Upload folder using huggingface_hub
Browse files- .gitignore +1 -0
- README.md +19 -0
- config.json +105 -0
- style_vectors.npy +3 -0
- tsukuyomi-chan_e116_s3000.safetensors +3 -0
- tsukuyomi-chan_e154_s4000.safetensors +3 -0
- tsukuyomi-chan_e193_s5000.safetensors +3 -0
- tsukuyomi-chan_e200_s5200.safetensors +3 -0
- tsukuyomi-chan_e39_s1000.safetensors +3 -0
- tsukuyomi-chan_e77_s2000.safetensors +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
.ipynb_checkpoints
|
README.md
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- jp
|
4 |
+
tags:
|
5 |
+
- つくよみちゃん
|
6 |
+
- StyleBertVITS2
|
7 |
+
license: "other"
|
8 |
+
---
|
9 |
+
|
10 |
+
# つくよみちゃんつくよみちゃんコーパス 学習モデル for StyleBertVITS2
|
11 |
+
|
12 |
+
[つくよみちゃんコーパス│声優統計コーパス(JVSコーパス準拠)](https://tyc.rei-yumesaki.net/material/corpus/)を使わせていただき、[StyleBertVITS2](https://github.com/litagin02/Style-Bert-VITS2)で学習を行ったモデルです
|
13 |
+
|
14 |
+
## 学習パラメータ
|
15 |
+
* バッチサイズ : 4
|
16 |
+
* エポック数 : 200
|
17 |
+
* 何ステップごとに結果を保存するか : 1000
|
18 |
+
* bfloat16 : true
|
19 |
+
* プロセス数 : 72
|
config.json
ADDED
@@ -0,0 +1,105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"model_name": "tsukuyomi-chan",
|
3 |
+
"train": {
|
4 |
+
"log_interval": 200,
|
5 |
+
"eval_interval": 1000,
|
6 |
+
"seed": 42,
|
7 |
+
"epochs": 200,
|
8 |
+
"learning_rate": 0.0002,
|
9 |
+
"betas": [
|
10 |
+
0.8,
|
11 |
+
0.99
|
12 |
+
],
|
13 |
+
"eps": 1e-09,
|
14 |
+
"batch_size": 4,
|
15 |
+
"bf16_run": true,
|
16 |
+
"lr_decay": 0.99995,
|
17 |
+
"segment_size": 16384,
|
18 |
+
"init_lr_ratio": 1,
|
19 |
+
"warmup_epochs": 0,
|
20 |
+
"c_mel": 45,
|
21 |
+
"c_kl": 1.0,
|
22 |
+
"skip_optimizer": false,
|
23 |
+
"freeze_ZH_bert": false,
|
24 |
+
"freeze_JP_bert": false,
|
25 |
+
"freeze_EN_bert": false,
|
26 |
+
"freeze_style": false
|
27 |
+
},
|
28 |
+
"data": {
|
29 |
+
"training_files": "Data/tsukuyomi-chan/train.list",
|
30 |
+
"validation_files": "Data/tsukuyomi-chan/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 |
+
"num_styles": 1,
|
43 |
+
"style2id": {
|
44 |
+
"Neutral": 0
|
45 |
+
},
|
46 |
+
"spk2id": {
|
47 |
+
"tsukuyomi-chan": 0
|
48 |
+
}
|
49 |
+
},
|
50 |
+
"model": {
|
51 |
+
"use_spk_conditioned_encoder": true,
|
52 |
+
"use_noise_scaled_mas": true,
|
53 |
+
"use_mel_posterior_encoder": false,
|
54 |
+
"use_duration_discriminator": true,
|
55 |
+
"inter_channels": 192,
|
56 |
+
"hidden_channels": 192,
|
57 |
+
"filter_channels": 768,
|
58 |
+
"n_heads": 2,
|
59 |
+
"n_layers": 6,
|
60 |
+
"kernel_size": 3,
|
61 |
+
"p_dropout": 0.1,
|
62 |
+
"resblock": "1",
|
63 |
+
"resblock_kernel_sizes": [
|
64 |
+
3,
|
65 |
+
7,
|
66 |
+
11
|
67 |
+
],
|
68 |
+
"resblock_dilation_sizes": [
|
69 |
+
[
|
70 |
+
1,
|
71 |
+
3,
|
72 |
+
5
|
73 |
+
],
|
74 |
+
[
|
75 |
+
1,
|
76 |
+
3,
|
77 |
+
5
|
78 |
+
],
|
79 |
+
[
|
80 |
+
1,
|
81 |
+
3,
|
82 |
+
5
|
83 |
+
]
|
84 |
+
],
|
85 |
+
"upsample_rates": [
|
86 |
+
8,
|
87 |
+
8,
|
88 |
+
2,
|
89 |
+
2,
|
90 |
+
2
|
91 |
+
],
|
92 |
+
"upsample_initial_channel": 512,
|
93 |
+
"upsample_kernel_sizes": [
|
94 |
+
16,
|
95 |
+
16,
|
96 |
+
8,
|
97 |
+
2,
|
98 |
+
2
|
99 |
+
],
|
100 |
+
"n_layers_q": 3,
|
101 |
+
"use_spectral_norm": false,
|
102 |
+
"gin_channels": 256
|
103 |
+
},
|
104 |
+
"version": "1.3"
|
105 |
+
}
|
style_vectors.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c3b8cda77c40697e8548eea0fa74bcf6bd107703fd9bb07acd4d169a6858a57b
|
3 |
+
size 1152
|
tsukuyomi-chan_e116_s3000.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:20f60dfc5893457280d9e2539380c189b4d1f292359f559f13f88b617d8c917b
|
3 |
+
size 198768188
|
tsukuyomi-chan_e154_s4000.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:badc75f775f2faf8a80bdddc01040bdcc6a19f5410834dae1936dda6885246b5
|
3 |
+
size 198768188
|
tsukuyomi-chan_e193_s5000.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f9f49f782ceccec5699675cafcb64a0ddd5f39cfd13be9b735c29891400d5714
|
3 |
+
size 198768188
|
tsukuyomi-chan_e200_s5200.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bdf558d05cc6c56fdc5fa295e290b6f916ae5fc367ffe0c8c0091850cdf17c8c
|
3 |
+
size 198768188
|
tsukuyomi-chan_e39_s1000.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:86b38d3a10aa0d70ee34b000264e746304aedbda1cf12a3c550c54ed365708a2
|
3 |
+
size 198768188
|
tsukuyomi-chan_e77_s2000.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bbb4641c022830be5c300c582341966d8549f2c549f7edb0ae2c8a6199653c10
|
3 |
+
size 198768188
|