cca6e913a7bf3951c01a6804076520a9d22b4b0685f08f451051ea121a24a
3
+ size 542789469
svs-fork/speak/G_26000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f583fec93bc559ddb4ab8a154ca64c0c984d8e56a37e0e9599b4c528644fa961
3
+ size 542789469
svs-fork/speak/G_40000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57a0c76214a225577ff579ec04e77d738fe59109026236c95d8eefe61d2fb799
3
+ size 542789469
svs-fork/speak/G_6000.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a469a160fcd31ea0f0b8186bbb931e19b549fe342dea96d66c23ab0da71fa03d
3
+ size 542789469